Situational Awareness bets $400 million on chip startup Source Foundry

Embattled AI hedge fund Situational Awareness poured $400 million this week into Source Foundry, a Stanford-founded startup building chip-making equipment to challenge ASML's grip on lithography, the most complex step in semiconductor production.

"If you trace the AI supply chain upstream, from models to chips to the machines that manufacture them, each layer becomes increasingly critical and constrained," Stephanie Zhan, a partner at Sequoia Capital, said. "Source Foundry tackles the tightest bottleneck: tooling for semiconductor manufacturing, starting in lithography."

The fresh infusion brings Situational Awareness's total commitment to $500 million, according to people familiar with the matter. Source Foundry, founded last year by Stanford researchers Abdulmalik Obaid and Joe Burg, recently raised funds at a $5 billion valuation, with Sequoia among its backers. The startup is developing machines, equipment and software to produce AI chips faster and cheaper than ASML's extreme ultraviolet (EUV) systems, which cost more than $400 million each and take months to build and install.

The bet comes weeks after the fund, run by 24-year-old former OpenAI researcher Leopold Aschenbrenner, sold most of its public stock portfolio to Ken Griffin's Citadel at a discount after a 67 percent July loss triggered margin calls. Assets under management fell to about $10 billion from $20 billion. The investment shows Aschenbrenner's willingness to fund technically ambitious bets that could take years to reach commercial viability — and his conviction that chip tooling, not just chips themselves, is the next bottleneck in the AI supply chain.

The fund's pivot from leveraged public-market bets to long-term private stakes in the physical machinery behind AI marks a strategic reset. Aschenbrenner had told investors he would allocate up to 30 percent of the portfolio to private companies. He retained a multibillion-dollar stake in Anthropic, plus investments in cloud startup Fluidstack and chip startup MatX, after selling the bulk of his public holdings — including memory-chip makers SK Hynix and Sandisk, AI cloud provider Nebius Group and fuel-cell maker Bloom Energy — to Citadel. At one point the fund discussed selling $3.5 billion of its Anthropic stake to Sequoia and Greenoaks to cover margin calls, but backed out.

Earlier this year, Situational Awareness co-led a roughly $1 billion round with Lux Capital in Physical Intelligence, a startup building an AI model to command robots, at an $11.5 billion valuation. Founders Fund, Thrive Capital, Kleiner Perkins, Accel and Iconiq also participated.

The $400 million machine

ASML's EUV lithography systems — which use 13.5-nanometer light generated by firing lasers at droplets of molten tin — are indispensable for producing the world's most advanced chips. The Dutch company reported sales of 32.67 billion euros in 2025. Source Foundry is betting it can deliver a faster, cheaper alternative as demand for chip-making equipment surges. The startup has not disclosed how its technology differs from ASML's approach, and any viable alternative would need to match the precision and throughput required for high-volume production.

Semiconductors account for more than 95 percent of the value in an AI server rack, according to an analysis by the Semiconductor Industry Association and Deloitte. Global foundry revenue is projected to rise 24.8 percent to $218.8 billion in 2026, per TrendForce, with AI chip demand keeping advanced-node capacity at consistently high levels. Investment in AI data center infrastructure could exceed $4 trillion by 2028.

Who wins, who loses

Source Foundry is not alone in targeting the gap. Startups including Substrate, xLight and Inversion Semiconductor are also working to bring advanced lithography technology to the U.S. A breakthrough in chip-making equipment would benefit the entire semiconductor industry, not just one accelerator maker — but it would threaten ASML's dominance in a market where its EUV machines face no real competition.

For investors, the question is whether Source Foundry can translate research into production-ready technology. The fund's bet on a pre-revenue startup at a $5 billion valuation carries substantial risk, but it also positions Situational Awareness ahead of what could be the next major constraint in AI computing. Aschenbrenner's conviction that AI demand is growing on an exponential curve while semiconductor manufacturing capacity grows on a linear one — a mismatch Sequoia's Zhan calls the "chip wall" — now backs the machinery that makes the chips.

This article is for informational purposes only and does not constitute investment advice.