David Weisburd moderated a Global Alts New York 2026 panel on endowment and foundation strategies with Boris Raykin, newly appointed CIO of a major university endowment, Jon-Michael Consalvo, CIO of Carnegie Mellon, and Bruce MacDonald, CIO of VCU. The session opened with a direct question about the Yale model David Swensen built over three decades and whether it is still the right template for the current environment.

The Yale endowment model is twenty years past its optimal environment, and the CIOs on this panel are building something different. Raykin opened with the framework. The core tenets still hold: align with managers, find them early, look for alpha in inefficient markets. But the environment has changed enough that the model has to evolve in three directions. First, quantitative strategies that exploit compute and data at a speed no discretionary manager can match. Second, global macro that invests long and short across asset classes including fixed income and commodities. Third, portable alpha that recognizes US large caps are now too efficient to earn alpha through stock selection alone.

Why the Yale endowment model needs to evolve

MacDonald made the environmental argument. The Yale model was “perfectly designed for the environment that ensued, which was low interest rates and a strong equity market.” That environment has ended. Secular inflation, higher structural rates, and the crowding of private equity and venture capital have compressed the alpha that was available when Swensen first articulated the model. The portfolio that compounded at 16% for 20 years was built for a world that no longer exists. The right response is not to abandon the model but to adapt it to the world where diversification requires more creative thinking about what truly diversifies.

Consalvo added the AI-first manager framework. Carnegie Mellon looks for managers who are “on the cutting edge and willing to be out-competed by people who are starting with AI.” The question in manager selection is no longer just whether a manager has a good process. It is whether that process survives in a world where AI can do the top-of-funnel analysis that used to take a team of analysts six weeks.

How CIOs are managing the AI diversification challenge

Raykin walked through how Carnegie has mapped AI exposure in its private portfolio into three buckets: AI-first managers whose entire strategy depends on AI staying at the frontier, AI-experimental managers who are testing new tools without full commitment, and AI-insulated strategies where returns are achievable regardless of AI outcomes. About 30% of the private portfolio falls into the AI-insulated category, anchored by distressed asset-based lending. The exercise of mapping the portfolio this way exposed concentration risks that were invisible under traditional strategy classifications.

What the right endowment and foundation governance model looks like now

MacDonald closed on governance. The endowment and foundation model that works requires investment committees that understand the strategy, accept that the portfolio will look different from peers, and give the CIO team enough rope to run the process. “You cannot over-communicate. It is impossible.” The institutions that built that culture have a structural advantage because they can underwrite conviction positions that peer-pressure-sensitive investment committees will not approve. For allocators mapping endowment and foundation managers and strategies, Allocator Intelligence on iConnections surfaces the deep specialists the Yale model originally sought.