By Sharon Weinberger
"USA250: The Story of the World's Greatest Economy" is a yearlong WSJ series examining America's first 250 years. Read more about it from Editor in Chief Emma Tucker.
In 1990, a Pentagon official had a bright idea.
The military wanted to guarantee a reliable domestic supply of a critical computer chip. So the Pentagon official had the agency he led make a direct $4 million investment in a company working on the chips, a Silicon Valley outfit called Gazelle Microcircuits. After all, the government was already pouring money into research on manufacturing semiconductors, and the official had broad authority to make investments in technology.
The idea didn't go over well. In fact, the official who suggested it, Craig Fields, lost his job as head of Darpa, the Defense Advanced Research Projects Agency. The idea that the government would pick winners and losers -- as part of an effort to support industries and companies deemed critical to the nation's interest -- was anathema to key Republicans in the George H.W. Bush White House.
How times have changed. Several years ago, the U.S. government started putting tens of billions of dollars of direct subsidies into the domestic semiconductor industry. And the Trump administration last year reached a historic agreement to take a 10% stake in Intel, a leading American manufacturer of semiconductors.
Away from basics?
The new strategy is part of a shift in the U.S. government's relationship with scientific and technological innovation that is giving priority to private-sector research over basic science. For decades, the federal government supported basic scientific research, hoping for breakthroughs that would eventually reach industry and the military.
Now, fearing competition from China and the loss of U.S. manufacturing, the new industrial policy is being embraced across Democratic and Republican administrations. Both sides point to a new reality about tech: Private-sector companies are doing a lot of the R&D work that the government used to support at universities and other research institutions.
But there is still a debate going on -- about whether the government should still keep backing basic research in a big way, or devote the lion's share of resources to helping industry.
Democrats say the U.S. needs to keep helping basic research, because it remains a pipeline for innovation. Private companies, they say, will mostly focus on R&D in areas that seem profitable at the moment -- and not always ones that may yield unexpected gains in the longer term. The Trump administration, on the other hand, has proposed cutting basic research deeply, saying the public has lost trust in scientists and the government should back only projects that meet a "gold standard" of credibility and usefulness. Critics of the administration, meanwhile, fear that this is simply code for targeting research -- and institutions like universities -- that the White House disagrees with politically.
A new frontier
Much of the conventional wisdom about the relationship between the government and innovation dates back to 1945, when Vannevar Bush, the head of the Office of Scientific Research and Development, delivered his report, "Science, the Endless Frontier" to President Harry Truman. In it, Bush argued that funding basic science, the "pacemaker of technological progress," was imperative for downstream advances.
"A nation which depends upon others for its new basic scientific knowledge will be slow in its industrial progress and weak in its competitive position in world trade, regardless of its mechanical skill, " his report stated.
What followed were decades of support for science, in everything from computers to biology. But by the late 1980s, some key research areas, like computer science, had moved beyond the labs, and Silicon Valley was no longer heavily reliant on the government for contracts or funding.
The emerging consumer market was more attractive for many companies than the military, but market forces sometimes weren't enough to guarantee survival for innovative companies, or protect those companies from foreign investment. That is part of the reason Darpa wanted to invest in Gazelle: The Pentagon was trying to ensure a foothold in a market that was increasingly dominated by commercial interests.
Then came the terrorist attacks of Sept. 11, 2001, which was a catalyst for many changes across the national-security establishment. The traditional defense manufacturers, operating under cumbersome Pentagon contracting requirements, were well behind the private sector in computer technology.
"When 9/11 hit, one of the discoveries in the commission report and many others was that our big defense contractors, the companies traded publicly, were woefully behind the times when it came to digital information technology, just way behind the times," says Paul Bracken, emeritus professor of management and political science at Yale University.
Startups step in
The 9/11 attacks created an opportunity for new companies to win business. Usually, when the government awarded contractors a deal, it shouldered a lot of other costs for the companies, such as the R&D involved in making the product. A small number of startups, though, figured they could win fat government contracts by funding the R&D themselves with private capital -- and offer the government lower prices than the established contractors could.
A couple of now-familiar names entered the fray. Palantir, for instance, got its start helping the CIA and others with counterterrorism software. More than 20 years later, what started as a trickle of private investment has become a flood: PitchBook estimates that venture-capital deals in defense startups reached a high of about $49 billion in 2025.
The Pentagon has in recent years largely embraced this new world of venture-funded defense companies, and has created a patchwork of new divisions, like the Office of Strategic Capital, which offers funding and loans to private companies developing cutting-edge technology important to national security.
Depending on venture capital for defense innovation carries risks, however. Even if they get early assistance from the government, these fledgling companies won't survive unless they win actual government contracts. And there is no guarantee that they will land those deals.
"I think the downside is, how much patience will they have before they demand to see production contracts?" says Heidi Shyu, who as the Pentagon's top technologist in the Biden administration pioneered some of the efforts to support defense startups. "And that is the difficulty for them. All these companies, VC-funded companies, struggle to try to find a production contract. And I can tell you also, some of them die on the vine."
The basic question
The bigger question may be what happens to the government's funding of basic science, which in the past has created a pipeline for research now being commercialized by private companies.
For instance, Elon Musk's SpaceX was built on the foundations of billions of government spending, particularly through NASA, on rocket science (the company also received a Darpa contract in its early days). Another Musk-owned company, Neuralink, owes a debt to Darpa's brain-computer interface work in the 1970s. And artificial intelligence, which is poised to reshape the U.S. economy, is built on decades of government support for computer science, from Darpa and other agencies.
President Trump's cuts to basic research -- which haven't been fully explained beyond asserting a crisis in public confidence in science -- have been largely reversed by Congress, but academics and scientists say they remain worried about steady funding. The White House this month has again proposed deep cuts to the science budget for next year.
The political atmosphere is making scientists realize that the postwar compact with the federal government is over, says Jonathan Moreno, professor emeritus at the University of Pennsylvania, who has looked at the intersection of science policy and national security.
"The longstanding assumptions that the world of science made about its relationship to the American government is never going to be the same," he says.
Write to reports@wsj.com
(END) Dow Jones Newswires
April 20, 2026 14:00 ET (18:00 GMT)
Copyright (c) 2026 Dow Jones & Company, Inc.
Comments