Shaun Maguire on why foundation model companies are the new operating systems, and his unpopular take that only ~10 defense tech companies will win big
May 1, 2025 with Shaun Maguire
Key Points
- Sequoia Capital partner Shaun Maguire argues foundation model companies capture OS-level value through both verticalized applications and capital intensity, positioning xAI, OpenAI, and Anthropic as structural winners over open-source competitors.
- Maguire predicts only ~10 defense tech companies globally will reach $10 billion-plus valuations, with winners vertically integrated like Tesla rather than subcontracting like legacy primes.
- Consistent multi-year defense budget allocation matters more than single-year funding figures; erratic spending kills contract-dependent startups, as Maguire experienced when sequestration cut a $10.5 million government contract to 60 cents on the dollar.
Summary
Shaun Maguire, partner at Sequoia Capital, argues that foundation model companies will capture far more value than the market currently expects — and that the next generation of defense tech will consolidate into a surprisingly small number of winners.
Foundation models as operating systems
Maguire's core thesis is that foundation model companies combine two historically durable moats: the application lock-in of operating systems and the capex moat of cloud infrastructure. Microsoft and Apple don't capture value through Windows or iOS themselves — they capture it through the verticalized applications built on top. Maguire believes the same dynamic plays out with AI, where xAI, OpenAI, and Anthropic sit at the OS layer and extract value both from the applications built on their models and from the sheer capital intensity of staying competitive. Open-source models, in his framing, follow the Linux trajectory — more installs, far less value capture. He points to Red Hat as a viable but smaller outcome, suggesting open-source AI will matter without dominating the economics.
Defense tech: ~10 winners
On defense, Maguire's unpopular take is that despite the category's importance, only roughly 10 companies across the US, Europe, and Israel will become genuinely large businesses, with market caps above $10 billion and several crossing $100 billion. The rest get rolled up or fail. His bull case for the winners rests on vertical integration and margin structure. The legacy primes — Boeing, Lockheed — subcontract almost everything, which means investors buying those stocks only ever captured a slice of the value chain. Next-generation defense companies that own their stack will look structurally closer to Tesla than to GM, a comparison the market got badly wrong for years.
On the $500 million earmarked for autonomous systems in the current defense budget, Maguire is less focused on the absolute number than on the trajectory. Consistent, predictable growth in allocation over a decade matters more than a single-year figure. Erratic funding — $500 million one year, zero the next — is what kills companies trying to build around government contracts. He learned this firsthand when his cybersecurity company, focused on defense, won a $10.5 million contract only to have government sequestration cut the payout to 60 cents on the dollar.
His read on which politicians actually understand the technology agenda points to Senator Todd Young, who he says distinguishes between strategically critical industries worth reshoring and lower-priority manufacturing that can sit with allied nations — a level of nuance he treats as rare in Washington.