Silicon Valley would not exist without government-funded research. Foundational technologies, including the semiconductor and the Internet, emerged from Cold War–era military research programs. As graduate students at Stanford, Larry Paige and Sergey Brin relied on funding from the National Science Foundation to develop the search algorithms that would eventually become Google. The touchscreens and lithium-ion batteries that we now carry around all day were likewise developed in university labs funded by government grants. Even generative AI—incessantly touted as the crowning achievement of the free market, upon which the fate of the American economy depends—emerged out of decades of research underwritten by the Department of Defense (DOD). Geoffrey Hinton, the Nobel Prize winner known as the Godfather of AI, left his academic position in the United States precisely because he wanted to avoid Pentagon contracts. Hinton nevertheless turned to the Canadian government to help fund his lab at the University of Toronto, which eventually produced leading AI researchers for OpenAI, Google, and Meta.

Given how much Silicon Valley has profited from government-funded research over the years, you might expect a certain amount of reverence for the system. At the very least, even the staunchest techno-libertarian rationalists should recognize the value in not killing their golden goose. Yet Silicon Valley elites are at the very heart of the Trump administration’s devastating assault on public science funding—and, not coincidentally, have positioned themselves to profit off the wreckage. In particular, conservative venture capitalists Peter Thiel and Marc Andreessen have parlayed their extensive ties with the president into an unabashed assault on universities and institutional science. In private text messages leaked to The Washington Post last year, Andreessen wrote that “universities are at Ground Zero of the counterattack.” He characterized Stanford and MIT as “mainly political lobbying operations fighting American innovation at this point” and vowed that universities would “pay the price” after “they declared war on 70% of the country.” Most troublingly, Andreessen called for the National Science Foundation to receive “the bureaucratic death penalty.”

Thiel has long set his sights on shifting federal research dollars from universities to private industry. In numerous interviews, Thiel has pointed out that we have 100 times as many science PhDs as we had a century ago, yet the rate of progress is about the same. The claim itself is dubious. He offers no clear benchmark by which to measure scientific progress, nor does he consider the possibility that science has become more complicated after a century of advancement. Could it be that more bureaucracy, however flawed, is needed to operate a Large Hadron Collider as compared to a microscope and Bunsen burner? For Thiel, the answer is a definitive no: “The average PhD is 99% less productive than people were 100 years ago,” he concludes with unwavering confidence. But even he cannot ignore the successes of Cold War research programs. However much it might pain his libertarian soul, Thiel acknowledges that DARPA—the research and development arm of the DOD—functioned well early on, but he has conveniently decided that it was a one-time acceleration that “came at the price of completely corrupting the institutions.”

In any case, the justifications now matter less than the actual results. Trump entered his second term with a plan to cut federal science funding and extort prominent universities with threats of targeted budget cuts. The attacks were orchestrated by Michael Kratsios, the director of the White House Office of Science and Technology Policy, who previously served as the chief of staff to Thiel at his venture capital fund. The proposed budget included funding reductions of 40 percent for the National Institute of Health, 57 percent for the National Science Foundation, and 24 percent for NASA.

Current Issue

May 2026 Issue

Although Congress has since attempted to roll back some of the cuts, the administration has already inflicted enormous damage. Over 10,000 federal workers with STEM PhDs left the federal workforce last year. University labs have been forced to fire researchers, cancel studies, or just shut down operations altogether. Some academics sought refuge in Europe; others retired early. An unmistakable chill has taken hold of the scientific establishment—one that will linger long after the Trump presidency.

Why would tech billionaires attack a system that made them enormously wealthy at virtually no personal cost? The most obvious explanation is that much of that newly freed-up funding can be redirected to the tech industry. Thiel and Andreessen position start-ups as the remedy to the supposedly bloated, inefficient scientific bureaucracy. They cast themselves as the true champions of science, locked in an existential battle against pencil-pushing charlatans. If Newton were around today, the thinking goes, he would be applying to Y Combinator and ordering swag for his B2B SaaS start-up. This grandiosity is coupled with a strong sense of paranoia. In a 2025 interview, Andreessen described the Biden administration as being preoccupied with “the raw application of the power of the administrative state, the raw application of regulation, and then the raw arbitrary enforcement and promulgation of regulation,” concluding: “Absolutely tried to kill us.”

When Trump took power, it was their turn to strike back. As science budgets got axed, portfolio companies backed by either Thiel or Andreessen—and sometimes both—received billions of dollars in federal contracts. The administration quickly deregulated crypto and threatened to punish states that enacted “onerous and excessive laws” relating to AI. This agenda was spearheaded by Trump’s policy advisers, including the billionaire venture capitalist David Sacks, who led PayPal alongside Thiel, and Sriram Krishnan, who was previously a partner at Andreessen’s investment firm.

The attacks on science also created a new talent pool for Silicon Valley to exploit: newly displaced STEM researchers. Within the AI industry, executives frequently cite the goal of creating models that are “PhD-level experts” across various academic disciplines. But training those models requires actual PhD-level experts to write relevant prompts, generate training data, and verify the output. How do you get someone with a doctoral degree in physics or math to sit down and solve hundreds of challenging problems? One way is to hire them, pay a competitive salary, and offer health insurance. Another, perhaps less obvious, approach is to kill off as many of the previous job opportunities as possible, such that highly credentialed researchers might be enticed to perform mind-numbing gig work for $30 an hour.

A multibillion-dollar industry emerged for precisely this function. The two most prominent companies, Mercor and ScaleAI, both received venture funding from Thiel. Those early investments seemingly paid off: Mercor most recently raised funds at a $10 billion valuation, while Meta bought a 49 percent stake in ScaleAI at a $29 billion valuation. This industry has grown rapidly by emulating the playbook Uber and Lyft used to appeal to drivers in the early days of ride-sharing. Their advertisements emphasize the flexibility and freedom of gig work. Jobless academics are shown hiking through the woods, reading books on hammocks, and playing sports with friends. Their testimonial voiceovers explain that, even though the academic job market has no opportunities, gig work allows them to make money while remaining in their field—even if not quite how they imagined it. “Finding jobs in academia has always been a struggle,” one contractor notes, explaining that he turned to Mercor after his institution cut off summer funding. In an advertisement for a competing platform, a Stanford-trained chemist expresses hope that working with AI will open future job opportunities, despite the otherwise bleak job market for PhD graduates.

As with the ride-hailing industry, initial promises of easy money often give way to a more complicated reality. One doctoral student in applied mathematics was tasked with solving tournament-level math problems for roughly $90 per hour. The questions were challenging, even for someone with his expertise. Yet the company would pay for only two and a half hours of work per question—and incorrect or incomplete responses received no compensation at all. This meant he faced a choice: either continue working on challenging problems for free, or give up and forfeit any payment. Other researchers reported similar experiences across gig platforms. A recent PhD graduate of an MIT engineering program recalled tabulating all of her unpaid hours for a project, only to realize the effective hourly rate was considerably lower than what had been advertised. “I originally thought that was [a] very fair, generous amount,” she explained, “but then I started keeping track of all the unlogged hours that I wasn’t really paid for and so it ended up not being super worth it.”

Silicon Valley libertarians might respond that this is merely the free market at work. After all, no one coerced the doctoral students or underemployed scientists into performing gig work. But this narrative ignores the very direct policy decisions that shaped that “free” market for researchers and academics. All of the researchers interviewed for this article described turning to gig work because Trump’s federal funding cuts made it much harder to find opportunities in their field. The cause and effect are abundantly clear: The Department of Energy cuts funding so a summer stipend disappears, or the Trump administration threatens a university and multiple postdoc positions close. “I would say it’s akin to being farmed,” the doctoral student in applied mathematics said. He described the ads for AI gig work as “clickbaity,” explaining that, as a grad student struggling to find work amid funding cuts, it is hard to resist a job description that touts remote, flexible, asynchronous work at seemingly high hourly rates.

As it stands, we increasingly rely on the altruism of young people who view science as a calling—but that paradigm can be pushed only so far before it breaks. Most scientists are not in it for the money; the same cannot be said of Silicon Valley. Hence, the 23-year-old founder of Mercor, Brendan Foody, became a paper billionaire in a matter of months by supplying AI labs with, among others, PhD researchers scrambling to remain solvent on their long and increasingly arduous career paths. “The wealthiest companies in the world are willing to spend whatever it takes to improve model capabilities, where Mercor is sitting at the forefront and sort of the primary bottleneck,” Foody explained in a recent interview. In reality, the humans training the models are the bottleneck—gig platforms, to their credit, just figured out how to profit from them.

The myth of the free market is often used to obscure what are ultimately value judgments. Even Peter Thiel’s libertarianism stops wherever Palantir’s interests begin. Basic science research has never been particularly profitable in its own right, but society has benefited enormously from its advancement. The new bargain struck by Silicon Valley conflates wealth generation with progress. It is akin to deciding that a tree’s roots no longer need to be watered because the fruit comes only from its branches. The tech industry may suffer in the long run, but several venture capitalists will make extraordinary short-term returns. In the meantime, a generation of scientists risks getting left behind.