Opening up the hearing to debate the 14th Amendment, the amendment’s author Senator Jacob Howard states:

The first amendment is to section one, declaring that “all persons born in the United States, and subject to the jurisdiction thereof are citizens of the United States and of the States wherein they reside.” I do not propose to say anything on that subject except that the question of citizenship has been so fully discussed in this bodes not to need any further elucidation, in my opinion. This amendment which I have offered is simply declaratory of what I regard as the law of the land already, that every person born within the limits of the United States, and subiect to their iurisdiction, is by virtue of natural law and national law a citizen of the United States. This will not, of course, include persons born in the United States who are foreigners, aliens, who belong to the families of ambassadors or foreign ministers accredited to the Government of the United States, but will include every other class of persons. It settles the great question of citizenship and removes all doubt as to what persons are or are not citizen’s of the United States.

This language is enshrined in the 14th amendment as

“All persons born or naturalized in the United States, and subject to the jurisdiction thereof, are citizens of the United States”

That phrase, interpreted by Justice Gray in the case of Wong Kim Ark is what gives us the broad birthright citizenship that the United States has today.

Law of the Land

So what was the law of the land that Howard mentions in his opening statement?

When Howard introduced the citizenship clause on May 30, 1866 the law of the land was the 1866 Civil Rights Bill which had passed roughly 7 weeks earlier (over President Johnson’s veto) on April 9, 1866.

The bill contained similar language to what would be enshrined in the 14th amendment. It states:

All persons born in the United States and not subject to any foreign power, excluding Indians not taxed, are hereby declared to be citizens of the United States.

So if they were passed in such a short gap from each other, why do they use different language?

Justice Alito had the same question at the Trump vs. Barbara Supreme Court hearing.

The Hearing

Early in his second term, Donald Trump issued an executive order effectively ending the broad interpretation of birthright citizenship that has been in place since the Justice Gray’s decision in the case of Wong Kim Ark in 1898. This argument reached the Supreme Court in Trump vs. Barbara.

At that hearing on April 1, 2026, the ACLU’s own lawyer, the person charged with arguing for broad birthright citizenship, Cecilia Wang conceded to Justice Alito’s question on whether the difference in phrasing was simply the author of the amendment, Senator Howard, switching the language of the law of the land from the Civil Rights Act into the affirmative. She conceded it was, but couched this in an argument that the authors wanted to capture the common law exceptions and the tribal exceptions.

JUSTICE ALITO: Ms. Wang, would you agree that the citizenship test in the Fourteenth Amendment is the same as the test in the 1866 Civil Rights Act?

MS. WANG: So the words are obviously different. What Wong Kim Ark tells us and what the debates tell us is that the Framers, there was — it was the same Congress, obviously, framing both. Congress was trying to do the same thing with both the 1866 Act and with the Fourteenth Amendment. They wanted to capture the common law exceptions and the Indian tribal exception. They started out with the two separate phrases, “not subject to any foreign power” plus excluding “Indians not taxed.” And as Justice Gray described it in his majority opinion in Wong Kim Ark, they decided to switch to the affirmative phrase, “subject to the jurisdiction” and —

JUSTICE ALITO: Yeah. Well, do they mean the same thing? And wouldn’t it be very odd if the citizenship test in the Fourteenth Amendment were broader than the citizenship test in the 1866 Civil Rights Act, particularly in the light of the fact that the 1866 Civil Rights Act was reenacted after the adoption of the Fourteenth Amendment and remained in place until 1940?

MS. WANG: Sure. I — the Framers were trying to do the same thing with the language in both.

JUSTICE ALITO: Okay. So then I think we can turn to the language of the 1866 Civil Rights Act because it’s more straightforward. You know, “subject to the jurisdiction thereof” is like the — you know, the puzzle wrapped in an enigma wrapped in a mystery.

(Laughter.)

JUSTICE ALITO: But “not subject to any foreign power” is pretty straightforward. So let me give you these examples. A boy is born here to an Iranian father who has entered the country illegally. That boy is automatically an Iranian national at birth, and he has a duty to provide military service to the Iranian government. Is he not subject to any foreign power?

MS. WANG: Not within the meaning of the 1866 Act, Justice Alito. And that’s clear from Wong Kim Ark, and it’s clear from the debates. What the Framers meant by the phrase “not subject to any foreign power” was [p.108] referring to the ambassador exception. If it meant what the government contends, basically, not a subject of any foreign power, that you were — that another country considers you a jus sanguinis citizen, then lawful permanent residents, all foreign nationals, should be excluded.

Put simply Wang’s argument is that the intended phrase is anyone born in the US is a citizen unless they are subject to a foreign power and by foreign power that means a foreign minister or an ambassador.

The Ambassador Exception

Justice Thomas assembled much of this evidence in his dissent in Trump vs. Barbara. He argued the drafters meant to exclude temporary residents. But he never asked the prior question, which we are asking right here. Where did the ambassador exception come from in the first place?

So where does that ambassador exception come from? It comes from Justice Gray’s 1898 opinion in Wong Kim Ark.

And where does Justice Gray derive that? Wang’s answer: “it came from common law exceptions.” Who applied those common law exceptions? It was Justice Gray who added them in his 1898 opinion on Wong Kim Ark. Here are the exceptions, referenced by Brett Kavanaugh in his opinion in Trump vs. Barbara.

This Court’s 1898 Wong Kim Ark decision interpreted the “subject to the jurisdiction” language of the Fourteenth Amendment. United States v. Wong Kim Ark, 169 U. S. 649. That decision adopted a general rule of birthright citizenship for those born in the United States—with four disparate exceptions for “children of foreign sovereigns or their ministers, or born on foreign public ships, or of enemies within and during a hostile occupation of part of our territory” and “children of members of the Indian tribes.” Id., at 693; see also Elk v. Wilkins, 112 U. S. 94, 102 (1884). Importantly, the Wong Kim Ark decision treated those four exceptions as a closed set for constitutional purposes, meaning that Congress could not create other exceptions to birthright citizenship, although Congress could eliminate one or more of those four exceptions. See 169 U. S., at 682, 693.

It’s worth noting that Kavanaugh’s opinion did not disqualify Trump’s EO on grounds that it violated the 14th. Instead Kavanaugh decided that Trump’s EO was invalidated because two laws passed by Congress after Wong Kim Ark contain this same “subject to the jurisdiction thereof” language and therefore enshrined the Wong Kim Ark decision over Trump’s EO. He said that if Congress was to change these to define additional exceptions, then Trump’s EO would be valid.

So where do these exceptions come from? There is no mention of invading armies or foreign ships in the records debates of the 14th. My reading of the debates for the 14th amendment shows only one true reference to anything resembling an ambassador exception. That reference is here, in the Howard statement I noted at the beginning:

That every person born within the limits of the United States, and subject to their jurisdiction, is by virtue of natural law and national law a citizen of the United States. This will not, of course, include persons born in the United States who are foreigners, aliens, who belong to the families of embassadors or foreign ministers.

Does this phrase indicate an exception for ambassadors to you? If foreigners is describing an ambassador, what is alien doing?

So where did Gray turn to? He turned to English common law.

Wang alluded to this in her statement to Alito.

What Wong Kim Ark tells us and what the debates tell us is that the Framers, there was — it was the same Congress, obviously, framing both. Congress was trying to do the same thing with both the 1866 Act and with the Fourteenth Amendment. They wanted to capture the common law exceptions and the Indian tribal exception. They started out with the two separate phrases, “not subject to any foreign power” plus excluding “Indians not taxed.”

Notice what Wang is doing here, the first thing she does is to reference Gray’s decision in Wong Kim Ark. This is interesting because the framers of the 14th didn’t seem to mention those exceptions in the debates. The one they did mention wasn’t exactly clear. The exceptions were pulled out of common law by Gray for his opinion in Wong Kim Ark.

Notably, Cecilia Wang does mention the 1866 Civil Rights Law and to re-iterate she says they are trying to do the same thing. But in her rendition, this “same thing” includes the common law exceptions.

Was this what they were trying to do?

The Civil Rights Act of 1866

All persons born in the United States and not subject to any foreign power, excluding Indians not taxed, are hereby declared to be citizens of the United States.

The broad birthright citizenship interpretation hinges on the foreigners exception from the Civil Rights Act being extremely narrow: only ambassadors. There is no clarification on this in the actual debate of the 14th Amendment.

Maybe they were simply tired of debating the topic? That seems to be the mood of the amendment’s author Senator Howard:

I do not propose to say anything on that subject except that the question of citizenship has been so fully discussed in this bodes not to need any further elucidation, in my opinion. This amendment which I have offered is simply declaratory of what I regard as the law of the land already.

Howard hands the Senate the final text and immediately tells them there’s nothing to debate, they’d already had the debate. Forty-one days earlier, on April 9, 1866, this same body had passed the Civil Rights Act over Andrew Johnson’s veto. The men in that room had voted on it, fought the President over it, and argued for weeks about how to word its citizenship provision. Howard sat through all of it.

Thirty two years later, Justice Gray took Howard’s words and told everyone what that meant.

In the forefront both of the Fourteenth Amendment of the Constitution and of the Civil Rights Act of 1866, the fundamental principle of citizenship by birth within the dominion was reaffirmed in the most explicit and comprehensive terms.

He continues by restating not the language from the 14th Amendment but the language from the Civil Rights Act.

The Civil Rights Act, passed at the first session of the Thirty-ninth Congress, began by enacting that "all persons born in the United States, and not subject to any foreign power

The connection from the Civil Rights Act to the 14th Amendment is very clear to Gray. His opinion continued:

The same Congress, shortly afterwards, evidently thinking it unwise, and perhaps unsafe, to leave so important a declaration of rights to depend upon an ordinary act of legislation, which might be repealed by any subsequent Congress, framed the Fourteenth Amendment of the Constitution

Gray stated that the authors of the 14th thought their declaration in the Civil Rights Act was so important that they did the amendment to lock it in place.

But the Civil Rights Act doesn’t contain an exception for ambassadors. So where did the ambassador exception come from? To make this work, Gray needed to redefine “not subject to any foreign power” to fit a much narrower group of people, so he added exceptions.

Where did these exceptions come from? They didn’t come from the Civil Rights Act debate, the law of the land that Howard spoke of. No, these exceptions came from “the law existing before its adoption” which turns out to mean Calvin’s Case, Blackstone, Cockburn, Dicey — three centuries of English common law.

The Framer’s Intent

I can’t speak for Gray, but we do have access to the debates from the actual Civil Rights Act of 1866. The text from those debates that so tired Howard. The debate that produced the bill that became the law of the land that Howard was merely trying to make declaratory.

The text of those debates, the text of what the framers of the 14th amendment actually thought, completely destroy any argument, made by Gray or Wang, that the intent was to narrowly apply “not subject to any foreign power” to only ambassadors (or invading armies or people on ships). No, the debate from the Civil Rights Act of 1866 is very clear.

This is how the Senators arrived at the precise wording for the Civil Rights Act of 1866, the wording that Howard was merely trying to make declaratory. From Senator Trumbull:

Mr. TRUMBULL. The Senator from Missouri assumes that here is a sort of property qualification to citizenship. Such is not the meaning of the provision. The Senator from Missouri and myself desire to arrive at the same point precisely, and that is to make citizens of everybody born in the United States who owe allegiance to the United States. We cannot make a citizen of the child of a foreign minister who is temporarily residing here. There is a difficulty in framing the amendment so as to make citizens of all the people born in the United States and who owe allegiance to it. I thought that might perhaps be the best form in which to put the amendment at one time, “That all persons born in the United States and owing allegiance thereto are hereby declared to be citizens;” but upon investigation it was found that a sort of allegiance was due to the country from persons temporarily resident in it whom we would have no right to make citizens, and that that form would not answer. Then it was suggested that we should make citizens of all persons born in the United States not subject to any foreign Power or tribal authority. The objection to that was, that there were Indians not subject to tribal authority who yet were wild and untamed in their habits, who had by some means or other become separated from their tribes and were not under the laws of any civilized community, and of whom the authorities of the United States took no jurisdiction. The Senator from California [Mr. CONNESS] told us that there were in his State Indians who had been placed upon reservations under charge of Indian superintendents who had been separated from their tribes and were not under any tribal authority, but they were there under the regulations of treaties which had been made with them, and were supplied and looked after by our Indian agents the same as other Indians who were perfectly wild, not submitting at all to the usages of civilized life, and it could not be intended to make that class of persons citizens.

Then it was proposed to adopt the amendment as it now stands, that all persons born in the United States not subject to any foreign Power, excluding Indians not taxed, shall be citizens.

Here’s the actual text from the Congressional Record.

That was the process by which they arrived at the that all persons born in the United States not subject to any foreign Power, excluding Indians not taxed, shall be citizens. The phrase made declaratory in the 14th amendment.

The Senators started by deliberately excluding ambassadors by filtering to only those who owed allegiance to the United States.

“That all persons born in the United States and owing allegiance thereto are hereby declared to be citizens;”

But the Senators determined that while that excluded ambassadors, it needed to go further, it needed to specifically exclude temporary residents. The exception needed to be bigger.

Trumbull says it in his own words.

But upon investigation it was found that a sort of allegiance was due to the country from persons temporarily resident in it whom we would have no right to make citizens.

From there the Senators changed the phrase to “not subject to any foreign Power.”

Then it was suggested that we should make citizens of all persons born in the United States not subject to any foreign Power or tribal authority.

This would not have been suitable to Gray’s argument. This would not have been suitable to Wang’s argument.

This language very clearly defines the scope of the 1866 Civil Rights Bill as only granting citizenship to persons born in the United States that aren’t subject to any Foreign Power.

The framers felt that they would have no right to make American citizens of the children of people who were only temporarily residing in the United States.

What’s Left to Debate?

This was the debate that formed the law of the land at the time. The law of the land that Senator Howard was seeking to make declaratory in the 14th amendment, the law he was merely trying to codify into an amendment.

The Senators weren’t reaffirming a common-law principle. They had spent February debating the actual phrase. They rejected two formulations before settling on the third. Howard was codifying the product of that debate. Gray says as much himself: the same Congress, “thinking it unwise, and perhaps unsafe, to leave so important a declaration of rights to depend upon an ordinary act of legislation, which might be repealed by any subsequent Congress, framed the Fourteenth Amendment.” They were protecting the Act. Which means the Act is what they were protecting.

And despite those efforts, Gray and others in the Supreme Court voided it all anyways. They inserted Common Law traditions into the amendment that was created for the purpose of codifying a law that deliberately left those common law traditions out.

You may prefer Gray’s reading, but his interpretation in Wong Kim Ark is not what the framers of the 14th intended.

The ambassador wasn’t the exception. The ambassador was a small piece of a much larger exception that covered every child born subject to a foreign power. As Trumbull said: “we would have no right to make citizens” of persons temporarily resident in it.

These are the facts. This is the closest reading. This was the intent of the framers of the 14th amendment. They may be uncomfortable. They may be hard to stomach.

They are not deniable.

Wong Kim Ark rests on a reading the framers never intended. It’s a common law gloss grafted onto an amendment that was written to keep this exact question out of common law, and out of Congress.

That last part is what Kavanaugh gets backwards. He says Congress has to pass a bill to add these exceptions. But there is nothing to add. The exclusion of those subject to a foreign power was in the phrase from the start. The Senators put it there on purpose. They rejected the broader formulations that would have left it out. Gray is the one who took it back out.

And the statute Kavanaugh leans on has no meaning of its own. §1401(a) borrowed “subject to the jurisdiction thereof” straight from the Amendment. It means what that phrase means, nothing more. Overrule Wong Kim Ark and the phrase goes back to the framers intent. That intent was “not subject to any foreign power.” That doesn’t mean an exception for Ambassadors, it means all of those who are “temporary resident” in America. This is how “subject to the jurisdiction thereof” should be interpreted in the Congressional statute and in the Constitution. There is no need for Congress to act. Congress built it’s laws on the phrase from the 14th Amendment, not on Gray’s gloss of it. A statute that borrows a constitutional phrase borrows the phrase’s meaning, and it can’t freeze a reading the Court has since said was wrong.

This is the Court’s mistake to fix, not the legislature’s. The framers took this out of Congress’s hands on purpose. Sending it back to Congress just defers to the same error that needs correcting.

The only thing that makes this hard is the part you can’t wave away. For 128 years people were born into citizenship under Gray’s reading and built their entire lives on it, having done nothing but be born. Gray was wrong, and we have lived under his mistake for 128 years. Both of those are true. An honest fix has to carry the weight of both. But the second one doesn’t make the first any less a fact.

They are correct that AI data centers increase utilization. But that’s where the reality of their story ends.

You don’t have to take my word for it. The Federal Reserve put it in the minutes of its June meeting:

Many participants noted that ongoing strong demand for AI infrastructure would likely sustain upward pressure on prices for technology products and electricity.

That’s the FOMC. Not an activist group. Not a utility. The people whose job is watching prices, writing down that AI infrastructure demand is pushing electricity up.

They didn’t explain the mechanism. So let’s do that.

It’s all about price.

The decision to run or not run a power plant comes down to one thing. Is the price of the output higher than the cost of the input?

For anything that burns fuel, that comes down to the heat rate: how much fuel it takes to make a unit of electricity. It’s an efficiency measure. Take the heat rate multipled by the fuel cost and you get the cost to make power. If the market price is above that, it’s in the money and they run. If it's below, it’s out of the money and they shut off (if they can).

That's the decision. Every hour. Every plant.

And because they all make it independently, the system sorts itself into a stack, cheapest to run at the bottom, most expensive at the top. Wind and solar sit at the very bottom, because their cost to produce one more unit is basically nothing. Nuclear next, for the same reason. Then efficient gas. And at the very top, the plants that only exist for the worst hours of the year: essentially jet engines that burn a lot of fuel. These are called peakers. As batteries advance those are taking market share from these peakers. Buy power when it’s cheap, sell when it’s expensive.

The Margin

Everybody gets paid the price of the last plant you have to turn on. Not their own cost. The last one.

So on a mild Tuesday when you only climb halfway up the stack, everybody earns the halfway price. On a brutal August afternoon when you have to reach all the way to the top, everybody earns the peaker's price — including the wind farm whose costs didn't change at all.

Demand doesn't just get served. Demand sets the price, for everyone, by deciding how far up the stack you have to climb.

Demand response is what it calls when the price causes demand to drop (shed) and as a result, the required supply drops, leading to lower prices.

“Steady”

The whole case for data centers rests on that one word, and it's doing a job it can't do.

A customer who shows up at four in the morning in April is free money. The plants are idling anyway. Demand is low, the power plants supplying the power are the cheapest to operate. These are called base load, think of a nuclear plant. You're at the bottom of the stack. Nobody notices. That customer genuinely does lower your rates. The power has to be used.

But steady doesn't mean off-peak. Steady means constant. Four in the morning and also five o'clock on the worst afternoon in August, when you were already scraping the top of the stack before they showed up.

They don't fill the valley. They raise the valley and the mountain by the same amount.

And the mountain is the only part anybody has to build for.

Traffic

AI proponents like to compare power grids to highways. They are out there filling in, increasing utilization of the road. They’ll pay for the highways too. They will make your commute cheaper, because they run at three in the morning when the road's empty. Fine. Night freight really does work like that.

But these trucks run at three in the morning and at five o'clock, in the jam, every day, forever. Because a truck sitting still doesn't earn.

You can't relieve congestion with traffic that shows up during the traffic.

No Time to be Idle

The obvious fix: run at night. Back off a few hundred hours a year when the grid is screaming. Everybody wins.

They won't. The expensive part of a data center isn't the electricity, it's the chips — billions of dollars of hardware that starts losing value the day it's installed and is close to worthless in a few years. They’re like airliners. Every idle hour is that clock running with nothing to show for it.

So they can't idle. The economics that make the buildout urgent are the same economics that make it inflexible.

The one behavior that would make their story true is the one behavior the business model forbids.

Renewables

Data Center proponents will argue that your beef is not with the data center, it’s with the liberals who pushed for more use of renewables to decarbonize the grid.

Liberals subsidized solar and wind. It worked. Renewables got built, and they went into the grid. And they have a property that changes everything: their cost to make one more unit is essentially zero. If the wind is blowing or the sun is shining they run. They are always at the bottom of the stack.

When they’re running.

And this is their problem, they are intermittent. Sometime they produce too much and price has to go negative to balance. Users get paid to use power.

But then sun sets or the wind drops and the power has to come from elsewhere. Batteries are doing some of this, but usually someone has to turn on something that burns something to cover that gap.

INewer combined-cycle units ramp up and down quickly. That’s what they’re built to do. Price goes negative, you shut down, you come back in three hours, no harm done.

If you’re a coal plant, that’s a catastrophe. A coal plant is not a light switch. There’s a minimum load below which the boiler is unstable. Ramping it up and down subjects it to thermal stress that wrecks the equipment. Shutting down and restarting takes many hours and costs real money.

Shale gas made natural gas fired plants much cheaper, these plants began to compete more and more with coal in the stack.

Coal powered plants were pushed down in the stack and onto the margin much more. Coal plants aren’t designed to be on the margin, when they are out of the money a coal plant often can’t leave. It runs anyway. It sells below its own cost, because eating that loss is cheaper than the alternative.

The system asked for flexibility and coal doesn’t have any.

Through the 2010s and into this decade, American electricity demand was flat. Basically didn’t move. Efficiency ate the growth as fast as it arrived.

Every new solar farm and wind turbine wasn’t serving new demand. There wasn’t any. It was displacing something that already existed.

The coal plants stopped covering their costs, and they closed.

You’ll be told this was a policy disaster inflicted by activists. It was arithmetic. Nobody keeps a plant running against demand that isn’t there. The market looked at what America actually needed and trimmed to fit, which is exactly what a market is supposed to do.

The nuclear closures were mostly a separate story with the same shape. Old plants, decades on the clock, needing serious capital to keep going, in a market where that cheap shale gas had crushed the price they could earn. They aged out into a market with no room for them.

Nobody was wrong. That’s the part that matters. The plants that closed should have closed, given the demand that existed.

Bad Timing.

Picture the perfect customer for a coal plant. It would buy power around the clock, never varying, never asking the plant to ramp. It would take every single megawatt, every single hour, at four in the morning and at noon and at midnight. It would be, in the exact technical sense, inflexible, which is normally a flaw, but paired with an inflexible plant it would be perfect.

That customer is an AI data center.

It is precisely, almost comically, the thing that would have kept those plants alive. A 24/7 baseload buyer arriving in the 2010s would have soaked up the exact output that coal couldn’t stop producing.

The grid optimized away the thing AI needed, because nothing needed it yet. Every decision was correct. The outcome is a grid with no slack.

What it all means

The story you’re told is that data centers are big, steady customers who fill up a half-empty grid, spreading the cost across more electricity and lowering your rate.

But steady doesn’t mean off-peak. Steady means always. Four in the morning, yes, and also five o’clock on the worst afternoon in August, when you were already scraping the top of the stack before they showed up.

The grid that came out of the 2010s is exactly the right size for us, with nothing to spare.

Then AI showed up with a massive demand for power, always more, some data centers demand more power than all of New York City. AI won’t wait for the cheap hours or cut its demand if prices get too high.

It doesn’t fill the valley, it jacks up the valley and the mountains alike. It shifts the whole curve up.

And when you shift the curve up, you climb further up the stack. You reach plants that cost more to run. And the last plant you turn on sets the price for everybody.

That’s it. That’s the whole mechanism.

AI data centers don’t just take our unused capacity. AI moves us into a more expensive part of the stack, in every hour, forever.

And everybody on the system pays the price for the amount of power AI demands.

Bill Gurley's response was that plenty of states have open arms, so he's cool with it. Let New York be a laboratory of democracy. And then this:

Since people in these places do use AI, I do wonder if they should have to pay more.

Read that again, because he just gave away the whole thing.

The premise he needs

That argument only works one way. New Yorkers are getting something without carrying the cost of it. They're freeloading. The states with open arms are hosting the load, and New York is enjoying the output for free.

Fine. But you cannot freeload on a benefit.

If hosting a data center makes your power cheaper, New York didn't dodge a burden. New York turned down a gift. There's nothing to freeload on and nothing to charge them extra for. Gurley's penalty needs a cost to exist. He built the cost into his own argument and then handed it to us.

So somebody near a data center is paying something. That's his premise, not my accusation. He can say he never wrote it. He can't run the argument without it.

Your Subscription Isn’t Covering The Cost

We already pay for AI. It's called a subscription. Twenty bucks a month. You provide money, they provide a service that is paid for with that money.

So if the subscription covers the cost entire, New York subscribers are paying the same as Virginia subscribers. Everyone is paying the same cost and receiving the same benefit. No freeloading.

Which means Gurley is telling you the subscription doesn't cover it. If New Yorkers need to pay more for their subscriptions that means someone else must be paying more than the cost of the subscription. Who would the New Yorkers be freeloading off of?

That's the quiet part. The price is wrong. Somebody else is eating the difference.

And that someone else is every single person in a state where a Data Center is built. Every single person who has to compete with those data centers and pay higher prices for power to than they would if those plants weren’t there. That someone is you.

Who’s the freeloader?

It’s the VCs, it’s the tech CEO, the guys who are taking your subscription revenue and then hitting you with other costs that you aren’t being compensated for.

Because if New Yorkers need to pay more for AI, it’s not because they are underpaying than these AI companies. It’s because the person who lives in the areas where these tech companies are operating their data centers is making a payment outside of their subscription.

Socialize the costs.

Privatize the gains.

He didn't confess to it. He just built an argument that doesn't stand up without it.

David Sacks was the White House AI Czar, he said it directly:

If the U.S. were to lose the AI race, the repercussions would be significant, and it would alter the balance of power in the world in a very unfavorable way.

This argument is how they turn a rising electric bill into a patriotic duty.

This argument is the one Kevin O'Leary ( building the world's largest proposed data center in Utah) made to Tucker Carlson:

If they don't think my motivation is to beat the Chinese, they're wrong. I am going to beat them.

This argument makes the cost sound like a duty instead of a bill. It’s the only way for them to get us to subsidize a technology that they’ve told us is going to put millions out of work. Fear of losing does the work that no promise of gain could do. It manufactures your consent.

But look beyond their words and you will quickly find that they don’t believe their own argument.

If you really wanted to beat China, you would never sell them your AI models.

If you really wanted to beat China, you would never let them build their own AI models here.

And yet they do.

Socialize the Costs

Start with what the buildout is costing you, because that part is no longer up for debate.

On July 9, the Federal Reserve released the minutes of its June meeting. The Fed said:

Many participants noted that ongoing strong demand for AI infrastructure would likely sustain upward pressure on prices for technology products and electricity.

The Fed put in writing that the AI buildout is pushing up the price of electricity enough to show up in national inflation.

You are paying the costs of AI development, but how is what you are paying for being used?

Selling the Rope

On the same week the Fed published its electricity finding, the Financial Times reported that OpenAI and Google have been selling their advanced AI models to Singapore-based subsidiaries of Alibaba, Baidu, and Tencent.

Those are not random Chinese firms. They are on the Pentagon's 1260H blacklist, a congressionally mandated list of companies the U.S. government ties to the Chinese military.

The sales are legal. That is the whole problem. Washington restricted the chips. It restricted a handful of individual models. It never barred Chinese-headquartered companies from using cutting-edge American AI software, even the ones on a military blacklist. So the same industry telling you to swallow the power bill to beat China has been quietly selling to the exact firms the argument names as the threat.

And it was not clean. OpenAI cut off Alibaba-linked access to its models only after the FT asked about it, over suspected "distillation," the practice of using one model's output to train a competing one. It doesn't even require a sale. Anthropic told Congress that Alibaba ran 25,000 fake accounts to pull more than 28 million exchanges out of its model to do exactly this.

Think about what that means. The models are trained on GPUs. The GPUs run on electricity. The electricity is the electricity you are now paying a premium for, the premium the Fed just confirmed. You are footing the power bill to build the lead that we’ve been told we must build. Then the AI companies went off and sold what you paid for in higher power prices to China.

Chris McGuire spent years working on export controls at the White House and the State Department. His verdict: the administration says "we need to beat China on AI all the time, but the problem is they haven't done anything on export control, which is the actual tool we have to slow China down."

The reason is loud where it makes you pay. It goes silent where it would cost them money.

Right Under Our Nose

Selling the models to China is bad enough. But the same government telling you to pay more to beat China has been letting China build the thing on American soil.

This isn’t new. In November 2024, The Financial Times reported that Alibaba, ByteDance, and Meituan have been standing up AI teams in California, expanding their offices specifically to poach engineers and researchers straight out of OpenAI and the top US labs. Alibaba recruited an AI team in Sunnyvale and told workers it planned to spin the operation into a separate startup. A former OpenAI researcher described being bombarded by Alibaba and Meituan trying to learn what they had worked on. ByteDance runs development on its Doubao model out of California, reporting to a boss in Beijing.

And the chip ban, the one piece of "beat China" that actually had teeth? There were "no restrictions on US-based entities related to or owned by Chinese tech companies accessing high-end AI chips through data centres located in the US." So the export control everyone points to was being routed around on American ground. Build a team in California. Reach the chips through a US data center. Hire the people who trained the American models.

So walk through what "beating China" actually looks like in practice. The chips leak, reached through US data centers by US-based firms owned by Chinese companies. The models leak, sold to blacklisted firms and distilled through fake accounts. The talent leaks, poached out of the frontier labs by Chinese teams built in California, while we import researchers to staff the same labs they are poaching from. Every single control point the security story implies is standing wide open.

How Stupid Do They Think We Are?

If we were serious about beating China, this wouldn’t be happening. But it is happening. Because we aren’t serious about “beating China.” That’s just the lie they tell us when they stick us with the bill. The lie they tell us to manufacture our consent.

The training runs eat enormous amounts of power. The Fed now says that power demand is pushing up your electric bill. You were told to accept that. You were told that AI would take your job, until that warning got inconvenient and they walked it back. And throughout it all, you’ve been told me have to beat China in AI. The Government has allowed China companies to build AI right here in the US, training their models on power from our power grid, jacking up the price you pay for power. The “American” companies also jacked up your power prices while the Government jusifies it by saying “we need to beat China”then those companies sold their models to China. "Beat China" is real when it makes you pay the bill and silent when export controls would cut the revenue. The power cost is spread across the public. The revenue, China included, stays private.

They sell the fear because fear of losing to an enemy is the only thing strong enough to make you pay for something whose only promised return, to you, is your own replacement.

That is what "beat China" is for. It is not a strategy. It is manufacturing consent. It is the mechanism that gets the American people to subsidize an industry that has told them, plainly, that it is coming for their work. The race is the wrapper. The bill is the product. And the leaks, the models sold to China, the labs built in California, the chips reached through US data centers, the talent poached out the front door, are how you know the people invoking the race do not believe in it themselves.

So ask the only question that matters. If we are being forced to pay more to beat China, why is the technology, sold, built, and poached, ending up in China anyway?

Because it was never about China. It was about getting you to pay. And this is what makes it clear.

It’s not. But what it very conveniently does is leaves birthright citizenship for the children of H-1Bs.

No one reads what Kavanaugh wrote and says that is what the Banks bill does.

Simply put, Banks is lying.

Kavanaugh

This is what Kavanaugh wrote in Trump v. Barbara.

Consistent with the Fourteenth Amendment, Congress could amend §1401(a) or otherwise enact new legislation establishing exceptions to birthright citizenship for children born to foreign citizens unlawfully or temporarily in the country.

The mechanism: amend the statute. Pass an ordinary law. No constitutional amendment, no emergency powers, no calling anyone anything. Write the exception into §1401(a) and go home.

The scope: unlawfully or temporarily. Two groups: kids of people here illegally and kids of people here legally but temporarily. The tourist visa, the student, the H-1B.

Banks

Banks first defines people here illegally as "invaders." Then he strips citizenship from the children of "invaders.

"Invader" isn't in Kavanaugh's opinion. It's not in the opinion Banks says he's following. He didn't get it from Kavanaugh. Kavanaugh's whole point was that a plain law is enough. This isn’t what Kavanaugh asked for. You don't build an invasion declaration to walk through a door someone told you is already open.

Look at his rankings by the tech bro run “Club for Growth.” Banks is a staunch defender of Economic Freedom.

The Opening Kavanaugh didn’t provide.

Kavanaugh's sentence covers people here legally but temporarily. Banks's bill can't touch them.

You can't call an H-1B worker an invader. The government gave him the visa. The word falls apart. So the entire lawful-temporary group, half of what Kavanaugh authorized, drops out of the bill.

Every H-1B, every student, every temporary worker who has a kid here still gets automatic citizenship under Banks's bill. That's the exact thing Kavanaugh said Congress could change. Banks left it out.

A bill that followed Kavanaugh covers both groups. They're both in the sentence. Banks covers one and drops the other, in the direction that sounds tough while sparing the visa labor nobody wants to touch.

There’s two reasons why that happens. A misreading doesn't quote the Justice by name, adopt a word he never used, and delete half his sentence. When you claim a document says something, and it says the opposite, and the gap runs your way every time, there's a shorter word for it.

Jim Banks is masquerading as someone who is protecting Americans.

Jim Banks is pretending to be America First.

But when you actually look at what he’s doing it becomes very clear.

Jim Banks is a liar.

Jim Banks is America Last.

The update that broke your Start menu. The login button that turned invisible and left you dragging your mouse across a blank screen, hunting for something you could not see. Dark mode flashing white every time you open a folder. Windows displaying half its menus in one language and half in another, like a machine having a stroke.

Microsoft is the most valuable software company on Earth. So why does its software feel like it is falling apart in your hands?

The answer is design. The design of the business.

Microsoft stopped selling you a product and started renting you a meter. Once the money arrives whether or not the software works, the incentive to make it work disappears. To widen their margin, Microsoft spent twenty years engineering a cheaper, more compliant workforce, through a visa pipeline it opened by routing around Congress. The bugs on your screen and the Chinese hackers in the nuclear weapons agency come from the same place.

That is the whole argument. The rest is the evidence.

It’s not your imagination.

2025 was the worst year for Windows in over a decade. It is the consensus of nearly every publication that covers the company.

Microsoft shipped bugs on a near-monthly cadence. It broke the Start menu. It broke the Taskbar. It broke File Explorer. The parts of the computer you interact with constantly broke after routine security updates.

There was the bug that split the screen across two languages. The invisible password button Microsoft told you to find by waving your cursor around and feeling for the edges. Dark mode strobing white. Task Manager spawning a fresh copy of itself every time you opened it and never closing any of them as it ate your memory.

Microsoft admitted, in its own documentation, that core features had been broken since the middle of the year. By 2026 it was pulling engineers off new features and throwing them at an emergency effort to make the operating system work again. IT departments reported a 30% jump in Windows support tickets compared to two years earlier.

One engineer summed it up in the trade press. Microsoft's push to ship features was "outpacing their testing capabilities."

Why waste money on QC when the customers will test it for free?

One man’s bug leads to another man’s breach

If buggy only meant annoying, you could live with it. But the same corner-cutting shows up in security, where the stakes are not a frozen Taskbar. They are national.

In 2023, a China-linked group forged authentication tokens and read the email of U.S. government officials, including the Commerce Secretary. The federal Cyber Safety Review Board investigated. This is the body the government created to examine cyber disasters the way it examines plane crashes.

Its verdict was not gentle. The intrusion "should never have happened." It was caused by "a cascade of Microsoft's avoidable errors." And the root was "a corporate culture that deprioritized enterprise security investments and rigorous risk management."

The Board told Microsoft to stop building features until it fixed the fundamentals.

They didn’t.

In July 2025, a flaw in SharePoint, the software organizations use to share internal documents, was exploited by Chinese state hackers. The campaign hit around 400 organizations worldwide. Among the victims was the National Nuclear Security Administration. That is the agency that designs and maintains America's nuclear weapons.

And here is the part that should end the conversation. According to a Reuters investigation, Microsoft knew about the underlying flaw and failed to patch it effectively. Its first fix was incomplete. The attackers walked through the gap it left.

The company whose software broke your Start menu is the same company whose software let a foreign government into the nuclear weapons agency. Same root cause. When quality control is a cost to be minimized, the failures do not stay small.

Microsoft’s greed knows no bounds. For decades, Microsoft used China-based engineers for U.S. military technical support. The "digital escort" program only ended last year following an investigation that revealed Chinese nationals had remote access to Defense Department (DoD) cloud.

The Rest of the Story

The old Microsoft sold you a product.

You bought Windows or Office once every few years. It sat on your shelf. If you never upgraded, Microsoft never saw another dollar from you. To get paid again, they had to make the next version good enough that you actually wanted it.

That was then.

Today's Microsoft rents you everything, forever. Microsoft 365 is a subscription. Azure is billed by the minute. Dynamics, LinkedIn, GitHub, Copilot. All recurring. All metered. Read the company's own financial filings and you will find them organized around the language of the landlord. Seat growth. Subscriber counts. Consumption growth.

The relationship is not a sale, it is a meter.

Everything is downstream of that. A drug dealer doesn’t require satisfaction from their customer either, they just need them to keep coming back. An addict is stuck because of addiction. You’re stuck on Microsoft because they’ve embedded themselves into your workflow. The cost of switching is too high. To keep you hooked they don’t need to make the product better, they need to increase the switching cost.

That means AI, shoved into Notepad and every other place you never asked for it because that can be used to capture your data and make the product sticker. Because that can be used by Microsoft to make more money off of you.

And when the revenue arrives automatically, every dollar not spent is a dollar that goes into your margin.

That is the machine. It is not built to make good software. It is built to make recurring software, at the widest possible margin. Good is optional. Recurring is not.

Who Builds It

Microsoft spent two decades engineering a workforce for what really matters to them: cost.

A large part of how it did that was by building a private immigration pipeline that mostly bypassed Congress.

In the 2000s, Microsoft lobbied hard to raise the cap on H-1B. They wanted more labor they could pay with a life in America instead of a wage alone. But in 2007, the immigration bill supported by the George W. Bush administration, the bill that Microsoft had pushed, the bill to double the H-1B cap collapsed.

Microsoft’s chief lobbyist, Jack Krumholtz urged Secretary of Homeland Security Michael Chertoff to consider a workaround that would bypass the congressional logjam. Krumhz laid the groundwork to expand a program that has become, in effect, the U.S.’s largest pipeline for skilled guest workers with no legislation from Congress. This program is STEM OPT and with no Congressional action at all, a program that allows Microsoft access to an uncapped number of tech workers was created.

With no real wage floor, and almost no oversight the program grew. By 2019, more than a million foreign students were supervised by fewer than 400 federal staff. One former official called it the Wild West of worker programs.

By 2015, this back door was issuing more work authorizations every year than the H-1B program it was built to route around.

Why did Microsoft want it so badly? Two reasons, and neither is that Americans could not do the work.

First, cost. Because OPT workers are technically still students, employers who hire them pay no Social Security or Medicare tax on them. That is a 7.65% discount off the top for the employer alone. If you include the employee it means that for every dollar in wage paid by Microsoft to a STEM OPT, that dollar goes nearly 15% further than it would if it went to an American. The government made a foreign graduate cheaper than an American doing the identical job.

Second, leverage. A worker whose legal right to remain in the country is tied to his job does not negotiate hard and does not walk across the street for a raise. That is not a criticism of those workers. They are doing exactly what any of us would do. It is a description of the incentive Microsoft built for itself. Cheaper. More captive. Higher throughput.

And Microsoft still leans on it. In fiscal 2025 it was the second-largest H-1B sponsor in the entire country, roughly 4,800 to 5,200 approved petitions, behind only Amazon. For the first time ever, four American tech companies swept the top four spots for new visas. Microsoft is also one of the largest OPT employers in the nation.

OPT feeds the front end. H-1B holds the back. Between them, Microsoft runs a two-stage conveyor belt for visa-dependent labor. One stage conjured by regulation. The other by a cap it stopped trying to win and started routing around.

Where The CEO Came From

You can see the whole system in one man.

Satya Nadella runs Microsoft. He was born in Hyderabad, India. He came to the United States in 1988 on a student visa, earned a master's degree, and went to work in tech on exactly the F-1 to OPT to H-1B path this piece has been describing. He is, quite literally, the pipeline's most successful product.

Nadella is the architect of the rented Microsoft. He ran the Server and Tools division. He built Azure. The shift from the boxed product you owned to the metered cloud you rent is his shift, more than anyone's. The business model that pays whether or not the software works, that rewards Copilot over a working Start menu, that treats testing as a cost to be trimmed, is his model.

And the pipeline points home. In January, Nadella announced Microsoft would invest $3 billion in India, including in artificial intelligence, to advance Prime Minister Modi's vision of a "Developed India." Bill Gates, for his part, has spent a decade as Modi's most reliable Western cheerleader, careful never to say a critical word, because his foundation's access to the country depends on staying on the government's good side.

Follow the whole circuit. American engineers priced out at home. A regulatory side door opened so foreign graduates could be hired cheaper and kept more compliant. A CEO who came up that door and now sends the investment, the AI work, and the praise back to the country the pipeline draws from. The talent flows one way. The capital and the flattery flow the other.

None of it is illegal. All of it is the plan.

The Insult On Top Of The Injury

If the story were really "America has a desperate shortage of engineers, so we must import them," you might grumble and understand.

Microsoft's own actions destroy that story.

In 2025, the same year the software fell apart and the nuclear agency got breached, Microsoft laid off more than 15,000 employees. One round alone was about 9,000. It did this while posting record cloud revenue and record profits and paying its executives lavishly.

You cannot claim a talent shortage so severe it justifies a twenty-year campaign to build a private visa pipeline, and then fire fifteen thousand of the exact workers that pipeline supposedly supplements.

One of those two things is false. It is not the layoffs. Those happened.

The Point

Nobody sat in a room and decided to make your software bad. It is worse than that, and more ordinary.

A company could not win an honest vote in Congress. It found a side door around it. The side door turned out to be cheaper anyway. And the company reorganized itself around a simple discovery: it gets paid whether or not the product works.

Treasury Secretary Bessent called the accounts “one of the president’s most enduring legacies.” Six million accounts opened in the first week. Michael and Susan Dell pledged $6.25 billion. Ray Dalio, Brad Gerstner, BlackRock, and dozens of other billionaires and corporations lined up to add their names to the list.

Everyone is calling this generosity. It isn’t. It’s the most elegant mechanism yet devised for making ordinary Americans permanently, structurally dependent on asset prices at the exact moment those asset prices are being propped up by a trillion-dollar bet on AI that may not pay off.

To see why, you need Karl Polanyi.

The fictitious commodity, updated

In his seminal 1944 work, The Great Transformation, Polanyi argued that market societies work by treating things that were never actually produced for sale: land, labor, money as if they were ordinary commodities. He called these “fictitious commodities.” Land isn’t manufactured; it exists. Labor isn’t a product; it’s a person’s life. Money is a social convention, not a thing with a cost of production. Pretend otherwise, Polanyi warned, and you get a market with no natural limits, one that will, left alone, “demolition of society”:

To allow the market mechanism to be the sole director of the fate of human beings and their natural environment… would result in the demolition of society.

Retirement security used to be delivered through wages, pensions, and Social Security’s pay-as-you-go transfer. Then it was 401Ks, retirement security from a mandatory personal equity position. Now we have Trump accounts. Every dollar in these accounts must be permanently invested in the stock markets.

Permanent froth to pump up asset prices.

Who actually benefits

Start with Michael Dell’s $6.25 billion “donation.” It sounds like philanthropy. It isn’t structured like philanthropy.

• The money doesn’t go to a child or a family. It must be in a broad US equity index fund and in these funds it must remain, untouchable, until the child turns 18.

• Dell gets to deduct the full pledge today, immediately, against his tax bill.

• The forced buying pressure of billions of new dollars hitting the market lifts the value of every share Dell already owns, immediately.

As of July 4th, there was $1.4 billon in government committed money going to these Trump accounts to be injected right into the stock market. More borrowed money from a Country with a $38 trillion national debt.

Bessent described the program as a “backdoor to privatizing Social Security.” Retirement outcomes for tens of millions of people become even more hostage to the market. More incentive to vote and do whatever it takes to keep the market going up.

In February 2026, Attorney General Pam Bondi, facing pointed questions from Congress about why the Justice Department had indicted none of Jeffrey Epstein’s co-conspirators, responded not with an answer but with a market update: “The Dow is over 50,000 right now. The S&P at almost 7,000 and the Nasdaq smashing records.” The Attorney General reached for stock index levels as a substitute for an answer on Epstein. That is where we are going. The market isn’t just an economic indicator anymore. It’s being deployed as a political shield. Trump Accounts are the institutional version of that same move: wire the public’s fortunes into the index, and “the market is up” becomes the only thing that matters.

Pumping the AI bubble.

Trump Account dollars are wiring children into a market that is not a diversified, boring, weigh-the-fundamentals market. It is, increasingly, a single concentrated bet: that AI capital expenditure pays off on schedule.

The numbers are stark. As one analyst put it, the industry is producing eleven digits of annual revenue to justify thirteen digits of spending. Months ago investor Brad Gerstner, a massive Trump account proponent who stood with Trump at today’s bell ringing, asked Sam Altman: “How can a company with $13 billion in revenues make $1.4 trillion of spend commitments?” Altman shot back: “If you want to sell your shares, I’ll find you a buyer… Enough. I think there’s a lot of people who talk with a lot of breathless concern about our compute stuff or whatever that would be thrilled to buy shares.” Perhaps Altman who just decided to push OpenAI’s IPO date to 2027 is a little bit less confident now.

Or maybe Brad Gerstner is pushing for these Trump accounts as the bag holder for the shares of OpenAI stock he is no longer as confident in.

This isn’t a side bet the market is making. It is the market, at this point. The “Magnificent Seven” carry the S&P 500’s returns. Hyperscaler capex funds Nvidia’s revenue, which funds Nvidia’s stock price, which funds the hyperscalers’ ability to keep raising capital to spend more. It’s a closed loop, and OpenAI’s and Anthropic’s planned IPOs are the loop’s next stress test. As the FT put it bluntly: the costs of staying at the frontier are punishing, and both companies must “pray they have their timing right,” because the last time investment cycles ran this hot — the dotcom boom — the Nasdaq fell 77% peak to trough and 4,800 companies went bust.

Every dollar from Trump accounts defaults into S&P 500 tracking funds. Which means it all flows into the same handful of AI-exposed mega-caps that are already carrying the index. The accounts don’t diversify away from the AI bet. They concentrate the country’s children inside of it, at the exact moment even Wall Street’s own bulls are warning the bet is unusually fragile.

This where Polanyi’s insight about self-interest becomes uncomfortable. Once your retirement, your child’s savings, and your portfolio all depend on the AI capex cycle continuing, you have a rational, self-interested reason to want it to continue even if it means tolerating spiking electricity bills from data centers in your town, even if the underlying revenue doesn’t yet justify the spending, even if commoditization from cheaper open models is already compressing the margins the whole thesis depends on. You’re not being irrational by defending the bubble. You’re doing exactly what a person with no other safety net would do. That’s the trap: the system doesn’t need to coerce anyone into supporting it, because it has already made their survival contingent on its success.

al

Index Capture

When SpaceX went public at a $1.8 trillion valuation, Nasdaq, FTSE Russell, and CRSP all rewrote their eligibility rules within months to fast-track it into their benchmarks: inclusion after just 5 to 15 trading days, minimum float requirements waived outright, no profitability test required. Only S&P Dow Jones Indices held the line, keeping its 12-month seasoning period and GAAP profitability requirement intact, which means SpaceX (a company that posted a $4.94 billion net loss in 2025) can’t enter the S&P 500 until at least mid-2027 (if it’s profitable then).

Those broad index funds the Trump accounts invest in aren’t static, they are a moving political battleground where the largest, most connected private companies lobby index providers directly to change the rules of entry in their favor, so existing pre-IPO shareholders get a guaranteed pool of forced buyers. Musk didn’t have to convince a single retail investor. He needed to convince a handful of index committees. Once he did, billions of dollars of everyone else’s retirement savings — soon to include Trump Account money by default — became contractually obligated to buy in, no judgment involved.

The index is supposed to be the passive, disinterested backdrop against which everything else happens. It’s not. It’s an asset the largest players actively lobby to shape, and Trump Accounts wire every American child into whatever comes out the other side of that lobbying.

New Nationalism

In 1910, Teddy Roosevelt delivered his “New Nationalism” speech. In that speech, he argued that the country’s core danger wasn’t wealth itself but wealth concentrated enough to capture the systems meant to check it. He thought unaccountable concentrations of economic power would, left alone, always find a way to dress self-dealing as public benefit. When he spoke of “malefactors of great wealth,” this wasn’t rhetorical flourish, this was a description of exactly this pattern. The use of philanthropy, patriotism, and opportunity-for-children to make entrenchment look like generosity.

At least with that generation of elites, their philanthropy went towards a hospital rather than adding more froth to the stock market.

Back to Polanyi

Polanyi’s history is full of the raw version of this: 19th-century industrialists who simply commodified labor and let wages be set by an unregulated market, full stop, and called the wreckage “progress.” What we’re watching now is a subtler, more durable version of the same move. Instead of just commodifying people’s labor and calling it freedom, the system gives people a personal equity stake in the commodification of their land and labor so that defending the market’s continued expansion becomes something people do for themselves, not something imposed on them from outside.

None of this is an argument against giving children a financial head start, or against markets generally. It’s an argument that we should be moving away from universal, mandatory market exposure as the substitute for social insurance.

Polanyi’s answer to unchecked commodification was never “abolish markets.” It was the “double movement” of society organizing to protect itself once the social costs of full marketization become intolerable. The question Trump Accounts pose is what the response looks like when the thing people need protecting from is the same thing their own retirement depends on.

That’s real disaster. They’ve made it individually rational for millions of Americans, including children to want the AI bubble to keep inflating regardless of the costs to our society.

Read that again, slowly, because there is a sleight of hand in it so smooth most people will never notice.

The American dream, in that headline, means the ease with which talented people from other countries can come here to work. The "10 out of 10 destination" is a score — a rating of how good a deal America is for the person arriving. When the piece mourns that the dream is imperiled, what it means is that America has become a slightly less convenient place to be hired into.

That is one American dream. It belongs to the arrival, and to the company that hires the arrival.

There is another American dream, and it belonged to a man named Kevin Flanagan.

Zero Sum Game

Kevin Flanagan was a software developer at Bank of America's technology center in Concord, California. He had been there seven years. He had studied computer science and philosophy at Cal State Long Beach. By every account he was bright, competent, and hard-working. Flanagan had done the exact things this country tells you to do if you want a good life. Go to school. Learn a skill. Work hard. Be loyal. He did all of it.

It wasn’t enough. For a year before the end, he watched it come apart. His father, Tom Flanagan, later told InformationWeek what his son had known: "He knew for a year that Bank of America was dumping programmers and tech people, and he knew it was coming. And he knew people from India were brought into his building and trained to do his job and sent back to New Delhi."

H-1B workers were brought into Flanagan’s building where Bank of America forced Flanagan and his colleagues to train them. Then once they were trained, Bank of America fired those American workers and ultimately sent the work overseas, where it could be done for less.

Kevin Flanagan was not obsolete. His skills were not out of date. There was no shortage of people who could do his work. He was simply more expensive than the alternative the bank had arranged, and so he was made to hand over everything he knew before he was let go.

Sadly this was too much for Kevin Flanagan who took his own life on that April 17, 2003, when Bank of America let him go. He was 41.

Flanagan’s father said that his son's replacement, the loss of his job to a system designed to underprice him, was "the defining event in his decision to end his life."

An H-1B may have achieved their American Dream but it came at the cost of Kevin Flanagan’s.

Whose American Dream?

When Bloomberg tells you the American dream is imperiled, that’s not an honest question. H-1B has been imperiling the American Dream since it was created. The only honest response is a question.

Whose?

Because there are two dreams, and they sit on opposite sides of the same transaction. There is the dream of the person arriving to take the job, and the dream of the person already here who holds it.

Milton Friedman called the H-1B visa what it is: a government subsidy program. A mechanism by which the public underwrites cheaper labor for private employers. That is the dream Bloomberg is worried about losing. The subsidy. The discount. The 10-out-of-10 convenience of a system tilted toward whoever pays the wage and against whoever earns it.

Kevin Flanagan was chasing his American Dream too. He had followed the rules to achieve success in his home. That dream did not fail him because it was naïve. It failed him because at some point, quietly, without anyone voting on it, this country decided that the worker was a cost to be minimized rather than a citizen to be kept.

The dream failed him because unlike the workers that were imported to replace him, his compensation came solely in wage. And that made him too expensive. The thing that made his replacements willing to accept a cheaper wage was something Kevin Flanagan did not need from Bank of America. It was something owed to Flanagan by birthright.

It was his birthright.

Will they ever consider what they’ve done to those of us like Kevin Flanagan and what has become of our American Dream?

What about our ancestors? What would have happened to them?

It’s usually said with real feeling, and it’s usually said by someone whose family came through Ellis Island a hundred and twenty years ago. It’s also a category error, and it’s worth being precise about why, because the precision is the whole argument.

Our Ancestors

There was no formal sponsorship requirement in the way we’d recognize one today, no signed affidavit, no binding legal document. But there was a real test, and immigrants had to clear it: whether they were likely to become a public charge. Officers wanted to see money in hand, usually around $25, or someone waiting on the other side who could vouch for them and take them in. Unaccompanied women were the one group held to something close to a hard rule; they generally needed a husband, father, or brother to meet them, or they were detained until one arrived.

The Alien Contract Labor Law of 1885 barred immigrants with pre-existing jobs. The intent was to stop companies from importing cheap labor en masse, undercutting American workers by pre-signing foreign workers to jobs before they even got here.

Roughly ninety-eight percent of arrivals got through. They landed, they were inspected, they were let in, and the overwhelming majority stayed and built lives here. That’s not a technicality. That’s the whole ballgame.

Lawfully & Permanently

Lawfully admitted. Intending to settle. Domiciled, in the language the Fourteenth Amendment’s own framers used for exactly this question. A family that cleared Ellis Island in 1905 and stayed wasn’t hiding from anyone. They weren’t here on a tourist visa with a return ticket already booked. They came through the front door, passed the inspection the law required, and put down roots. Many of them naturalized. Their children were raised as Americans in an American household, by people who intended to stay Americans.

Kavanaugh’s Request

Kavanaugh’s concurrence in Trump v. Barbara didn’t propose stripping citizenship from the children of anyone who ever crossed an ocean to get here. His language was specific:

exceptions for children born to foreign citizens unlawfully or temporarily in the country.

Two categories. People here without legal authorization, and people here on a basis that was never meant to be permanent.

Every serious version of the statute built on that opening, whatever else you think of the politics around it, draws the same line: citizens and lawful permanent residents in, unlawful or temporary presence out. Domiciled, settled, intending to stay: protected. That’s not a new distinction invented to justify the modern proposal. It’s the same distinction the framers themselves used when they wrote “subject to the jurisdiction thereof” instead of just “born here,” and it’s the same distinction that ran the immigration station at Ellis Island for sixty years.

The Truth

Ask what actually happened to an unlawfully present family in 1905, or a family here on a status everyone understood to be temporary. The answer is they didn’t get past the interview. Two percent didn’t. The ones who did get through, the overwhelming majority, the people whose photographs sit on your mantle, cleared a public-charge test, had someone vouch for them or money in their pocket, and settled in for good. They are not the population any version of this proposal touches. They are the population every version of it protects.

The invocation of ancestors isn’t wrong because immigration was harder then, or easier, or different in some vague historical sense. It’s wrong because it assumes your family belongs in a category they were never in.

Say it plainly instead of gesturing at the past: your ancestors didn’t come here unlawfully or temporarily. That’s exactly why they’re still here, and exactly why the comparison doesn’t hold.

Those were the people whose children received birthright citizenship then and those are the people whose children should receive it today.

Consistent with the Fourteenth Amendment, Congress could amend §1401(a) or otherwise enact new legislation establishing exceptions to birthright citizenship for children born to foreign citizens unlawfully or temporarily in the country.

This is all Congressional Republicans should be focused on. Anything else is an admission of guilt. Let’s force every member of Congress to walk the line.

The mechanism

Congress inserted the “subject to the jurisdiction thereof” language from the Fourteenth Amendment into federal statute in 1940, and carried it forward into the Immigration and Nationality Act of 1952. This was after the Wong Kim Ark decision that interpreted “subject to the jurisdiction thereof” as meaning birthright citizenship. Kavanaugh’s ruling was that these Congressional statutes would supersede Trump’s Executive Order and therefore Trump’s EO was invalid.

He went further and this is where it really gets interesting. Kavanaugh wrote that Congress could, consistent with the Fourteenth Amendment, amend those statutes or enact new legislation establishing exceptions to birthright citizenship. The exceptions seem very reasonable. Children born to foreign citizens who are either unlawfully present in the US or temporarily in the country do automatically become permanent American citizens.

Reasonable right?

Why he thinks that’s allowed

Wong Kim Ark recognized four narrow exceptions to birthright citizenship in 1898: children of diplomats, of invading armies, born on foreign public ships, and of tribal members owing allegiance elsewhere. Kavanaugh’s point is that those exceptions weren’t meant to be frozen at four forever. They were narrow in 1868 because the categories of people whose presence wasn’t meant to be permanent were narrow then. There was no mass illegal immigration. There was no visa system. There was no such thing as lawful but temporary status at scale. There was no intercontinental air travel.

Kavanaugh’s argument is that the same underlying principle, applied honestly to conditions the framers never had to classify, supports new exceptions for exactly those two categories now. It’s not a new theory of the Constitution, it’s the same theory, extended to circumstances that didn’t exist yet.

What it isn’t

It isn’t an amendment. Kavanaugh is explicit that a statute is the path, not a two-thirds-and-three-fourths process that hasn’t succeeded on anything contested in over fifty years. That’s a dead end.

It isn’t a reason to give up. Four justices, Kavanaugh plus the three dissenters, are on record against the majority’s constitutional holding. That’s not a majority. It’s also not nothing, and it’s not the “settled, 6-3, don’t bother” story running in most of the coverage.

Already in motion

Within hours of the ruling, Sen. Eric Schmitt announced he was filing legislation to walk through the door Kavanaugh may have left open, posting that “Justice Kavanaugh MAY have left Congress a door. I’m filing legislation to walk through it.” He’s running two tracks at once: a statute aimed at the Kavanaugh opening, and, separately, a constitutional amendment as a backup if the statute isn’t enough on its own.

Lesser Evils.

We find ourselves in a lesser evils trap between the Republicans and Democrats. The Republicans claim to be on our side but always seem to fall short. The dog that can never catch its tail. But were they ever really on our side? Or was it all nice sounding words to secure our vote to get elected where their actions could focus on benefiting their donors? It’s time for an actual test.

A clean bill, matched to Kavanaugh’s language, is just that. Every member has to explain their vote in public, not in a tweet. The ones who vote against have to explain why children born to people unlawfully or temporarily in the country should be citizens by default, on the record, to their own voters.

Let’s see which Republicans are for the people of their district and not their donors.

Let’s make Democrats explain why the children of people in the US on a temporary basis should become citizens forever.

Trump vowed to end birthright citizenship. Let’s see who is really on board with him.

“We tried” isn’t good enough anymore.

I cannot agree with, indeed I am appalled by, the Court’s suggestion that the decision whether to stand by an erroneous constitutional decision must be strongly influenced, against overruling, no less, by the substantial and continuing public opposition the decision has generated. -Antonin Scalia (calling out stare decisis in Trump vs. Barbara)

The Supreme Court ruled 6-3 today in Trump v. Barbara. Executive Order 14160, the one denying birthright citizenship to children of parents unlawfully or temporarily in the country, was struck down.

The New York Times called it “reaffirming the long-held principle that the Constitution guarantees that nearly all children born on U.S. soil are citizens.”

Read the actual opinions and a different picture emerges. The case for the broad reading is thinner than the vote count suggests. The case for the narrow reading is stronger than the coverage is giving it credit for.

Roberts’ argument does not rest on the Constitution itself but rather on stare decisis: a fear of the consequences of eliminating the birthright citizenship policy.

It rests on stare decisis.

While some Republicans are now speaking of the need for an amendment, Trump is not. He posted on Truth Social.

His instinct that this isn’t over is right, even if the tweet alone doesn’t prove it. Here’s why.

What the Court actually did

Five justices, Roberts, Sotomayor, Kagan, Barrett, Jackson, held that the 14th Amendment’s “subject to the jurisdiction thereof” language guarantees the broad birthright citizenship rule America has had since the Supreme Court ruled on Wong Kim Ark in 1898.

Kavanaugh refused to join that holding. He said so directly: “I respectfully disagree with the Court’s constitutional holding. In my view, the Executive Order does not violate the Fourteenth Amendment.”

Kavanaugh sidestepped the interpretation of the 14th Amendment entirely. He cited a 1940s statute that reused the Fourteenth Amendment’s “subject to the jurisdiction thereof” language after Wong Kim Ark had already defined it, and argued Congress quietly adopted the Wong Kim Ark interpretation by reference. That statute, in his view, superseded Trump’s EO on its own, no need to reinterpret the 14th at all.

Kavanaugh didn’t stop there. He explained that the idea that broad birthright citizenship is constitutionally guaranteed doesn’t stand on as firm a ground as one might think: “As revealed by the Court’s opinion with its detailed account of history and precedent, and by the weighty and thoughtful dissents, the constitutional issue is far more complicated than the statutory issue. After reading those scholarly opinions, one thing seems evident: the constitutional issue is not straightforward, much as we might want it to be.”

To paraphrase Kavanaugh: “subject to the jurisdiction thereof” is not a guarantee of birthright citizenship for all.

Sorry, New York Times.

The opening Kavanaugh left

Kavanaugh didn’t stop at disagreement, he laid out the next steps. He said Congress could change the statute referencing “subject to the jurisdiction thereof” and noted that Congress hasn’t done so yet. Write a new statute, one that doesn’t lean on old language tied to old precedent, and the technicality that let him sidestep the constitutional question disappears. At that point his own stated view governs, and he already told you what it is: “that interpretive principle would support additional exceptions for children born to foreign citizens unlawfully or temporarily in the country.”

Add Thomas, Alito, and Gorsuch, who said so in dissent, and a statute to eliminate birthrigjt citizenship as we know it enters Supreme Courr debate with the support of four justices. That’s the real number coming out of this case. Not three dissenters and a lost cause. Four, with a roadmap from the fourth telling you exactly how to get there.

Before you attack my math skills, I know four does not a majority make. We’re going to need someone to flip and my bet would be on Roberts, the author of the majority opinion.

Why a Congressional bill is a different problem for Roberts

Roberts has spent two decades on record worrying about exactly this kind of moment. At his confirmation hearing he said judges “are like umpires” who “don’t make the rules, they apply them.” In 2007 he warned that repeated 5-4 splits along predictable lines risk “a steady wasting away of the notion of the rule of law.” He said he’d push colleagues toward narrower, more unified opinions specifically to protect the Court’s standing as legal rather than political.

That’s what makes today notable. The narrow, low-heat option was sitting right there, and Roberts didn’t take it. Kavanaugh wrote it for him, basically. Resolve the case on the 1940s statute, avoid the constitutional question, let the next case sort it out. That’s the classic Roberts move, the one he’s built his whole reputation on. He passed on it. He went further than he had to and wrote a full constitutional holding instead, engaging Thomas, Alito, and Gorsuch point by point rather than ducking the fight.

This isn’t the first time. In Citizens United, Roberts had a narrow off-ramp available too. The Court scheduled reargument specifically to consider whether to overturn a prior precedent, McConnell v. FEC, and he didn’t have to go all the way there. He went anyway, and used the case to reshape campaign finance law for a generation.

But what if Congress was able to follow Kavanaugh’s instructions and return with a bill that would actually exclude children born to people in the US unlawfully or on a temporary basis?

The executive order was Trump’s decision. Striking it down put the Court against a single administration’s unilateral action. A statute that passed both chambers of Congress is a different animal entirely. Striking that down means telling the elected legislature, after debate and a recorded vote, that it lacks the power to shape a major part of government policy. If you’re going to do that, you better have some strong Constitutional backing.

A thorough review of Roberts’ argument suggests he actually doesn’t have that constitutional support. Instead, Roberts’ argument ultimately relies on stare decisis. The fact that for over 100 years, the official policy of our government has been that we have a broad, constitutionally guaranteed birthright citizenship. Roberts’ biggest issue seems to be not the interpretation of the Constitution but rather what the implications of a change would be.

Chief’s Pillars

Pillar one: the textual jurisdiction argument.

Roberts’ argument: “subject to the jurisdiction” reaches anyone physically present, grounded in Marshall’s 1812 opinion in Schooner Exchange v. McFaddon, a case about whether a foreign warship could be seized in a U.S. port. Not about citizenship, but Roberts uses it to establish what “jurisdiction” meant as a legal term in the era: full territorial power over anyone present, with only narrow exceptions.

The counter: Justice Alito makes the first move here. He argues Schooner Exchange has no bearing on the Fourteenth Amendment at all, it’s a case about whether the United States implicitly waived its own sovereignty over a foreign warship in port, a sovereign-immunity question, not a citizenship question. Neither the holding nor Marshall’s reasoning, in Alito’s words, has anything to do with the Citizenship Clause.

Justice Thomas presses a sharper problem underneath that one. The Fourteenth Amendment supplies its own contradiction. The Equal Protection Clause, drafted by the same Congress in the same section, protects anyone “within” its jurisdiction. Plainly territorial. The Citizenship Clause uses different words: “subject to” the jurisdiction thereof. If the framers meant the same territorial concept in both places, why phrase it differently? Roberts’ opinion never identifies a single person in the 1866 debates who used “subject to the jurisdiction,” as opposed to the differently worded “within the jurisdiction,” to mean bare territorial presence. The Amendment’s own drafters used two different phrases in two adjoining clauses. That’s presumed to be intentional, not interchangeable.

Bottom line: Roberts says the phrase means physical presence, full stop. Alito and Thomas say he’s leaning on the wrong case and ignoring the Amendment’s own word choice to get there.

Pillar two: “declaratory of the law of the land.”

Roberts’ argument: Senator Howard and Representative Bingham both described the Citizenship Clause as merely “declaratory” of existing law. Roberts treats that as proof the framers meant unconditional common-law presence, nothing new, just writing down the rule that already existed.

The counter: neither man said which law was being declared. And if nobody in the debates actually used “subject to the jurisdiction” to mean territorial presence, “declaratory” has nothing left to point to on Roberts’ side.

Bottom line: Roberts says the framers were just writing down settled law. The counter says he never proves which settled law they meant.

Pillar three: the domicile-isn’t-birthright recharacterization.

Roberts’ argument: the Reconstruction-era domicile language belongs to naturalization and state-citizenship law, not to who’s a citizen at birth. Domicile mattered for becoming a citizen or losing citizenship, not for the separate question of a child’s status at the moment of birth.

The counter: This one has two parts, and they work together. Domicile and allegiance aren’t the same concept, but in the framers’ vocabulary domicile is what allegiance looks like when you can actually see it. You can’t inspect someone’s loyalty directly. You can see where they’ve settled and built a life. That’s why the two show up together throughout this record rather than as competing theories.

Allegiance: The Citizenship Clause was written to carry forward the Civil Rights Act of 1866, passed by the same Congress months earlier, which made citizens of “all persons born in the United States and not subject to any foreign power.” Senator Trumbull, that Act’s author, explained what it meant on the floor: “subject to the complete jurisdiction,” meaning “not owing allegiance to anybody else.” Senator Howard, the Fourteenth Amendment’s own drafter, said he “concur[red] entirely.” And Bingham’s 1866 floor statement lands in the same place: “every human being born within the jurisdiction of the United States, of parents not owing allegiance to any foreign sovereignty, is… a natural born citizen.” Not naturalization language. The clearest birthright-citizenship term available in the era’s vocabulary.

Domicile: The evidentiary proxy for that allegiance: Bingham had already tipped his hand years earlier. In 1859, arguing against Oregon’s proposed constitution, he described citizenship this way: “all free persons born and domiciled within the United States are citizens of the United States from birth.” Domicile wasn’t a stray word he reached for once in 1866. It was already how he thought about the question, seven years before he helped write the Amendment. Same man, same underlying concept, described twice, once through domicile and once through allegiance, because in his mind they were pointing at the same thing.

Bottom line: Roberts treats domicile as irrelevant to birthright citizenship. The counter treats domicile and allegiance as two names for the same requirement, both traceable to the same drafters, both showing up in the birthright context specifically, not the naturalization context Roberts tries to file them under.

Pillar four: Wong Kim Ark’s actual holding.

This is the deepest crack in the majority opinion, and it’s also the one place Roberts saw the attack coming and answered it directly instead of leaving it exposed.

The setup: the 1898 decision states its own rule in Gray’s own syllabus. A child born here becomes a citizen where the parents “have a permanent domicil and residence in the United States.” Not a later gloss added by treatise writers. Gray’s own formulation. In the years right after the decision, nearly every major legal authority read it that way. Treatises in 1898, 1901, 1904, 1905. Bouvier’s Law Dictionary. The Supreme Court itself, twice, in Chin Bak Kan v. United States (1902) and Kwock Jan Fat v. White (1920), both describing Wong Kim Ark as covering children of parents “permanently domiciled in the United States.”

Roberts’ argument: Attempts to narrow Wong Kim Ark based on its domicile facts “fail because the holding’s underlying reasoning cannot be squared with a domicile requirement.” His position is that Gray’s actual analysis, the common-law jus soli reasoning, never depended on domicile as a limiting element. The domicile language just described Wong’s specific facts. He also points out “domicile” shows up only twice in the Civil Rights Act debates and once in the Citizenship Clause debates specifically. Too thin, in his telling, to be an intended constitutional requirement, especially for a Congress that wanted the citizenship question settled “once and forever.”

Worth noting: This isn’t a new argument Roberts came up with. It’s the same one ACLU counsel Cecillia Wang made at oral argument. The Republican Supreme Court justice is sharing arguments with the ACLU.

The counter: Roberts’ (or should I say the ACLU’s) answer is real, not a dodge, but it’s not obviously the better reading either. It has to explain away Gray’s own choice of words in stating his own holding. It has to outweigh how nearly every legal authority who touched this case for two decades afterward read it, including the Supreme Court’s own later characterizations. This is the one pillar where “Roberts is wrong” isn’t demonstrated. It’s a genuine, live dispute between two competing readings of the same 1898 case. Arguably the single most contested point in the whole opinion.

Kavanaugh adds a different piece here, and it’s worth keeping separate from Thomas’s argument rather than blending it in. Wong Kim Ark recognized four narrow exceptions to birthright citizenship: children of diplomats, of invading armies, of those born on foreign public ships, and of tribal members owing allegiance elsewhere. In Kavanaugh’s own words, Wong Kim Ark “treated those four exceptions as a closed set for constitutional purposes.” His argument is that treating them as permanently frozen as of 1868 is inconsistent with how the Court normally reads the Constitution. Rules stay constant, he says, but their application to new situations doesn’t. In 1868 the exceptions were narrow because the categories that would justify them barely existed. There was no mass illegal immigration to speak of. There was no visa system, no such thing as “temporary lawful status.” The handful of people whose presence was understood to be genuinely temporary were the diplomats and the tribal members, so that’s what the exceptions covered. Kavanaugh’s claim is that the same underlying logic, applied honestly to a world where millions of people now fall into categories the framers never had to name, supports adding new exceptions for exactly the groups the executive order targeted. Not because domicile already excluded them in 1868. Because the class of people whose presence isn’t meant to be permanent has grown, and the exceptions should grow with it.

Bottom line: Roberts says domicile was incidental to Wong Kim Ark’s facts, not part of the rule. The record from the two decades after the case say otherwise. This one’s a real toss-up. And Kavanaugh’s separate point sits on top of it either way: even if the four exceptions were a closed set in 1868, he argues the Court doesn’t have to treat them as permanently frozen, since the same underlying rule can be extended to circumstances the framers never had to classify.

Pillar five: the “you can’t stomach the consequences” reductio.

Roberts’ argument: If divided foreign allegiance defeats citizenship, that principle would also exclude the children of lawful permanent residents, a result the dissenters don’t accept and don’t explain. Treated as proof the narrow theories are incoherent, this is Roberts’ rhetorical centerpiece.

The counter: Look at what it’s actually objecting to. Not that the principle is wrong. That an opinion doesn’t state where the line is or why it stops where it stops. That’s a drafting gap, not proof no line exists. Congress can close that gap directly: define citizens and lawful permanent residents as within the protected class, children of parents unlawfully or temporarily present as outside it with findings and a legislative record behind it.

There’s a deeper problem with the reductio too. The Court’s job is to say what the Constitution means. Managing the fallout of that meaning, deciding where exactly the line falls between a green-card family and an undocumented one, is the kind of line-drawing the Constitution leaves to the elected branches: Congress and the President. Roberts’ whole argument treats an opinion’s failure to fully solve that policy problem as proof its constitutional reading must be wrong. But that’s backwards. If the original meaning is narrower than universal presence, the Court’s job ends at saying so. Who gets excluded and exactly where the line sits from there is Congress’s problem to solve, not a defect in the interpretation for the Court to fix by expanding it.

Bottom line: Roberts says the narrow theories collapse under their own logic. The counter says that’s a gap in how the dissents are written, not in the underlying idea, and a statute can close it. It’s also not the Court’s job to solve in the first place. That’s what Congress is for.

When All Else Fails

Strip away what actually falls and one thing remains. The survivor is stare decisis, the legal doctrine that courts stick with a settled precedent rather than reopening it, even if they might rule differently on a blank slate. A hundred and twenty-eight years of reliance. Millions of people whose legal identity as citizens rests on an unbroken line of practice. Not a textual argument. Not a historical argument. An institutional one. It doesn’t need Wong Kim Ark’s holding to be free of dispute to still be a reason to leave it alone.

Roberts’ doesn’t say it directly but that’s what his argument truly rests on. It’s why the ACLU is running ads with Springsteen playing in the background. Because if you can’t win in the Supreme Court, the next step, the logical choice is to take it to the court of public opinion.

https://youtu.be/aNe4E-uynfg?si=P2jrePl90A0OeaVj

The Best Defense

The mainstream framing today is exactly what you’d expect. Settled. Decisive. Over. Roberts himself supplied the line reporters are running with: “Citizenship, then and now, was the right to have rights… We keep that promise today.” Cecillia Wang called it a fight where clients “have been under this cloud where the president of the United States was trying to undo this foundational right that all Americans have relied on for 150 years.” The ACLU’s own immigrants’ rights deputy director, Cody Wofsy, went further: the organization doesn’t “anticipate that there will be a round two of this fight over birthright citizenship, the Supreme Court has rejected it and rejected it emphatically.” NBC, CBS, CNBC, all of it framed as a “major victory,” a “landmark,” a reaffirmation of a “long-standing constitutional principle.”

None of that is fabricated. It’s a real 6-3 outcome and the executive order is dead. But “emphatic” and “settled” describe the vote count. Does Kavanaugh’s statement seem settled to you? It reads more like wishful thinking by liberals, mainstream media, and the ACLU, trying to convince you that this fight is over. The coverage isn’t wrong that Trump lost this battle. It’s incomplete about the state of this war. Four justices on record against the constitutional holding. A majority opinion whose strongest pillar is stare decisis: we’ve had this rule in place for a long time so we should keep it in place, regardless of whether it rests on a flawed interpretation of the Constitution. A Supreme Court decision not based on an interpretation of the Constitution, but rather what the consequences of that interpretation would be. That’s not the Supreme Court’s job. It’s Congress’s. And that’s exactly who can solve this issue. No amendment required.

The fix

Pass a statute that closes the one gap Roberts’ opinion actually got right about the dissents: no stated line between citizens and lawful permanent residents on one side, and unlawful or temporary presence on the other. Write it into the bill itself. Citizens and lawful permanent residents in. Everyone else, the categories the executive order targeted, out. That single addition, an explicit, debated, voted-on definition, takes away the one objection in Roberts’ opinion that wasn’t built on sand.

Remove Kavanaugh’s statutory excuse and the real question lands on exactly two desks: Roberts and Barrett. It puts Roberts in the position his own record says he’s least comfortable in. Not striking down one administration’s order. Striking down an act of the elected Congress.

None of this guarantees an end to birthright citizenship but it guarantees the fight actually happens. Not an amendment that will never go anywhere. Not another executive order that hands the Court an easy target. A statute, built to answer the one real gap in the narrow reading, aimed at the two justices who are actually in play.

At the exact same moment, in counties across the country, those same people are fighting to keep data centers off their land.

These two facts are not unrelated. They are the same story.

PJM

PJM Interconnection manages the largest electricity grid in the United States, serving 67 million people across 13 states and Washington, D.C. This week, amid the heat dome, PJM is projecting electricity demand that could surpass its all-time record, set in 2006.

The reason the record is in danger isn’t just the heat. It’s what’s already on the grid before the heat arrives.

PJM’s own independent market monitor concluded that data centers were responsible for 63% of the increase in electricity capacity prices in 2025/2026. That’s $9.3 billion in costs passed directly to customers in higher electric rates in a single year. Capacity prices have more than 10x’d from roughly $29 per megawatt-day in 2024 to $329 per megawatt-day in 2026/27. Starting this month, ratepayers across the PJM region are paying an additional $1.4 billion annually in capacity market costs, driven largely by data center demand.

PJM has also acknowledged that its grid is essentially full. Starting this summer, it has just enough power to maintain reliability. Data centers are connecting faster than new supply can be built. PJM currently has no authority to stop those connections — data centers are allowed to plug in even when there isn’t enough electricity to go around. By June 2027, the region may fall below reliability standards. That means a greater risk of rolling blackouts, especially during heat waves exactly like this one.

This isn’t a projection. It is the grid operator for the Eastern United States saying, in official filings, that data center demand is the primary reason for tight supply conditions and record consumer prices.

The Floor is Gone.

The energy markets confirm it in real time.

ICE daily PDP futures contracts on PJM West Hub, the benchmark pricing point for the mid-Atlantic grid. Pre 2026, those settled above $487 per MWh three times. Two were December 23rd and 24th, 2022 when winter storm Elliott hit without warning and caused grid emergencies across the Eastern US. One was a summer 2022 heat event. It was always on the last trading day when traders came face to face with weather’s reality. Nobody saw them coming days out.

In 2026 we’ve exceeded $487 twenty times already and not just on the final trading day. Today, July 3rd is already trading above $487. The market is pricing extreme stress in advance because everyone can already see the math. The grid has so little headroom that a heat wave produces the same price signals as a historic winter emergency, and does it days early because the constraint is structural, not seasonal.

In 2022, prices at this level meant a catastrophic surprise. In 2026, they mean a Tuesday in late June when it’s going to be hot.

The difference isn’t the weather. It’s the floor, the permanent, 24/7, never-stops, doesn’t-care-what-season-it-is data center load that has consumed every gigawatt of headroom the grid used to have for exactly these moments.

When your power bill spikes this weekend, it’s not the heat dome doing it. The heat dome is just the trigger. The data centers loaded the gun.

Who is it all for?

The people who built these AI companies spent years telling us — from conference stages, in newspaper profiles, in congressional testimony — that AI was going to take our jobs. They said it openly. They were proud of it. Dario Amodei, CEO of Anthropic, recently described a scenario of 5-10% GDP growth alongside 10% unemployment as “not impossible.” The CEO of AWS announced tens of thousands of layoffs while simultaneously launching an AI recruiting tool that interviews candidates with no human involved.

They told us the jobs were going.

Now those same companies are bidding against us for power. Power that they consume at gigawatt scale. More power than we consume in massive cities. It’s the reason why they are here, they need the power. Power from a grid the American public built over the course of a century through utility bills, municipal bonds, and rural electrification programs to feed something that will take jobs from the American public.

This 4th of July weekend, ordinary families are competing against data centers for the electricity that will keep their elderly parents cool in 105-degree heat. That will keep their pregnant wives comfortable. That will keep their children safe.

For what? So the AI that is coming for your job can keep running.

That is the deal on the table. You subsidize the infrastructure. You absorb the price increases. You lose the job. They keep the AI.

The dominant response from the industry has been that this is a PR problem: people just don’t understand the benefits, AI companies need community tours and listening sessions. But the PJM numbers aren’t a perception issue. $9.3 billion in extra costs passed to ratepayers in a single year isn’t a messaging failure. It’s a transfer of wealth, from the American people who built the grid to those who want to consume it to train their models that will take jobs from those same American people.

What’s the fix?

The fix isn’t complicated. It isn’t a ban on AI or a rejection of technology. It’s the principle every other capital-intensive industry has to live by: pay the full cost of what you consume. Build your own generation. Don’t jack up the power bills of the people who built the grid so your servers can run cool while their houses bake.

Theodore Roosevelt — who broke up the railroad trusts and the oil monopolies — said it plainly over a century ago: “The rights of the public to the natural resources outweigh private rights, and must be given its first consideration.” And he was equally clear about who the obstacle was: “Our duty to the whole, including the unborn generations, bids us restrain an unprincipled present-day minority from wasting the heritage of these unborn generations.”

He wrote that in 1912. It fits 2026 perfectly.

PJM projects that peak demand will grow by 32 gigawatts from 2024 to 2030. All but 2 gigawatts of that growth comes from data centers.

This week, in 105-degree heat, with the grid at record strain, those numbers stop being abstract.

When this bill come’s due, it will like always be the American people who are on the hook to pay it.

Speaking at Rice University in Houston, Kennedy told the assembled crowd:

Those who came before us made certain that this country rode the first waves of the industrial revolutions, the first waves of modern invention, and the first wave of nuclear power, and this generation does not intend to founder in the backwash of the coming age of space. We mean to be a part of it — we mean to lead it.

American innovation had built the modern world that Kennedy, the youngest person ever elected President, was leading America into. America was not alone of course, the Soviet Union, America’s protagonist, had been winning the Space Race but Kennedy was determined that America would win the final challenge of this race: putting a man on the moon.

Kennedy rallied Americans to believe in themselves even if the challenge might seem hard, because it was in that challenge where we would discover and strengthen ourselves. He famously said:

We choose to go to the moon. We choose to go to the moon in this decade and do the other things, not because they are easy, but because they are hard, because that goal will serve to organize and measure the best of our energies and skills, because that challenge is one that we are willing to accept, one we are unwilling to postpone, and one which we intend to win, and the others, too.

Less than seven years later, America landed men on the moon. Americans (with some help from a small team of former German scientists) had achieved the objective that. One small step for man, one giant leap for mankind. What couldn’t America achieve?

Kennedy, the President who had reminded Americans of their capabilities and whose by his actions displayed his belief in the American people’s innovative capabilities declared on that day in 1962

This country of the United States was not built by those who waited and rested and wished to look behind them. This country was conquered by those who moved forward.

President Kennedy whose optimism and belief in his fellow Americans led America to the moon did not live to see the landing. He had been assassinated barely one year after his Houston speech.

That is what a confident nation does when it is challenged. It invests in its own people. It pays them, trains them, trusts them, and points them at the frontier.

Maybe we lost something more on that day in November 1963, because only 15 years later, the office once filled by Kennedy was working on other plans.

The City on a Hill

In 1984, the White House Office of Science and Technology Policy under President Ronald Reagan sent a request to the National Science Foundation.

Reagan loved to quote John Winthrop, the first Governor of Massachusetts, and the man who originated the “City on a Hill” quote that Reagan used so frequently. But he never gave the full quote. Indeed the full quote from Winthrop seems almost antithetical to the ethos of the decade that Reagan would oversee. Yes, Winthrop did say we must be a city on a hill, but how did Winthrop say we should achieve that goal?

We must be knit together, in this work, as one man. We must entertain each other in brotherly affection. We must be willing to abridge ourselves of our superfluities, for the supply of others’ necessities. We must uphold a familiar commerce together in all meekness, gentleness, patience and liberality. We must delight in each other; make others’ conditions our own; rejoice together, mourn together, labor and suffer together, always having before our eyes our commission and community in the work, as members of the same body.

It was the heirs to that city on a hill, that would, 145 years later, take up arms against the British at Lexington and Concord, the event that led to the creation of the Government that Kennedy and later Reagan would represent. There is a direct through line from Winthrop to the document that would form the basis of that Government and it is in the Constitution’s Preamble where we find the reason for the existence of that Government: “Promote the General Welfare” and “Secure the the blessings of liberty to ourselves and our posterity.” In short, the purpose of the US Government was to optimize for the people of America.

In 1984, the Country was afraid. It was a moment not different from the one Kennedy found himself. Instead of the Soviet Union, the protagonist was Japan. Japan had taken the lead in semiconductors. The United States was losing and something had to be done.

Reagan named someone from industry rather than academia to run the National Science Foundation. Erich Bloch, a corporate vice president from IBM had run technical personnel development and chaired a semiconductor research consortium. Bloch did not think of science the way his predecessors had, it was “no more international than commerce”: a competitive weapon.

The same year Bloch arrived, the White House Office of Science and Technology Policy sent the National Science Foundation a request. Analyze the supply and demand for scientists and engineers. Tell us what is coming.

It was a reasonable question to ask during a competitiveness panic. On its face, there was nothing wrong with it. The Government wanted to know whether America would have enough scientists and engineers for the fight ahead.

There were two ways to answer that question and they all depended on one thing: the relationship between the Country and the market?

Two Doors

NSF assigned the analysis to an economist named Myles Boylan. Boylan did his job well. He modeled both sides of the question, supply and demand, the way a competent economist would.

What he found was that the supply of American scientists was a market like any other. The rate at which American graduates went on to a science or engineering PhD depended, in the document’s own words, on “both real and relative salaries.” He ran the numbers. He calculated that the supply of American PhDs was responsive to wages, while the supply including foreign students was three times as responsive. In plain terms, he had measured the thing precisely: if you let the wage rise, Americans come; if you hold it down, they go elsewhere.

He even forecast the wage the market would demand if left alone. Real PhD-level pay would have to climb from roughly $64,000 toward something approaching $100,000 by just after the year 2000 to draw enough Americans into the field.

These were the options:

Door One: Pay American scientists what the market said they were worth, and the shortage solves itself.

Door Two: Hold the wage down by flooding the supply with foreign students who would accept less.

The study’s conclusion was:

It was much more cost effective to the nation than the free market solution of allowing real and relative salaries of NS&E PhDs to rise.

The document was explicit about why foreign students were the tool. Not because they were more brilliant. Because they did not respond to the price signal. It noted that a growing influx of foreign PhDs “will hold down the level of PhD salaries,” and that foreign students were useful precisely because they were “not responsive to small decreases in real starting salaries.”

That is the arbitrage, written into a government document years before anyone coined the phrase “H-1B abuse.” The foreign worker is not chasing the wage. He is chasing the life. He will accept a wage an American would refuse, because the wage is not what he is being paid in. The rest is made up in residency.

The Constitution said the government existed to secure the blessings of liberty for Americans and their posterity. In 1985, an analyst inside that government weighed paying Americans against importing people who would underbid them, and put on paper that the second was more cost effective to the nation.

Someone had quietly redefined what the purpose of our Government was. It was no longer for the benefit of the American people. It was now for the benefit of the American corporation.

The Tell

This is where the people who claim to worship the free market give themselves away.

Thomas Sowell once said: “To an economist, it is incredible to speak about shortages without talking about prices, in this case wages.” A shortage that never resolves is not a shortage. It is a price somebody refuses to let rise.

The document looked at that outcome and chose door two. It recommended overriding the market, not to protect Americans, but to hold their wages down. Flood the supply with people who would not respond to the price. Jam the signal.

Reagan’s free market approach encountered two options: one option that was literally called the free market solution and another. They chose the other. The free market was selective: free for capital, managed for labor.

Milton Friedman saw it for what it was. He called the H-1B program “a government subsidy.” That is what a subsidy is: a transfer from the public to a private interest, laundered through policy. The public was the American technical workforce. The private interest was the employer. And the mechanism was a manufactured shortage and forty years of suppressed wages.

What Happened

The wage model was finished in 1985. The study built on it was published in 1987 with the demand side stripped out, projecting a shortage of hundreds of thousands of scientists and engineers that never arrived. The number went to the Senate. It justified budgets. It justified the H-1B visa program. And it is still running.

We know how it ends because we are living in the ending. The signal that would have told American kids to become scientists was jammed, exactly as the document described. The wage never rose. The Americans who would have filled the pipeline went into finance and law instead, because the reward for building things had been deliberately held down. And the visa built on the manufactured shortage was handed to firms like Tata and Cognizant, who perfected the model the document only sketched: bring the worker in, take what he knows, send the work to wherever labor is cheapest.

Here we are again.

The men selling artificial intelligence today are making the same argument the NSF made in 1985, stripped of even the pretense. In 1985 they at least claimed a shortage. They claimed talent. The AI pitch does not bother. It does not tell you how the technology serves the American worker. It tells you the worker must serve the technology. Do not regulate it. Do not slow it down. Do not ask what it is for. Get out of its way.

Polanyi’s quote from his seminal 1944 work “The Great Transformation” seems apt.

To allow the market mechanism to be the sole director of the fate of human beings and their natural environment… would result in the demolition of society.

Notice what is missing. No one has made the case for how this helps us. They have only made the case that we should help it. When the people selling you something can no longer explain what you get, and are reduced to demanding you not stand in its way, they have already told you which door they are standing in.

Our City on a Hill will turn 250 years old later this week. Perhaps it is a good time for us to give this question some thought.

Who is our city on a hill for?

September 2008: mortgages in Florida, Arizona, Nevada, California are failing in waves and the contagion is spreading. Two companies: Fannie Mae and Freddie Mac stand behind roughly $5 trillion of it, companies created to make the American Dream of homeownership a reality for Americans.  To enable this, these companies had been given an implicit guarantee from the U.S. Government.  An implicit guarantee backed by the American taxpayer.  If they collapse, mortgage credit in America disappears: no conforming loans, no refinancing, no floor under home prices.

The federal government seized both and placed them in conservatorship.  The United States Treasury, or put differently the American taxpayer, ultimately injected over $190 billion to keep them standing.

Americans of the present took on this debt that would be carried by Americans of the future and they did that for Americans.

Today, eighteen years after the Global Financial Crisis, Fannie Mae and Freddie Mac are still in government conservatorship.  The companies come under the supervision of the Federal Housing Finance Agency (FHFA) which is led by Bill Pulte.  The number one goal in FHFA’s 2026-2030 strategic plan is to “responsibly oversee Fannie Mae and Freddie Mac for the American People.” 

But behind the scenes, the actions of these companies suggest a different motivation: a motivation to boost the profits of American home builders while increasing American housing costs and adding billions of additional risk onto the balance sheet of the American people.

Fannie Mae’s Selling Guide directs the enterprise to purchase mortgages made to non-permanent residents such as H-1B holders on the same terms available to U.S. citizens.  If the visa lapses two years into a thirty-year note, the taxpayer holds the next twenty-eight.

We bailed out an institution whose purpose was American homeownership. It is now being used for something else entirely.

HUD Makes Its Move

In March 2025, the Department of Housing and Urban Development (HUD) issued Mortgagee Letter 2025-09. As of May 25, 2025, non-permanent residents were made ineligible for Federal Housing Administration (FHA) insured mortgages. This included H-1B, L-1, and F-1 visa holders. Only U.S. citizens and lawful permanent residents qualify.

HUD’s General Deputy Assistant Secretary for Housing Jeffrey D. Little wrote:

The Administration has reaffirmed its commitment to safeguarding economic opportunities for U.S. citizens and lawful Permanent Residents while ensuring that federal benefits, including access to FHA-insured Mortgages, are reserved for individuals who hold lawful Permanent Resident status.

Further, the Assistant Secretary wrote:

Currently, non-permanent residents are subject to immigration laws that can affect their ability to remain legally in the country. This uncertainty poses a challenge for FHA as the ability to fulfill long-term financial obligations depends on stable residency and employment.

In short: HUD’s goal is to provide economic opportunities for Americans, and non-permanent residents create a risk because their ability to remain in the country is not guaranteed.

In response to HUD’s policy change, and fearing for their business, the head of the National Association of Mortgage Brokers sent a letter to Director Pulte. The goal: ensure that Fannie Mae and Freddie Mac do not enact a similar policy to HUD’s.

The Letter

The letter from James L. Nabors, II, President of the National Association of Mortgage Brokers, is notable because the risks Nabors identifies in an attempt to dissuade Pulte from making a similar change are in fact the exact reasons why this change IS required.

The risks identified by Nabors were:

1. Market Contraction

2. Risk Concentration

3. Community Impact

I’ll dismiss the community impact argument now. That is not the purpose of Fannie Mae and Freddie Mac. We’re not subsidizing mortgages so that American communities can be made vibrant by the inclusion of foreigners.

Let’s focus on the other arguments.

Market Contraction

Mortgage Brokers: “Reducing the pool of eligible homebuyers may lead to decreased demand in the housing market, potentially affecting home values and slowing economic momentum.”

Translated: The American taxpayer should subsidize home mortgages for foreigners, because ending this policy would decrease demand for housing, which would cause home prices to decline.

Every day we hear that young Americans cannot afford to buy homes given the massive increase in housing prices we have witnessed since Covid. Director Pulte was in fact investigating the creation of a new 50-year mortgage that would allow young homebuyers to extend the term of their mortgage, spreading the payments across twenty more years just to make the monthly numbers affordable enough.

The National Association of Mortgage Brokers is telling you directly: you are subsidizing foreigners to buy homes in America, and if you stop subsidizing this, housing prices will fall.

God forbid!

Risk Concentration

Mortgage Brokers: “Excluding non-permanent residents could concentrate mortgage lending among a narrower borrower base, potentially increasing systemic risk.”

Translated: We should continue to subsidize mortgages for foreigners, because excluding them would concentrate Fannie and Freddie’s risk to just Americans.

Isn’t that exactly why we have Fannie and Freddie in the first place?

Furthermore, as HUD identified, allowing non-permanent residents to have their loans subsidized by Fannie and Freddie actually creates systemic risk.

This is a poison pill: something that makes an otherwise rational action prohibitively costly by design of the target’s own structure. Put simply, the current policy puts American citizens on the hook for mortgages provided to foreigners. If the President or Congress were to end H-1B, thousands of homes would need to be sold as their owners return home. Given the geographic concentration of H-1B holders, this could lead to localized housing price declines. Sellers might be forced to sell at a loss, or simply return to their home country, leaving the house to be handled in foreclosure. And who would pick up the tab on that? Fannie Mae and Freddie Mac. Who would pick up the tab for them? The American taxpayer.

This poison pill sits on the balance sheet providing a deterrent to ending H-1B.

And the worst part? The Government can’t even quantify the risk.

Flying Blind

In February 2026, Pulte announced that “for the first time” the Government would get visibility into the loans being sold to them by Mortgage Companies who have provided mortgages for purchasing homes from big homebuilders.

For the first time!

Eighteen years into conservatorship, the conservator announced, as news, that it was starting to look at what it was buying.

And what are those home builders saying? Who are they saying is buying those homes?

Taylor Morrison (recently acquired by Berkshire) told investors on its Q3 2025 earnings call that uncertainty related to H-1B policy had weighed on nonresident buyer activity, with Dallas, Austin, Atlanta, and the Bay Area feeling the greatest impacts. Toll Brothers told investors in early 2026 that the same uncertainty had paused demand from certain buyer segments. These are companies that sell enormous loan volume into Fannie Mae. The sellers disclose the risk to their shareholders, as securities law requires. We, the taxpayers who ultimately carry the risk on the mortgages used to buy the houses, get nothing.

So how much of a problem is this?

HMDA, the federal mortgage disclosure database, does record borrower demographics, including “Asian Indian” as a specific category. It’s an imperfect proxy: plenty of those borrowers are citizens or green-card holders. But as a rough outline of where the exposure lives, the 2024 data is striking. My analysis of 2024 HMDA data shows that in Forsyth County, Georgia, the Atlanta tech exurbs, Asian Indian borrowers received 43% of all conforming home-purchase originations, up from roughly 16% in 2018. In Middlesex County, New Jersey; Loudoun County, Virginia; and Collin County, Texas, the share exceeded one-third. These are conforming loans: the kind Fannie Mae buys. The kind you guarantee. Whatever fraction of that is visa-contingent, it is not trivial, and it is concentrated in a handful of counties.

I thought it was just hypothetical, but it’s starting to get real.

It’s Happening

This month, Bloomberg showed us what it looks like when the music slows.

Their feature on the suburbs north of downtown Dallas (Collin County) describes one of the fastest demographic transformations in American history, driven by visa-holder demand in the tech sector. Frisco’s Indian population grew from roughly 6% to 20% in a decade; locals took to calling the corridor “Dallaspuram.”

Home prices were bid up on the assumption that the pipeline would never stop.

Now it’s stopping. Tech layoffs, six-figure visa fees, tighter immigration rules, AI disruption. The market built on that demand is cooling in real time, with prices in the corridor already correcting.

How dependent had the market become?

One builder north of Dallas told Bloomberg that South Asian buyers had peaked at 70% of all its sales and have now fallen below 30%. Seventy percent. An industry analyst put the question plainly: with those buyers leaving, who is there to replace them? When seven in ten of your buyers come from one visa-dependent pipeline, you don’t have a customer base. You have a single point of failure. And Fannie Mae bought the loans.

And then there’s the part of the article that should be taped to the wall of every FHFA risk office. A Frisco real estate agent described a client sitting on an $800,000 underwater mortgage who had a simple plan: hand over the keys and fly home. In the agent’s words: “He will not lose anything. The banks will lose.”

On that, the agent is wrong. It’s not the banks who will lose, because it’s not the banks guaranteeing that mortgage.

It’s the American Taxpayer.

And Now They Want to Take It Public

President Trump is actively seeking to take Fannie Mae and Freddie Mac out of conservatorship and IPO back onto the public markets, possibly the largest IPO in history.

Think about what that means. The government is preparing to sell the public shares in enterprises carrying a concentrated, policy-correlated, visa-contingent credit exposure that the seller’s own regulator admits it cannot see. You cannot price an IPO, set a capital rule, or write a post-conservatorship guarantee around a risk that has never been counted. In 2008 the taxpayer was ambushed by risks the enterprises didn’t understand. In 2026 we are preparing to re-privatize them with a known unknown sitting on the books.

By listing these two companies Trump wouldn’t be taking the risk away from the American people, he’s just allowing existing investors like Bill Ackman to take their profits while leaving the bag with the people who held it in 2008 and are wearing it now: the American taxpayer.

What Has to Happen

Nobody needs to be denied a mortgage. Non-permanent residents can buy homes through jumbo, portfolio, and foreign-national lending. Private capital takes risk and receives a return. What they should not get is the subsidy Americans paid $190 billion to preserve.

FHFA should require the enterprises to measure and disclose exposure by residency status, stress-test it against visa-policy scenarios, and follow HUD’s lead: restrict the conforming guarantee to U.S. citizens and lawful permanent residents going forward. And no release from conservatorship should proceed until that exposure is quantified and disclosed.

We did not bail out Fannie Mae so it could subsidize mortgages for foreign workers.

The idea that Boston could produce anything of value in the space was so unbelievable that people could not accept it.

Slater Victoroff, cofounder of a small Boston startup called Indico had just watched Nvidia’s Jensen Huang present his company’s keynote in April 2016. Huang demonstrated a remarkable new AI technology that could generate realistic images in the style of Romantic-era oil paintings from a simple text prompt. This tech was according to Wong, the product of “incredible progress” by Yann LeCun’s laboratory at Facebook.

But it wasn’t LeCun’s progress.  The technology was developed by Victoroff and his colleagues Alec Radford and Luke Metz, young Americans, in their startup above a wine bar in Boston’s Leather District.

“It just gutted us,” Victoroff told the Boston Globe.

Shortly after the demo, Radford left for the Bay Area to join OpenAI. In 2018 he was the lead author on the paper that introduced the phrase “generative pre-training” or as we now refer to it: GPT.  OpenAI president Greg Brockman credited Radford for the initial breakthrough behind ChatGPT. Wired compared his role at OpenAI to Larry Page’s invention of PageRank.

Non-Competes Kill Innovation

There is one labor market policy question on which Silicon Valley, the FTC, labor economists, and both political parties have reached rare consensus: non-compete agreements are anti-competitive and anti-innovation.

The argument is not complicated. A non-compete agreement prevents a worker from leaving their employer to join a competitor, start a new company, or pursue a better idea. It captures labor below market value. It traps the worker’s best thinking inside the institution that holds the agreement rather than letting it flow to where it would be most productive. It kills new business formation. It concentrates innovation surplus in incumbents who can use captive talent to defend their positions rather than compete for it.

The evidence is so overwhelming that the FTC moved to ban non-compete agreements outright in 2024. The commission’s own research found that non-competes suppress wages, reduce job mobility, reduce new business formation, and reduce innovation. 

Silicon Valley….brought to you by Free Movement of Human Capital.

California built Silicon Valley in part by refusing to enforce non-compete agreements. The engineers who left Fairchild Semiconductor to found Intel, who left Intel to found startups, who left startups to found the next generation. They had mobility.  Free movement of people led to free movement of ideas and free movement of ideas led to Silicon Valley. 

Where workers are free to move, ideas flow. Where they are captive, incumbents calcify.

Non-competes kill innovation.

Pulling Up the Ladder.

Whatever happened to Fairchild Semiconductor?

Silicon Valley was built off of non-competes and the movement of people.  But what happens when the companies that get built want to maintain their position?

American engineers are expensive.  They have rights. They can quit. They can negotiate. They can organize. They can take their knowledge to a competitor, start a company, pursue a better idea. They have leverage precisely because the labor market is supposed to be free.

What if you could opt out of that?  

Similar skills but no optionality, no leverage. 

Someone willing to do what the employer wants without creating the risk that they will take that knowledge and take it elsewhere for their benefit.

That would really solidify the position of the incumbent.

Enter H1B.

A worker whose immigration status is tied to a single employer cannot leave that employer without risking deportation. 

That’s exactly the situation H-1B workers find themselves in. For workers from India, as many H-1Bs are, the green card backlog now stretches beyond a decade.  Walk out the door and those years disappear. The employer has not just hired a worker. They have acquired a captive one.

This is a non-compete agreement. The worker never negotiated it. They cannot escape it. And unlike a standard non-compete this one is enforced by the federal government through the immigration system itself.

Every single mechanism that makes non-competes anti-innovative applies identically here:

The worker cannot leave for a competitor. The worker cannot start a new company without losing their status. The worker cannot pursue a breakthrough idea if it means leaving the sponsoring employer. The worker’s best thinking is captured by the institution rather than flowing freely to where the market would value it most.

New business formation is suppressed.

Innovation surplus concentrates in incumbents.

If non-competes are anti-competitive and anti-innovative then the employer-tied H-1B visa is anti-competitive and anti-innovative.

Instead of innovation, we get whatever the H1B sponsoring employer wants.

Don’t Take My Word for It.

If tech companies wanted H1Bs not because they needed workers but rather because they want to stifle innovation what would we expect to see?

Think about your experience with Silicon Valley recently, outside of what Radford did with AI.

In a 2021 Breitbart interview, JD Vance put it bluntly: “Most of Silicon Valley, they’re actually parasites off of human attention. They get us to stare at our phones instead of spending time with our families.”

Sound familiar? 

Alex Karp, CEO of Palantir, believes that Silicon Valley has lost its way, spending its time on “inconsequential things” while failing to deliver the foundational innovation its resources and talent should produce. 

Maybe it’s not about innovation at all?

Palmer Luckey, the founder of Oculus and builder of Anduril, watched it from inside. “You would not believe what I saw in Silicon Valley.  The H-1B abuse is insane. It is obviously a scam to replace US workers with, basically, slave labor that can’t ever escape.”

The pattern is consistent. Captive labor produces compliance and benefits the incumbents who get what they need built.  It does not produce the innovative technologies that we need for our future.

The Companies That Prove It

Jensen Huang, Chris Malachowsky, and Curtis Priem were engineers at Sun Microsystems and LSI Logic. In 1993 they left their employers, combined what they knew, and founded Nvidia.  All three had the freedom to walk out the door.

You know the rest of the story, Nvidia built the chips that power the entire AI revolution.

Imagine they were H-1B workers tied to Sun Microsystems. They can’t leave. The rational choice is to stay, collect the salary, and build what Sun needs built. Nvidia never gets founded. The GPU never gets built. The AI boom never happens.

The program that claims to drive innovation would have prevented the single most consequential act of technological entrepreneurship of the last thirty years.

Now let’s do SpaceX. The most innovative technology company in America and arguably the world is legally prohibited from hiring H-1B workers for its core work. ITAR, the International Traffic in Arms Regulations, restricts access to defense and aerospace technology to US persons. SpaceX cannot staff its rocket program with visa holders. It has to hire Americans with full labor rights, full mobility, and full leverage.

The result: reusable orbital rockets, Starship, Starlink, a launch cadence that has transformed the entire space industry, and a cost-to-orbit reduction that no other entity on earth has matched.

The constraint the H-1B defenders label as anti-competitive turns out to be the condition under which the most ambitious work gets done.

Free workers build the future. Dependent workers like those on H1B maintain the incumbent.

The problem wasn’t greed. Every individual operator was making a perfectly rational decision. Pump as much as you can, as fast as you can, before the next man takes it. The result of all those rational individual decisions was collective destruction. Oil prices crashed from $2 a barrel to three cents. The pressure depleted. The field that had seemed infinite was exhausted within years.

This is the tragedy of the commons in its purest form. A shared resource, rational individual actors, catastrophic collective outcome. No villain required.

Texas eventually solved it by creating the Texas Railroad Commission, which regulated extraction rates to preserve the long-term value of the resource. Nobody called this socialism. Nobody called it anti-capitalist. They called it sound resource management — protecting a commons from being strip-mined into worthlessness by short-term incentives.

America’s labor market is facing its own Spindletop moment. And the mechanism destroying it goes by the name H-1B.

What the Free Market’s Own Champions Said

Milton Friedman, the Nobel Prize-winning economist and the intellectual godfather of modern free market capitalism, said it plainly:

There is no doubt that the [H-1B] program is a benefit to their employers, enabling them to get workers at a lower wage, and to that extent, it is a subsidy.

Thomas Sowell, perhaps the most influential conservative economist of the last half century, addresses any skill shortage claim with characteristic precision:

To an economist, it is incredible to speak about shortages without talking about prices, in this case wages.

If there were a genuine shortage of skilled American workers, wages would rise until the shortage cleared. That is how markets work. The fact that wages have not risen — that they have in fact been suppressed — tells you everything.

Both men spent their careers fighting government distortions of free markets. The people defending the program today as free market capitalism have not read their own canon.

The Arbitrage Nobody Names

The standard defense of the H-1B program goes like this: America faces a shortage of skilled technical workers. Importing talent fills that gap, drives innovation, and grows the economy. The workers come voluntarily. The contracts are legal. What’s the problem?

The problem is what the compensation package actually contains.

When a company sponsors an H-1B worker, they are not simply offering a wage. They are offering a wage plus a life in America: access to US infrastructure, institutions, safety, opportunity for children, and the path to permanent residency. That package has enormous value, value neither created nor paid for by the employer.

The worker, rationally, accepts a wage below what a comparably skilled American would demand, because access to America itself is making up the difference. The employer captures the vast amount of surplus from this transaction. The American public that built and maintains the commons being receives nothing.

This is not a voluntary market transaction between equals. It is a corporation trading something it does not own in exchange for a labor discount it pockets entirely.

Harvard economist George Borjas documented the outcome in a 2026 working paper: H-1B workers earn approximately 16% less than comparable American workers, after controlling for education, occupation, age, and geography. The gap is largest — nearly 29% — for the most skilled workers. The people with the most to offer are the most underpaid, because they have the most to lose by leaving.

The arbitrage is not incidental to the program. It is the program.

Who Actually Gets the Visas

The public debate about H-1B centers on a particular image: the brilliant engineer at Google or Meta, the person whose unique skills justify an exception to normal labor market rules. This image is carefully maintained and almost entirely false.

Federal data from the US Citizenship and Immigration Services tells a different story. Among the top ten beneficiary firms, Infosys, Tata Consultancy Services, and Cognizant collectively received more visas than Apple, Microsoft, Google, and Meta combined. These are not innovation companies. They are the Walmart of the consulting world — outsourcing firms that place workers at client sites to perform functions that are oversupplied domestically.

As one researcher’s paper put it bluntly: the results are “more supportive of the narrative in which H-1Bs crowd out alternative workers, are paid less than the alternative workers whom they crowd out, and thus increase the firm’s profits despite no measurable effect on innovation.”

More tech support at suppressed wages. That is what the program delivers.

The Spindletop Dynamic

Back to Beaumont, 1901.

Every company drilling at Spindletop was extracting from a shared pressure system. One company’s extraction reduced the reservoir available to everyone else. The rational response to this was to drill faster — which accelerated the depletion for everyone. No individual actor had any incentive to slow down. The collective result was ruin.

America’s labor market works the same way. Every company that imports an H-1B worker at below-market wages is drawing down the pressure in the shared reservoir.

When you flood the market with workers who accept below-market wages because America itself is subsidizing the discount, you destroy that price signal for Americans to enter the market. The ROI on a STEM education collapses. Young Americans rationally redirect their energy elsewhere. Then the same companies that suppressed domestic wages turn around and claim there is a domestic talent shortage to justify importing more H-1B workers.

California, home to Silicon Valley, systematically dismantled advanced mathematics education in the name of equity. The state that generates the most H-1B visa demand destroyed the educational pipeline that would produce domestic technical talent. It is difficult to look at this and not notice that a workforce that cannot do advanced math is a workforce that must be imported.

The Constitutional Obligation

The preamble to the United States Constitution identifies who the government exists to serve: “ourselves and our posterity.”

Not the world. Not global labor markets. Not the shareholders of multinational corporations. Our posterity: the Americans who will inherit whatever we leave them.

The Texas Railroad Commission was not anti-capitalist. It was the government fulfilling its basic obligation to protect a shared resource from being strip-mined by private actors whose individual incentives pointed toward collective ruin. The TXRRC preserved the long-term value of the oil commons for Texans and their descendants.

The government’s job is not to maximize short-term corporate profits. It is not to provide a labor supply subsidy to companies that find domestic workers inconveniently expensive and inconveniently empowered. It is to secure the blessings of liberty for ourselves and our posterity by protecting the labor market conditions that allow Americans to build lives, form families, and pass something forward.

This is not socialism. It is the most basic function of legitimate government, derived from the founding document of the Republic, supported by the oldest precedent in American resource management.

What the Texas Railroad Commission Actually Did

After Spindletop collapsed, Texas did not nationalize the oil industry. It did not cap profits. It did not tell companies where to invest or what to produce.

It set extraction rates that preserved the long-term value of the shared resource. It prevented individual rational actors from collectively destroying what they all depended on. It was, in the most precise sense, a market correction. Government addressed a failure that unregulated markets could not fix themselves because no individual actor had any incentive to act differently.

The analogous intervention for H-1B is not complex and the mechanism already exists in American energy policy.

When the federal government sells crude oil from the Strategic Petroleum Reserve, it does not run a lottery. It runs a competitive auction with a minimum acceptable price. The government owns the resource. The floor reflects the strategic value of what is being sold. The market bids above it. The public captures the value rather than handing it to private actors for free.

This is precisely the mechanism H-1B requires: an auction with a floor.

The floor establishes the minimum value of access to the American labor commons because below that price the arbitrage against American workers is simply too large to permit. The auction above the floor forces employers to reveal what the labor is actually worth. If the skill is genuinely scarce and the worker genuinely irreplaceable, the employer bids accordingly and wins the visa. If they won’t bid market wages they have answered the scarcity question themselves.

A wage floor alone does not solve this. A floor simply raises the quality of the worker the employer gets to underpay. The employer still captures the residency discount, now on a more productive worker. The arbitrage continues, slightly repackaged.

An auction eliminates the arbitrage. The residency premium the worker would have accepted gets captured instead by the Treasury, its rightful owner. Borjas modeled this in his working paper and found the revenue-maximizing fee between $118,000 and $264,000 per visa. An auction finds that number dynamically rather than requiring a regulator to guess.

None of this is anti-immigrant. The H-1B workers themselves are also victims of the system: bound to employers who exploit their visa dependency, underpaid relative to their skills, held in a form of labor control that would be unacceptable if applied to any other worker. The problem is not the workers. The problem is the mechanism that converts their desire to live in America into a subsidy for corporate margins.

Fix the mechanism. Protect the commons. Secure the blessings of liberty for our posterity.

That is what the Texas Railroad Commission did in 1919. It is what the federal government should do now.