Al Goldstein was eight years old when his family arrived in the United States, refugees from a country that had made it clear it no longer wanted them. Uzbekistan, in 1988, was not a place for a Jewish family to stay, and so the Goldsteins came to America with, as he later put it, absolutely nothing, and two small children in tow.
He did not speak English. He did not know that he would go on to found not one but two companies each worth more than a billion dollars.1 I think about Goldstein’s story often, not because it is remarkable, but because it is, in a country of nearly three hundred million people, almost banal in its repetition. Nearly six in ten of America’s unicorn companies — the privately held firms valued at a billion dollars or more, the closest thing capitalism has to folklore — were founded or co-founded by someone who was not born here.2 I want to spend the length of this essay asking what that number is actually telling us, and why, in an era anxious about borders, it is worth listening to.
I. The Arithmetic of Ambition
The most recent analysis by the National Foundation for American Policy found that immigrants founded or co-founded 455 of America’s 775 unicorn companies — fifty-nine percent, up from fifty-five percent just four years earlier. Add the companies founded by the American-born children of immigrants, and the share climbs to roughly two-thirds. Collectively, these companies are worth about five trillion dollars.3 Anthropic, OpenAI, SpaceX, Stripe, Epic Games, Chobani, and Perplexity — a peculiar and revealing list, spanning artificial intelligence, rockets, yogurt, and video games — each counts an immigrant among its founders.4 Almost a quarter of these companies were founded by someone who first set foot in America as a student, on a visa meant only for study, not for empire-building.4
I confess a certain impatience with the version of this argument that stops at the impressive total and calls it a day, as though a big number were itself an explanation. It is not enough to say that immigrants built valuable companies; the more interesting question is what it is about the condition of being an immigrant that seems, again and again, to correlate with founding one. A 2025 study by the National Bureau of Economic Research, examining nearly ninety-one thousand American startups, found that companies with mixed founding teams — immigrants working alongside the American-born — grew faster than native-only teams, employing twenty percent more people within three years of launch.5 My own suspicion, which I offer as a suspicion and not as settled science, is that migration is itself a kind of sorting mechanism. The people willing to leave everything familiar — language, weather, the specific smell of a childhood kitchen — for the uncertain promise of a different life are, definitionally, a self-selected population of risk-tolerant people. Founding a company is also an act of leaving something safe for something uncertain. It should not surprise us that the same people are disproportionately willing to do both.
II. The Machines We Are Building, and Who Is Building Them
Nowhere is this more consequential, or more contested, than in artificial intelligence, the field on which so much of the coming decade’s economic and geopolitical order is expected to turn. MacroPolo’s Global AI Talent Tracker, which follows researchers accepted to the field’s most competitive conferences, found that American institutions now employ fifty-nine percent of the world’s top-tier AI researchers, up from fifty-one percent less than a decade ago. That dominance is built almost entirely on importation rather than domestic production: the United States trains a comparatively modest share of these researchers itself.6 Strip away researchers of Chinese origin alone, and America’s lead over China in elite AI talent would shrink by roughly two-thirds.7 I do not think it is an exaggeration to say that the United States’ position in the defining technology race of this century rests less on any innate advantage than on a policy choice: whether the country continues to let the world’s most ambitious young scientists come and stay.
That policy choice has, in the past year, wobbled. The H-1B visa, the primary legal channel for skilled foreign workers, remains dominated by applicants from India, who accounted for roughly seventy-one percent of approvals in fiscal year 2024.8 In September 2025, the administration imposed a new hundred-thousand-dollar fee on petitions filed for workers outside the United States — the most significant cost increase in the program’s history. Approvals in the following quarter fell forty-five percent below the same period a year earlier.9 I do not think a hundred-thousand dollars is, by itself, an outrage; large institutions absorb costs like this without much trouble. What troubles me is the signal it sends to the individual researcher, the twenty-six-year-old in a lab in Bangalore or Beijing deciding where to spend the next decade of a short working life. Policy is read as intention. A fee like this reads, to the person it is aimed at, like an answer to a question they had not yet finished asking.
III. The Argument Against, and Why It Deserves a Fair Hearing
It would be dishonest, and beneath the seriousness of the subject, to present only one side of this. The case against the H-1B program in its current form is not a fringe position; it is held by serious economists and institutions. The Heritage Foundation has documented that most H-1B positions pay wages below the occupational median, and that only one in six pays at the highest wage tier — evidence, its authors argue, that the program has drifted from recruiting elite talent toward supplying discount labor.10 A widely discussed empirical study, conducted by economists using confidential firm-level data, found that additional H-1B hires did not raise a firm’s patenting output, “substantially crowd out employment of other workers,” and left average wages lower while raising profits — a finding one labor economist called consistent with what companies themselves say the program is for.11
The rebuttal, also empirical, is that the picture is more complicated than either side’s slogans allow. Research from the National Foundation for American Policy has found that, on average, every H-1B position requested by an employer is associated with several additional American jobs created around it.12 And when Congress previously cut the H-1B cap from 195,000 to 65,000, employment of comparable American workers did not rise — undercutting the simplest version of the substitution story.13 I suspect the honest answer sits, unglamorously, in the middle: the H-1B system, as administered through a lottery until recently and riddled with loopholes around prevailing-wage calculations, has almost certainly been used by some employers to underpay some workers, while simultaneously functioning, at the frontier of the labor market, as the pipeline through which a disproportionate share of the world’s most consequential companies and technologies have entered the United States. Both things are true. Good policy would discipline the abuse without strangling the pipeline, and I am not persuaded the current approach — a blunt, universal fee — is capable of telling the difference between the two.
IV. What I Think We Owe Each Other
I do not make this argument because I think immigrants are more virtuous, more intelligent, or more deserving than anyone born here. I make it because the evidence, read plainly and without the rhetorical inflation that attends most immigration debates, points to a specific and unglamorous truth: a country that makes itself difficult to enter, for the ambitious and the skilled, is choosing to compete with one hand behind its back, at precisely the moment when the game — artificial intelligence, biotechnology, energy — has never mattered more or moved faster.
Al Goldstein could not have known, on that flight from Tashkent, what he would build. Neither could the country that let him in. That, I think, is the real argument for immigration as a competitive advantage — not that any single newcomer is guaranteed to become a Goldstein, but that a nation confident enough to make the bet, again and again, on people it cannot yet measure, tends to win more than it loses. The countries currently tightening their doors are, in effect, betting the opposite: that they already know everything they need to know about who is worth letting in. I am not sure any country has ever been that wise.