20VC: How LPs Allocate to Venture in 2026: What They Want, What They Do Not Want | Why Fund Multiple Does Not Matter Without a Timeline | Why Velocity of Cashback is the Most Important Thing with David Morehead, CIO @ Baylor · podgraph
20VC: How LPs Allocate to Venture in 2026: What They Want, What They Do Not Want | Why Fund Multiple Does Not Matter Without a Timeline | Why Velocity of Cashback is the Most Important Thing with David Morehead, CIO @ Baylor
David Morehead, CIO of Baylor's endowment, says funds stretching to fifteen to eighteen years misalign with endowment math and should compound capital faster through shorter funds.
Baylor allocates about 45 percent to private assets within a 35 to 55 percent range, reserving the upper end for market drawdowns and focusing on venture, growth equity and buyout.
Morehead says Baylor bought software in February and March while it was down 50 to 60 percent, adding capital in roughly 10 percent decline increments so it never goes all in.
Baylor's CIO says velocity of capital beats fund multiples
Baylor runs its endowment around how fast capital comes back, not only how large the multiple is. Morehead argues that fifteen to eighteen year funds do not fit endowment math, so Baylor uses separate accounts, conservative marks and staged buying to control timing and risk. The results so far are a 9.4 percent return in 2025 and an expected 18.5 to 19 percent this year.
Takeaways
Returns mean little without a time dimension. Morehead says Baylor wants velocity of capital, not just multiples. Venture funds have stretched from ten to twelve years to fifteen to eighteen, which he says misaligns with endowment math. Compounding capital through shorter funds would produce more money for students.
Private assets exist only to make money. Baylor holds about 45 percent in privates within a 35 to 55 percent range, with the top reserved for drawdowns. It wound down real assets for venture, growth equity and buyout. Growth equity is largest and annualizes around 30 percent against an eight to nine percent target.
Separate accounts give control over overlap and concentration. Baylor asks general partners to run strategies just for it, for example declining more Nvidia or asking for three times as much. This has worked well over two or three years. A manager who changes strategy without a conversation would not be re-upped.
Baylor bought the software sell-off. It allocated in February and March with software down 50 to 60 percent, adding capital in stages so it is never all in. Morehead held four weeks of daily calls pushing manager Sean Barris toward concentration, arguing vertical software will deliver AI.
Public marks carry more information than private ones. Morehead says tens of millions of traders price public markets, while a few people set private prices. Baylor pushes managers toward conservative valuations, citing pre-exit marks averaging 60 to 90 percent gains versus 30 to 50 percent for the market.
Venture is a diversifier with concentrated AI exposure. Asked by Stebbings why an endowment does venture at all, Morehead points to laddered returns. Baylor holds about 2.5 percent of its endowment in Anthropic and has no exposure to SpaceX or OpenAI.
Permitted, powered land is the scarce data center asset. Local permitting pushback on AI data centers, now seen internationally, makes permitted and powered land valuable. Baylor's data center sites are up 50 percent in six months. Morehead also prefers equity to private credit, which he says has equity-like downside without the upside.
Tap a name to go there · tap a line to see why they are linked