Scott Wilson

1. Key Themes

Concentrate Hard, Then Let Winners Compound

WashU's playbook is deliberately unclever: find great partners, and when they surface something unique, scale it aggressively. The result was SpaceX growing to an outsized share of the whole endowment, enabled by governance that allows concentration.

Scott Wilson describes the core approach: "We try and find really good external partners. When they find something that we think is really unique, we try and double, triple, quadruple down on it and then hopefully let that compound over long periods of time. That's it." 00:05:40

On SpaceX specifically: "At one point north of 20% of the portfolio is in SpaceX. And at that point it becomes like, we had conviction as a team, and we had many chances to take lots of liquidity in that name, but we advocated for holding it and letting that compound." 00:00:00 He credits the board's governance: "you can't do that without the right governance structure." 00:13:30

Direct and Co-Investments Are the Edge, Funds Alone Were Average

Wilson says the direct/co-invest capability is what separated WashU from peers, and it requires a radically different time allocation than most LPs have. His team spends the bulk of its time with companies rather than managers.

"Upwards of, you know, 80-90% of our time is just out meeting individual companies. And that's very different than probably how most of your LPs spend their time." 00:11:46 And on performance attribution: "If we didn't have any direct or co-investments, we would have been right at the average of our peer group. You know, as it is, I think we're at the top, if not the very top of our peer group over the last decade." 00:12:13

Offense Beats Defense: Proactive Sourcing Over Inbound Deal Flow

WashU treats its managers as a signal layer, then goes out to meet the underlying portfolio companies itself, rather than waiting for syndicated deals.

"We're not waiting for them to bring us a deal... 'Harry, like, who are the five companies in your portfolio that you're most excited about?' And then we'll just go out and meet those five companies." 00:24:37 He warns about adverse selection: "I actually think like there's a huge negative selection bias if you're just waiting for someone to give you a call and say, 'Hey, we have, you know, $500 million of OpenAI that we're syndicating out.'" 00:24:37 The biggest takeaway from hard lessons: "We've done much better on this like offensive strategy than defensive strategy." 00:25:47

Look-Through Exposure Is the Unit of Analysis

Rather than evaluating fund performance, WashU aggregates underlying holdings across all partners and manages exposure to individual companies.

"We hardly ever talk about like fund performance. We only really talk about the underlying exposures in the portfolio. So we aggregate everything up, and we know that we have X dollars of SpaceX or Cerebras or Anthropic in the portfolio." 00:00:00 The target: "Half of our portfolio is in our top 50 look-through positions." 00:00:00 He notes diversification benefits fade fast: "The quantitative benefits for diversification trail off really quickly when you get more than 20-30 names in the portfolio." 00:18:05

Small Funds Over Mega Funds

WashU's model structurally favors smaller funds where a small check can be a fund-returner and look-through concentration is achievable.

"We'd much rather have someone who raises, you know, a billion-dollar fund when they could raise 4 billion and comes back every 18 months than someone who raises 10 billion and comes back every 3 years." 00:36:48 On mega funds: "Our model doesn't really work with mega funds. They don't need our capital." 00:00:00 They are in funds above $1 billion but "not in any that are north of 10 billion." 00:22:54

Private Markets Are Overheated; Public Markets Can't Absorb the Valuation Premium

Wilson sees a valuation gap between late-stage private pricing and what public markets will pay, and argues that gap, not market closure, explains the lack of DPI.

"There's just too much capital chasing these kind of late-stage champions in the private markets. And the valuations are... I think that's why there's no DPI. It's not because the public markets aren't open. The public markets are pretty close to all-time highs. It's just you can't go public at the valuation that the private markets will give." 00:26:27 He cites weak post-IPO trading: "Cerebras... has traded really poorly since the IPO. SpaceX is basically back to really close to its IPO price." 00:26:27

AI Reckoning: Model Layer Commoditization and Uncertain Profit Pools

Wilson is skeptical that foundation model companies have defensible economics, and expects a shakeout.

"Probably half of these companies don't exist 10 years from now. And it's not clear how many have like a defensible moat around their business." 00:28:48 On commoditization: "If you take the top five kind of large language models and they all had the same user interface, for 99% of problems or tasks, like you probably couldn't tell the difference... I can't describe another industry like that that has a big kind of deep profit pool." 00:28:48 On willingness to pay: "When there's a free alternative, customer willingness to pay drops very, very quickly." 00:29:45

Valuation Discipline Is the Top Reason to Drop a Manager

When WashU churns a venture manager, the most common reason is not performance but price discipline.

Asked for the dominant reason: "Discipline around valuation." 00:35:21 He elaborates: "We had one partner that I'm not sure he found a 100 times revenue deal that he was willing to turn down... For us, that's not what we want our capital to be used for." 00:35:49

Price vs. Value: Europe and China as Mispriced Opportunity

Wilson frames non-US markets as a value opportunity where equal-quality assets trade at dramatically lower multiples.

"If they were based in the US and had the same kind of KPIs, growth, and unit economics, they would trade at huge revenue multiples. And in Europe, they trade at, you know, small earnings multiples." 00:46:35 On the principle: "Price is what you pay, value is what you get... over long periods of time, like what the core KPIs of the business will drive returns more than the original entry multiple." 00:00:00 On China, WashU has two people permanently in Shanghai and "probably more than our peers in China." 00:48:17

The Hardest Lesson: Not Adding to Winners and Not Selling a Bubble

Wilson's two biggest self-identified mistakes are behavioral: anchoring on cost basis, and being greedy at the 2021 peak.

"Our biggest mistake was not adding to our winners just because... 'Oh, we invested in this company at 2 billion and now it's 5 billion, so we didn't add to it.'... that was my mistake and it's cost us billions of dollars." 00:43:30

2. Contrarian Perspectives

Private Markets Should Be 70-80% of the Portfolio, Because Public Markets Are Where Alpha Is Hardest

Most institutions cap privates because of liquidity worries. Wilson says the only real constraint is the 5% payout, and without it he'd go heavily private, citing returns: "We've probably compounded our privates by in the low- to mid-20s, and we've compounded, you know, our public equities at 10%." 00:09:39 His belief: "I think it's much harder to create alpha in public markets than it is private markets." 00:10:28 He even reframes the distinction: "The only difference between those two asset classes is how often they're marked to market." 00:08:20

Fire-Selling 75% of an Inherited Portfolio, Including All 36-37 Hedge Funds

Rather than gradually transitioning, Wilson restructured fast: "We had 36, 37 hedge funds in the portfolio and we got rid of them all. So we certainly weren't popular in certain parts of Manhattan. Ultimately, we turned over more than 75% of the portfolio through secondary sales in the private side." 00:00:00 His rationale: the board gave a three-year window to show improvement, and "changing around an endowment portfolio is more like turning around an aircraft carrier than a speed boat." 00:06:38

Declining Sequoia's Evergreen Structure and Mega-Platform Fund Bundling

Wilson passed on Sequoia's evergreen structure because it didn't fit his co-invest model: "It didn't fit super well into our kind of co-investment model, so they weren't great partners for us at that point in time." 00:32:54 He admits it was likely a mistake on returns ("I'm sure it was a mistake if you look at the returns that they've generated" 00:33:48), but holds a firm line on platforms that bundle products: "We just won't do that. We'll just find some other place to put capital." 00:34:30

Concentration Limits at Venture Funds Are "Artificial"

Wilson pushes back on the standard 10% per-company limits: "I do. They're mostly artificial... to have a 10% concentration limit in any individual fund, I don't think it makes sense." 00:14:28 His logic is that it's only a small position at the endowment level: "If you're 1% of our portfolio or 2% of our portfolio and you have a 50% position, that's only 100 basis points for us." 00:14:28

The Frontier Model Layer Is a Weak Business, Yet He Holds Anthropic Exposure

Wilson holds "pretty decent exposure to Anthropic through a few of our partners" 00:29:31 while believing the model layer lacks defensibility and that Anthropic could have raised "north of 2 trillion" privately but may not get that publicly. 00:26:27 His reasoning for not selling is structural: "It's one exposure in a fund of 40 other investments, so we wouldn't be selling just one thing in isolation." 00:30:34

3. Companies Identified

SpaceX

Private space and launch company that went public. It was the defining win of the WashU portfolio. "At one point north of 20% of the portfolio is in SpaceX." 00:00:00 WashU used continuation vehicles to hold through the IPO. Wilson notes it is "basically back to really close to its IPO price" 00:26:27 and that WashU will "gradually sell out of it over, you know, a multi-year period." 00:32:04

Anthropic

AI model provider. WashU has "pretty decent exposure to Anthropic through a few of our partners." 00:29:31 Wilson says it "could have raised capital in the private markets... at north of 2 trillion if they wanted to" 00:26:27 but questions whether public markets will match that.

OpenAI

AI model provider, mentioned as a late-stage private champion whose valuation may not translate publicly: "Same with OpenAI, you know." 00:26:27 Also used as the example of syndicated secondary deals that carry negative selection risk: "$500 million of OpenAI that we're syndicating out." 00:24:37

Cerebras

AI chip company, accessed via Eclipse. WashU held "a really big position at the IPO, and so we took some chips off the table and put some hedges on." 00:38:30 Wilson likes the team but says "it's tough to justify the valuation, even at 25-, 30 billion where it's at today... you're betting on the come." 00:38:30

Nubank

Brazilian digital bank. WashU owned it as a private company and still holds it in the 13F nearly five years post-IPO: "It's in our 13F. We've owned that since it was a private company. It was distributed and we still own it." 00:31:15 Harry also called it amazing and said "David is a rock star." 00:32:04

Stripe

Mentioned as the example of a private company that should not change an asset's classification on going public: "It's not like if Stripe goes public tomorrow, it should be in a different asset class." 00:09:05 Harry noted the Collisons proved staying private works. 00:27:40

Benchmark

Venture firm. Wilson's answer to "What fund are you not in that you wish you were in?": "Benchmark, probably." 00:51:12

Founders Fund

Venture firm. Wilson wishes WashU "had like a bigger allocation to the early-stage Founders Fund as well." 00:51:19

Eclipse

Venture fund (Lior Susan) through which WashU got access to Cerebras. 00:51:29

Sequoia

Mentioned for the evergreen fund structure WashU declined, with Wilson acknowledging "they're great investors. They're great people. And they've obviously made LPs a ton of money over long periods of time." 00:32:54

Thrive Capital

Harry cites Thrive as having "absolutely crush it with OpenAI and Cursor" 00:21:02 as evidence of outcome expansion supporting large funds.

Cursor

AI coding product, referenced by Harry as a position that drove outsized outcomes for Thrive. 00:21:02

Crosby

AI-enabled law firm and a sponsor, which Harry says "turned redlines around in 3 hours" and combines AI with "some of the best attorneys in the world from top-10 law firms." 00:00:35

OneMind

GTM "superhumans" product used by HubSpot, Alteryx, and ZoomInfo, per the sponsor read. 00:00:35

Framer

AI website builder used by Perplexity and Miro, per the sponsor read. 00:00:35

Uber and Nvidia

Harry cites them as "comparatively very cheap companies based on a fundamentals basis, actually, today" relative to privates. 00:40:45

Grinnell College Endowment

Wilson's prior employer, which "ran a similar strategy" with direct co-investments and a "super-concentrated" portfolio. 00:04:58

Other mentioned entities

Ramp, Stripe (as engineers' origin), Cognition, and Clay appear in the Crosby ad as customers/talent sources. 00:00:35 Harry also references OpenRoad ("buy OpenRoad for $8 billion") as an example of private-company M&A capability. 00:27:40

4. People Identified

CIO of WashU's endowment, who took it from just under $8 billion to just under $20 billion. "In the 9 years I've been here... we've generated $17 billion of profits on that original 8 billion, and we've paid out roughly 5 billion to the university." 00:42:11

David Clay

Wilson's former boss at Grinnell, credited as the person he learned from: "David was a great guy to learn from, because when I first landed at Grinnell, I really had no idea what an endowment actually did." 00:04:58

Lee Fixel

Investor (Addition), a mutual friend of Harry and Wilson, and Wilson's pick for best investor in the portfolio: "I think he's the best investor in our portfolio. He's just very pragmatic. He's disciplined around price, but will pay up for super high-quality assets. And I think his underwriting process and just the way he dissects companies and business models and management teams is really impressive." 00:51:42 Wilson also credited "Lee and the board and Adam" for the SpaceX win. 00:13:30

Andrew Feldman

Cerebras CEO, referred to as "Andrew": "We like Andrew and the team... we have lots of confidence in Andrew and his team to produce good results over the next decade." 00:38:30

Lior Susan

Founder of Eclipse, the fund that gave WashU access to Cerebras. 00:51:29

Charlie Munger

Quoted via Wilson on concentration: "Charlie Munger would say, 'If you have three, you're fine.'" 00:18:00

David (Nubank)

Referred to by Harry as "a rock star" in connection with Nubank. 00:32:04

Elon Musk

Wilson's first pick for a board member he'd add for "real industry expertise." 00:52:16

Thomas Reardon

Also named for a board seat on the basis of deep industry knowledge to help "diligence a variety of investments." 00:52:16

Fellow endowment CIOs

Wilson names peers whose multi-decade performance he respects: "what Erik's accomplished at Michigan, what Robert's accomplished at Virginia, Seth at MIT, Jane, now Josh, at Brown. Like those teams have just done unbelievably well over the last 20 years." 00:52:40

David Morehead

Baylor endowment leader, whom Harry referenced as having laughed at the idea of investing in Europe. 00:46:12

The Collison brothers

Stripe founders, cited by Harry as proof that staying private "and having a great business is very beneficial." 00:27:40

Jamie Dimon

Referenced for calling China "uninvestable," a statement WashU has invested against. 00:48:17

5. Operating Insights

Build Governance That Permits Concentration Before You Need It

The SpaceX outcome was possible because the board structure allowed it. Wilson explains: "At cost, you know, if we were going to put more than 2 or 3% of the portfolio in a single name, we'd probably... need to notify the board. I think it's 4%, we need their approval. But we have a lot of leeway to concentrate the portfolio quite a bit." 00:13:30 The tactic is to pre-negotiate tiered thresholds (notify, then approve), so conviction positions don't require re-litigating your mandate each time.

Test Any New Position Against "Would I Do 1% of the Portfolio?"

Wilson uses a simple conviction filter: "I always tell, if we don't have enough conviction to do 1% of the portfolio, then we probably don't have enough conviction to do it at all." 00:19:10 That is close to $200 million at WashU's scale, so the approach is to start with a smaller check (e.g., $50 million) and scale it as conviction builds.

Kill the Cost-Basis Anchor When Deciding Whether to Add to a Position

Wilson's costliest error was anchoring on how a position grew: "We had times where we could have invested in stuff that I think was still probably our highest-conviction investment, and we didn't do it either because it had grown in the portfolio or the price had just gone up." 00:43:30 The operating rule is to evaluate follow-on capital as a fresh decision based on forward expected return, not on position size relative to the original check.

Underwrite Using the Five-Part Direct Investment Filter

Wilson's direct-investment bar: "We have to be able to independently verify the thesis, understand the thesis, have the ability to underwrite the company, ability to monitor it going forward, and we have to think it's our best marginal use of capital." 00:16:17 Pass rate: "more like 0.1% of those investments than 1%." 00:16:17 Use "can I distinguish this company from five competitors?" as a gating question.

Sell When the Gain Is Real, Not When It's Perfect

Wilson's biggest regret on timing: waiting for Q1 write-ups in 2021 when a large secondary sale could have cleared at "$0.80-$0.90 on the dollar." 00:45:33 His takeaway: don't hold out for an incremental mark when you believe the market is a bubble. His general principle on size: "It's better to be not stupid than clever." 00:32:04

6. Overlooked Insights

Funds-Only Performance Was Merely Average, Which Means the Manager-Selection Layer Isn't the Alpha

In passing, Wilson states that without directs and co-investments WashU "would have been right at the average of our peer group." 00:12:13 That implies a top-tier allocator with elite managers still earned average returns from the fund layer, and that the edge came entirely from selectively deploying more capital behind the best underlying companies. The significance for fund managers is that their LPs' value-add is often in co-invest allocation, which is a negotiated, not default, part of the relationship. For allocators, it argues that adding more fund relationships is the wrong response to the scaling problem: "each manager means less and less." 00:17:13

Payout Rules Make Illiquidity a Compounding Constraint, Not Just a Timing Issue

Wilson's math is easy to miss: "If you have a 5% payout on and half your portfolio's liquid, that's 10% of your liquid portfolio that you're paying out every year." 00:09:39 That means the effective drawdown on liquid assets doubles as privates grow, which forces forced sales or limits private allocation regardless of conviction. It also explains why WashU is at 74% private (above its 50% target) and currently redeploying liquidity into public markets, specifically "non-AI, AI infrastructure-related publics" in "sectors completely that we think are relatively protected, if not immune from AI." 00:41:09 The real insight is that for any payout-constrained institution, private allocation capacity is a function of the spending rate and liquid sleeve, so a lower payout would translate directly to a materially higher return profile.