What’s next for defense startups?
Experts disagree on Silicon Valley’s future relationship with the military
Silicon Valley and the United States military go way back.
According to the Quincy Institute for Responsible Statecraft, the wealth and technological accomplishment of Silicon Valley today can be traced back to crucial U.S. government contracts in the 1960s. In the early years of that decade, the government was responsible for nearly all semiconductor chip sales coming out of firms like Fairchild’s, General Instruments, and Motorola.
As Silicon Valley companies became more established and accumulated wealth, many companies entered the defense technology sector, where they were bolstered by the high-price, small-volume contracts that only government projects like NASA’s Apollo missions could provide. By 1990, Silicon Valley boomed as the nine biggest defense companies raked in $11 billion per year in defense contracts. The Cold War was good for business.
However, the military’s relationship with defense companies shifted as global tensions changed. At a 1993 dinner colloquially known as “The Last Supper”, military officials told over 50 of their regular contractor companies that federal budget cuts meant they would have to merge or lose their contracts altogether. As the 90s dragged on, Lockheed Martin, RTX Corporation (formerly Raytheon), Boeing, General Dynamics, and Northrop Grumman consolidated their grasp over large government contracts, becoming the “Big Five” of the defense sector.
As federal funding became more available in the following decades, the national security apparatus shifted once again towards investing in early-stage defense technology companies. For example, in 2005, the venture capital arm of the CIA lent $2 million to an uncertain, early-stage company promising revolutionized surveillance. That company has since grown to become the multibillion-dollar corporation known today as Palantir.
One recently-released book, Valley of Death: How Big Tech is Fueling the Future of War by Sharon Weinberger, argues that the investment dynamic between the U.S. military and emerging defense startups extends to this day. Weinberger, a nationally acclaimed author, former correspondent for the Wall Street Journal and Foreign Policy magazine, and the editor in chief of the Bulletin of Atomic Scientists, writes in Valley of Death that established Silicon Valley companies aren’t the only ones offering their technology to the military anymore. She claims that emerging startup ventures with increasingly uncertain ideas are receiving more government contracts and support, and that the military-industrial complex is gaining unwarranted influence over the direction of American security.
Two MIT experts counter that the topic is more nuanced than Weinberger suggests. They claim that support for emerging defense businesses is a necessary part of defense technology innovation, and that the entire infrastructure of the military-industrial complex may be forced to shift towards adaptability at scale.
Dual-use technology at MIT
MIT is no stranger to the idea of promoting the military-industrial complex. In fact, the Institute has given its students the opportunity to learn how they can contribute to the dynamic by developing dual-use technology strategies.
The U.S. Department of Commerce defines dual-use technologies as commercial products with both civilian and military applications. MIT’s Defense and Dual-Use Ventures Lab at the Sloan School of Management prepares students to tackle the complexities of technology development, regulatory requirements (such as navigating differing field testing standards for commercial and defense products), and the process of drawing civilian and government contracts.
MIT’s Mission Innovation X (MIx) program is a related initiative that supports MIT student startups navigating the dual-use space through five levels of operational capacity: Commercial Funding Readiness Level, Commercial Customer Readiness Level, Mission Funding Readiness Level, Mission Customer Readiness Level, and Technology Readiness Level. There’s an emphasis from MIT programs on treating dual-use as a strategy rather than a category of products.
Professor Gene Keselman MBA ’17, the executive director of MIx, is a lecturer at the Sloan School of Management, a member of the Reagan Institute’s National Security Industrial Base Advisory Board, and a senior reservist for the Air Force Reserve. In his lectures, he doesn’t want students to think that defense products, civilian products, and dual-use products are three completely separate categories that can’t overlap.
“The whole concept of dual-use is that [companies don’t] have to set out to be for the security of the people,” Keselman said. “They can be for creating more resilient energy infrastructure, health care, or for creating a new type of motor that is more efficient.”
Once excellent technology is developed, Keselman teaches students to ask, “If we have a crisis, can it help solve that crisis?”
MIx aims to guide startups to become the best of both worlds. Several active companies built their foundation through the Defense and Dual-Use Ventures Incubator, including cybersecurity company Blue Cloak, AI fact-checking tool Jaxon, and brain health company Neurogeneces. The wide diversity of companies coming out of MIx speaks to how dual-use strategies allow them to access both military and civilian markets.
Do defense startups serve American interests?
Not everyone agrees on the dual-use nature of today’s startup defense companies.
On Oct. 15, Weinberger held a book talk at the Cambridge Public Library to discuss Valley of Death with her readers.
“The Silicon Valley ‘fake it ‘til you make it’ [ethos] has come to the defense industry,” she remarked.
Weinberger expressed concern that Silicon Valley corporations are receiving vast military contracts for projects they may not be able to practically complete. It’s certainly true that Silicon Valley is receiving vast military funding and deepening its ties with the government: in 2026, Anduril Industries signed another multibillion-dollar contract with the US military, Google and Microsoft finalized agreements to integrate their AI into “warfighter decision-making” for the Department of Defense, and Palantir expanded its role in automating Pentagon targeting capabilities.
In Valley of Death, Weinberger points to failed startups that raised billions of dollars of investment and won government contracts on impossible pseudoscientific ideas, driving home her claim that military investment in private business — especially the recent wave of emerging startups receiving contracts through political connections — doesn’t serve the American people.
However, MIT William Porter Professor of Entrepreneurship Fiona Murray disagrees with this perspective. Murray, who is also the Chair of the NATO Innovation Fund and a former member of the British Prime Minister’s Council on Science and Technology, believes that startups are a cheaper and faster way to develop military technology as compared to long, slow government contracts with the “Big Five.”
“The economic argument is that startups are more efficient,” she explained.
It’s also worth noting that government investments in failed ideas are not particularly new. During the same era in which early-1960s Silicon Valley companies enjoyed the success brought by U.S. investments in their semiconductor production, many scientists also championed government investment in “molecular electronics,” which are electronics powered through current-carrying organic molecules. However, molecular electronics research was ultimately an expensive and failed military investment. Instances like these suggest that military investment in startup science, while being successful at times, has always come with the costly risk of failure.
Still, Weinberger commented at her Cambridge Public Library book talk that the idea of emerging Silicon Valley companies serving American interests through their unique strategy is “brilliant marketing that just means nothing.”
Valley of Death points to the billions of tax dollars spent on failed defense startups, as well as recent instances of civilian deaths caused by the implementation of early-stage artificial intelligence into wartime decision-making, to illustrate Weinberger’s concerns about “supporting bad science.” When asked at the book talk if she thought the benefits of startup efficiency outweigh their potential for setbacks, she responded, “It’s only efficient if it works.”
Is the political influence of technology leaders a new trend?
“Could a rogue billionaire buy and sell a nuclear weapon?”
According to Weinberger, this question occupied the minds of analysts across the Pentagon in 2013. With up-and-coming tech mogul Elon Musk gaining more influence, it was a hypothetical that seemed increasingly plausible. Eventually, the analysts concluded that, yes, wealthy individuals could have the means to privately build and trade weapons of mass destruction.
Valley of Death highlights that private power over warfare continues to grow because of the newfound role of privately-owned technology in military operations. For example, when Elon Musk didn’t agree with Ukraine’s decision to attack a Russian fleet in Crimea in 2022, he cut off Ukraine’s access to Starlink coverage in several border regions.
Weinberger argues that business leaders like Musk shouldn’t unilaterally be making consequential decisions in war. Anthropic leaders announced they would not allow the Pentagon to have unrestricted use of Claude for military purposes, and the action was seen as a private company reining in military action based on their own private ethical guidelines. In her book talk, Weinberger gave the controversy as another example that military operational control lies less with elected officials — and more with billionaires — than ever before.
Murray countered Weinberger’s claim, saying that “at the end of the day, the government, as a customer, ultimately gets to decide how technology is used.” The nature of government military contracts, she stated, gives the military the final word.
Furthermore, commercial interests have always played into military decisions. She argues there’s been no shift — that Cornelius Vanderbilt and Nikola Tesla likely had the same influence centuries ago as Musk and OpenAI CEO Sam Altman do today.
Murray claims both that the political power of commercial firms today is not materially different than in history, and that the sense that billionaires are getting more politically powerful has more to do with high-profile people using social media to reveal powerful connections than an actual shift in power.
In her words, “we’re just starting to see something that’s always been happening.”
How will the American defense industry evolve next?
According to Keselman, these powerful private companies are heading toward “an extinction event” as artificial intelligence becomes more widely integrated into defense innovation.
For instance, as soon as relatively cheap autonomous drones can effectively and consistently destroy F-35 planes mid-flight, F-35s will suddenly become obsolete; entire production systems would have to change. It will become economically infeasible to maintain our current military-industrial complex, Keselman stated, and the “Big Five” will lose contracts if their system engineering strategy isn’t adaptable enough.
Keselman argued that this phenomenon would force the model of government technology acquisition to change, as “it no longer makes any sense for you to continue to buy something that costs $15 million apiece when the other guy now can take it out for a thousand bucks.”
Global warfare has already begun evolving in light of this new challenge. In Ukraine, the government contracts small producers that change their output based on demand on a month-to-month basis. Having foresight into where the enemy may take technological production next is a lifeline. Keselman stated that the war in Ukraine looks technologically different every six months, and that soon global war may similarly evolve faster than ever. He predicted that as soon as the United States’ rivals begin to innovate beyond our current capacity, America will be forced to move faster as well.
Right now, the defense industry places a large emphasis on systems engineering and massive, technology-specific manufacturing capacity. Keselman argued that industries are going to have to restructure.
“Fundamentally, the shift is going towards manufacturing capacity and adaptation capacity rather than system capacity,” Keselman explained. “Questions like, ‘How many can you make? How cheap is it? How adaptable is it? Even more so, how do you now fight a war?’ won’t be asked, strategized, and implemented once; they’ll be asked repeatedly as innovative capacity in organization and production becomes more crucial.”
It’s not just the products that have to change, he stated. “Your company has to be able to adapt as well because everything’s moving so fast. You cannot possibly be structured in a way that you cannot respond to demand signals anymore.”
When asked if the “Big Five” could rise to the challenge of changing their production model, Keselman was quite clear. “They can’t and they won’t,” he said.
He explained that that’s not how things are done at large, established military corporations, which is why defense startups at the university level are so crucial. He noted that in Ukraine, the most adaptable and successful defense production companies are led by “incredibly young” people. The American counterpart to that — in Keselman’s view — is college-aged founders who are less burdened by “the way things have always been done.” Valley of Death balks at the inexperience of some military startup founders, but Keselman sees unburdened young leadership as a strategic advantage.
For better or for worse, the defense industry is having its moment in the spotlight. Dual-use strategies bring in more cash for innovative companies, but they are criticized for drawing attention away from civilian products. Wealthy figures in the field may have always wielded political power, but their influence is much more apparent in the modern day. The American defense industry may be forced to completely restructure to keep up with evolving global threats.
According to Keselman, it is “impossible to tell” when the next defense technology revolutions will happen; the current pace of innovation is both exponential and inconsistent. There’s a lot about emerging defense technology that remains to be seen.