Elite universities that made early, targeted investments in artificial intelligence and cryptocurrencies are now seeing strong returns, according to new research from Markov Processes International.  

With strong public markets and a rebound in venture capital, Ivy League schools posted returns of around 11 to 12 percent. MPI found that exposure to AI and digital assets — via venture capital funds and direct stakes — boosted performance by up to 300 basis points. 

Michael Markov, MPI’s founder and CEO, told Institutional Investor that for the first time, these investments are showing up as key drivers of overall performance in data from MPI’s Transparency Lab. “It is plausible that AI and digital-asset exposure is now large enough to affect overall results,” he said, noting implications for liquidity stress and governance. 

Top performers included the University of Michigan (15.5 percent), M.I.T. (14.8 percent), and Stanford (14.3 percent), all of which had substantial AI and crypto holdings. MPI estimates Michigan’s digital asset exposure added 2.9 percent to its return, while AI-related investments contributed 2.8 percent. 

Brown, Cornell, and Harvard also had significant crypto exposure, while Princeton, Penn, and Yale held notable AI-related assets, according to MPI. 

Although endowments rarely disclose direct crypto holdings, many have been limited partners in crypto venture funds since at least 2018. MPI’s analysis suggests Brown, Harvard, M.I.T., and Stanford have steadily increased their crypto allocations. Cornell holds equity in crypto firm Ava Labs. 

In AI, Michigan invested in OpenAI and is an LP in Alt Capital’s AI-focused venture fund. Some gains are tied to late-stage venture valuations, meaning they’re unrealized and based on paper value rather than actual sales. 

The surge in AI and digital assets — up over 50 percent in the past two years — has been sufficient to lift endowment returns by several percentage points, provided allocations are meaningful. Markov emphasized that these assets have evolved from niche bets to significant holdings, bringing new challenges in liquidity, governance, and optics. He also cautioned that many AI-related gains are unrealized.  

The research states that when a diversified $30 billion endowment “is 200-300 bps ahead of peers” in FY25, “it is unlikely to be generic ‘manager selection,” but instead “looks consistent with targeted exposure to AI and/or digital assets or other themes that produced outsized gains in FY25, plus some residual manager-specific alpha.”