Markets & Investing

WashU CIO details concentrated private market strategy and AI valuation risks

Scott Wilson has grown the WashU endowment to nearly $20 billion by prioritizing direct co-investments over traditional fund commitments. He warns that current private market valuations for AI companies are unsustainable compared to public market realities.

· 1 source · Endowment strategy

Scott Wilson, chief investment officer at Washington University in St. Louis, has expanded the university's endowment from just under $8 billion to nearly $20 billion over nine years, generating $17 billion in profits. Wilson restructured the portfolio upon his arrival, redeeming over 50% of assets in the first week and turning over more than 75% of the private portfolio via secondary sales over 18 months. The current allocation targets a 50/50 split between public and private assets, though Wilson noted that without the 5% annual payout constraint, he would target 70-80% private exposure.

The portfolio is highly concentrated, with half of assets held in the top 50 look-through positions. SpaceX represented north of 20% of the portfolio at one point, a position Wilson advocated holding through the IPO using continuation vehicles to balance liquidity needs with upside potential. Wilson argues that private markets are less efficient than public markets, allowing for greater alpha generation, but contends that current valuation premiums are unsustainable. He cited Anthropic and OpenAI, which could raise capital at valuations north of $2 trillion in private markets, but may not achieve similar levels in public markets, as evidenced by the post-IPO performance of Cerebras.

Wilson’s strategy relies heavily on direct co-investments rather than fund-of-funds structures. He stated that WashU does not invest in mega-funds north of $10 billion because they do not fit the co-investment model or require the university's capital. Instead, the endowment favors smaller funds where a single check can significantly impact returns. Wilson also expressed skepticism toward the business models of frontier AI providers, arguing that high commoditization of large language models will erode profit pools and lead to an industry shakeout.