20VC：LP在2026年如何配置风险投资：他们想要什么、不想要什么｜没有时间线，基金倍数为何毫无意义｜为什么现金回收速度最重要——Baylor首席投资官 David Morehead
- Morehead的核心观点是：没有时间线，基金倍数就没有意义——捐赠基金真正需要的是资本周转速度。 一只VC基金用15–18年实现15X回报，可能还不如连续3只、每只6年实现3X回报的成长型股权基金，后者复利后达到27X——“比15X好得多，大概是2倍”。他的办公室原则是：“谈回报时，不允许不谈时间”，因为“如果30年赚了5X，那很糟糕；如果5个月赚了5X，那就太棒了。我想这就是SpaceX。”
- “私募资产存在的唯一理由就是赚钱，仅此而已，故事结束。” 因此，Baylor正在逐步退出实物资产，把配置集中到VC、扩张型/成长型股权和收购型投资。成长型股权是最大的私募配置，年化回报约30%，对比基准为8–9%；相比之下，VC对Baylor而言“纯粹是分散化配置”。Baylor捐赠基金约2.5%通过管理人持有Anthropic，没有OpenAI或SpaceX敞口。
- 2025年10月至2026年初软件板块下跌50–60%时，Baylor开始基于对人类行为的判断，而非技术判断，增加软件配置。 Morehead打电话给经营500人规模私营企业的朋友，问他们是否会用vibe-coded应用替换CRM——答案是“绝不可能”；他还不太确定地回忆，Salesforce CEO曾说最好的AI会达到“93%的正确率”，但“软件的问题在于必须100%正确”。他认为，在部分垂直行业，值得信赖的既有软件可能成为“AI的交付机制”。
- 风险纪律是：“我永远不想满仓，事情总是可能变得更糟。” Baylor以市场每下跌10个百分点为机械配置节点：下跌20%时投入约20%的资金，下跌30%时再投入20%；接受在底部前无法满仓、从而错过部分收益，以避免在市场触底前就用尽资金。现金的定价是8.5%：3.5%的收益率加5%的机会成本，依据是未来4年内找到一个20%机会的概率很高。疫情前现金占比为15–16%，目前处于低位。
- AI基础设施的可交易判断是：稀缺资产已经从土地，迁移到有电土地，再迁移到“获批的有电土地”；地方政府和居民对审批的反弹已成为新的瓶颈——“6个月前还不存在”。 Baylor账面上的数据中心用地6个月上涨50%；英国一个项目“仅仅因为我们有许可证”就很有价值；未来5–7年，在供给约束解决前，电价可能继续上涨。他整体看空欧洲——“国防……俄罗斯……在AI上落后，因为、因为、因为”——并在欧洲指数上配置宏观对冲。
- 管理人纪律遵循棒球总经理的类比：无论回报如何，风格漂移都会让你被解雇。 “如果我走上场，发现队里有2个二垒手、却没有三垒手，那三垒手就得被解雇……我不在乎你的回报是多少。”不过，真正的策略切换与人为划分的类别边界需要区分：从投后产品市场匹配阶段的公司转向“车库里的2个人”，就属于策略切换。仓位规模从每家公司应投入的金额出发：扩张型股权和收购型投资中，每个底层公司配置250万–300万美元，确保5X回报真正有意义；VC则有所不同。
- Baylor在资产配置上投入的时间多于管理人筛选：私募目标占比为45%，在35–55%的区间内运行，确保分母效应不会迫使组合卖出资产（“第一要避免的是欺诈，第二要避免的是被迫卖出”）。 Baylor通过single-manager基金补充混合型基金，因此可以把对NVIDIA等单一公司的敞口调高或调低。2025全年回报为9.4%，低于Dartmouth的10.8%，原因是2020年、2021年及随后几年私募承诺额提高了60–70%，带来了第2条J曲线；与此同时，一只single-manager基金和另一类资产刚开始向上拐点。Morehead预计今年回报为18.5–19%。捐赠基金规模从14亿美元增至27亿美元。
1. 人口趋势是一场缓慢驶来的灾难，并决定Baylor如何投资
- Morehead的出发点是结构性的：全球金融危机后，美国高中毕业生人数减少，加上签证政策摩擦影响全额付费的国际学生，意味着“全国很多学校都没有达到”2030届的招生目标；因此，未来10–15年，捐赠基金分配必须填补收入缺口。Baylor大约5年前开始围绕这一趋势重组投资体系。
- 办公室过去的优势在于控制下行风险：2026年Q1，在标普下跌4%的情况下，Baylor持平；2018年Q4、2016年Q1以及2012年也有类似的相对收益。过去5年的项目，是补上另一条尾部：“市场70%的时间都在上涨。如果你在上涨行情中落后，那会很成问题。”他向金融学教授描述自己试图同时赢下上涨和下跌两端时，“他们真的会笑我”。
- 机制是single-manager基金。混合型基金提供的是让100个或1,000个LP都能接受的“平均风险收益特征”；Baylor则要求GP在独立账户中执行同一策略，并让Baylor看到账面组合。这样，当下一位管理人想买入NVIDIA时，Baylor可以说“我们已经有很多了”，或者说“把仓位做大到3倍”。“过去2、3年，这套机制的效果极其好。”
2. 先框定私募资产，区间的存在就是为了避免被迫卖出
- Baylor在组合构建上花的时间多于管理人筛选，“这在业内很独特”。目前组合大约为45–47%私募、53–55%公开市场资产。私募决策必须先做，因为“私募那一侧会吞噬你的流动性，束缚你的配置能力”；先把它设定好、框住，然后接受“私募组合真的非常难调整”。
- 围绕45%目标设定35–55%的区间，是为了确保分母效应不会迫使组合清算：“第一要避免的是欺诈，第二要避免的是被迫卖出。这是一场灾难。”2022年最后一段时间，科技板块下跌，私募占比升至约51–52%，虽然令人不舒服，但从未真正限制配置。
3. 资本周转速度：为什么15X可能是错误答案
- Morehead最核心的批评是：基金期限已经从10–12年拉长到15–18年，“这让所有LP都很恼火”；而且“我不确定GP的激励是否与捐赠基金运行的数学规律一致”。一只基金用15–18年实现15X，可能不如把资金连续投入3只6年期、每只实现3X的成长型基金，最终达到27X。“学生不能用回报支付学费，他们必须用美元支付。”
- 他完全理解GP为什么要长期持有赢家：6X在营销材料里比3X好看，而且“这意味着下一只基金可能会被募集”。但“我优化的不是对GP最有利的生意，而是我们学生能拿到的最大资金池”。当资本周转速度“开始渐近地接近它最终能达到的水平”时，他希望继续向前配置。
- Stebbings直接追问：既然成长型股权可以在6年内实现3X，为什么还要做VC？Morehead坦承，这个问题在内部一直有争论，但不同回收期限的梯次配置很重要：部分回报在6–10年到账，部分在3–5年到账，还有部分在1–3年到账。办公室的原则是，谈回报时永远不能不附带时间维度。
4. VC负责分散化，成长型股权才是发动机
- 当被直接问到VC是否“只是纯粹的分散化配置”时，Morehead回答：“对我们来说，是。”他同时指出，Baylor约2.5%的捐赠基金通过管理人投向Anthropic，没有SpaceX或OpenAI敞口。他明确把功劳归给管理人，而不是自己。
- Baylor进入那些顶级品牌较晚——“你敲门时，他们基本不会开门”——因此VC配置集中在“更新、更偏初创的新名字”上；与此同时，“我们办公室的女士们在扩张型/成长型股权上取得了非常、非常出色的回报”，年化约30%，对比基准为8–9%。成长型股权在避免归零方面也更有优势：“如果归零的项目更少，其他项目就不需要替那些失败项目补洞。”
- 对于2021年和2022年的糟糕基金年份，他的态度是：“这就是这个行业的一部分。”Baylor先在PE、扩张资本和VC之间设定配置，再看整个组合能否超过预期回报门槛，而不是孤立地评判某一个年份。
5. 软件交易：人类行为胜过工程知识
- Morehead认为自己的优势不在技术：“硅谷出来的很多东西我都看不懂，但我知道人是怎么想的。”面对“软件已经死了，有人会用vibe coding替代它”的叙事，他打电话给经营500人规模私营企业的朋友——其中包括一家他认为可能是全球唯一垂直一体化的百花香制造商——问他们是否会因为一个未经验证的产品而拆掉CRM，答案是：“绝不可能。”他不太确定地回忆，Salesforce CEO曾说最好的AI会达到“93%的正确率……但软件的问题在于必须100%正确”。
- 由此形成的判断是，在部分垂直行业，值得信赖的既有软件可能成为“AI的交付机制”。市值200亿–500亿美元的SaaS公司“并不愚蠢”，它们不太可能简单地眼看着现有软件归零。
- 2025年10月以来，软件“打了50%、60%的折扣”，Morehead的应对方式是打电话给企业，确认软件颠覆论并不适用于它们的运营，然后说：“那我就持有它。”
- 执行由管理人Sean Barrett负责：连续4周每天通话，不分时段交换文章，Morehead则推动组合集中：“你有这家公司，还有另一家公司……哪一个的风险调整后机会更好？”两人的分工是：“我基于人类行为做决策……但我依赖管理人成为各自领域的专家。”这就是配置人的工作，类似Buffett和Munger：决定增量的1美元交给谁。
6. 永不满仓：机械买入与8.5%的现金价格
- 从过往经验中留下的教训是：“每次交易，你肯定都会亏钱……有时会亏很长一段时间……结论就是，我永远不想满仓。事情总是可能变得更糟。”即便是这次软件交易，也没有一条明确的底线：已经下跌50–60%，但“谁说它不会跌70%、80%？”
- 机制是按10个百分点分批投入：下跌0–10%属于无限存续期组合的“正常波动”；下跌20%时投入约20%的资金；下跌30%时再投入20%。“现实是，在市场反弹前，我们实际上从来没有完全投资进去……我们会把一部分钱留在桌面上。这是真的。”换来的好处是，在下跌延续时永远不会完全押满。
- 现金被明确计价：未来4年内找到一个20%机会的概率“非常高”，因此现金的收益由3.5%的收益率加5%的机会成本构成，合计8.5%。疫情前，由于找不到有吸引力的机会，Baylor持有15–16%的现金；如今现金余额很低，“因为我们一直能找到年化20%、30%的事情去做”。
- 在市场下跌时，Baylor没有集中度风险。“我们什么都持有，从防晒霜到氦气，再到科技……我们的分散程度远高于标普500。根本不是一个量级。”
7. 公开市场价格合理，私募估值应保持保守
- Stebbings追问公开市场的“赌场化”——SpaceX在Elon溢价推动下达到1.8万亿美元——Morehead承认市场可能不理性，但不接受其因此失去有效性：“有数千万人在根据这些信息交易，而私募市场大概只有3个人……这不代表他们是对的，只代表价格纳入了所有可获得的信息。”这场交锋也说明了私募估值的问题：只因为几个人说自己试过产品、觉得不错，估值就被重新设定。
- 他过去管理交易账簿的经历塑造了估值纪律：错误定价会扭曲心理——一项被标记为3,000万美元、实际只值1,000万美元的仓位，会让你拒绝一份2,000万美元的溢价收购报价。Morehead认为Baylor估值保守的证据是：在被收购前6–9个月，Baylor账面上的平均收益约为60–90%，而他认为市场普遍水平是30–50%。
- 至于把VC当作学习学院，他的回答是：“对我来说不是。我其实从公开市场管理人那里学到很多。”他以2016年的自动驾驶热潮为例：“10年过去了，我们有什么？路上大概5万辆车？拜托。”
8. 棒球总经理：风格漂移会被解雇，按公司金额确定仓位
- 风格漂移原则原话是：“如果我走上场，发现队里有2个二垒手、却没有三垒手，那三垒手就得被解雇，彻底解雇……我不在乎你的回报是多少。”一个原本满仓股票的管理人，如果突然持有10%现金，也会被解雇——“我不想拿自己的钱去当未经验证的宏观判断的小白鼠。”但Morehead并不打击“人为制造的类别限制”：从产品市场匹配后的公司转向“车库里的2个人”，属于应被解雇的策略变化；从B轮转向后期A轮，“谁在乎”。Stebbings原本预期双方会有分歧，最后承认：“我们其实完全一致。”
- 仓位规模从捐赠基金真正关心的结果出发：一个管理人卖掉一家7X公司，却只为Baylor带来40万美元回报，得到的反应是：“什么？谁在乎。”因此，Baylor如今在扩张型股权和收购型投资中，目标是每家底层公司投入250万–300万美元；投10家公司意味着承诺3,000万美元，5X回报就是1,500万美元：“这已经足够重要。”VC有所不同，因为其底层公司数量更多。
- 谈到超大规模配置者，Notre Dame目前规模为200亿美元；他们原本以为在100亿–150亿美元时会遇到增长墙，但那道墙并未出现。不过，在接近Harvard或UTIMCO的规模之前，2,000万美元的支票如果变成10亿美元，也只占组合的2%，这正是a16z式平台模式正在利用的空间。至于大型VC平台本身：“只会越来越难……这就是大数定律。”
9. 获批的有电土地、看空欧洲，以及快速问答
- AI带来的反弹已经传导到土地层面：数据中心建在“我生活的地方，而不是硅谷”；稀缺资产从土地变成有电土地，再变成获批的有电土地。德州、亚利桑那等干旱地区的居民因电价和水价上涨而竖起抗议标牌，地方审批委员会为了连任而投票否决项目。由于足够多的项目无法落地，“电力公司开始找上那些已经拿到许可证的人，说：‘我们可以比原先预计的更早给你供电。’”Baylor账面上的数据中心用地6个月上涨50%；英国一个项目“仅仅因为我们有许可证”就很有价值。Morehead说，中国没有面临同样的流程，因为当地政府“需要建在哪里，就直接建在哪里”。
- 对欧洲，他的态度明确不乐观：“它的防务结构……俄罗斯……在AI上落后，因为、因为、因为。”但Baylor仍配置欧洲多空管理人，正是因为市场中会有赢家和输家；同时，Baylor“在欧洲指数上配置了一些规模更大的宏观对冲”。
- 快速问答：在软件上加仓；美伊冲突和霍尔木兹海峡局势导致原油涨破100美元时，降低能源多头；在3月或4月增加私募股权管理人的配置；他认为最被高估的资产类别是私募信贷——“下行时看起来、表现得都像股票一样的信贷敞口，却没有股票上行的回报”。最欣赏的同行是Brown和Jane的团队——“真正的投资人……敢于做出需要巨大勇气的事情”。最想投资的基金是Benchmark。未来10年的兴奋点是生物科技“解决疾病，而不仅仅是治疗症状”，以及应对办公室从10亿美元迈向50亿美元的拐点。2025全年回报为9.4%，低于Dartmouth的10.8%，原因是2020年、2021年及随后几年私募承诺额提高了60–70%，带来了第2条J曲线；与此同时，single-manager基金和另一类资产刚开始向上拐点。“今年我们的回报会达到18.5%、19%，而且没有任何SpaceX或Cerberus。”
完整逐字稿
The single reason that privates exist is to make money—period, end of story. I'm a little perplexed by the length of some of these funds. It's not clear to me that GP incentives are aligned with the math that runs endowments. What we're really after is the velocity of capital, not just returns on capital. There's a rule in our office: you're not allowed to talk about returns without also talking about time.
We happen to have about 2.5% of the endowment in Anthropic. I never want to be all in. Things can always get worse. You are seeing the pushback on AI at the data center level.
I'm a venture investor for a living, and something that's frustrated me for a long time is that we don't get to hear from the greatest CIOs—chief investment officers—who invest in the venture funds that we run. We don't know how they think, what they like to invest in, what worries them when they're invested in a manager and see what they're doing, how they think about the market today, or how they think about allocating to managers. Today, I sit down with one of the best in that business, David Morehead. He's the CIO of Baylor University Office of Investments, and he's one of the most respected CIOs in the business.
Baylor's endowment is around $2.6 billion. David is quite outspoken, which makes this conversation one of the most refreshing, but also articulate and clear, for managers thinking about raising and for managers now wondering how they should operate with their LP base. David was incredible, and I'm really proud of this show because it shines a light on a part of the industry that I feel needs a lot more transparency.
David, I am so excited for this. I've done so much stalking over the last 24 hours, it's untrue. Thank you so much for joining me today. This will be a lot of fun.
I would love to start with an overview of Baylor and how you think about investing today from Baylor as an institution.
1. Baylor Faces Higher Ed Pressure
It's a pretty important job, particularly in the place that we are with higher ed. We obviously have fewer high school students in the U.S. coming out of the Global Financial Crisis. As the number of high school students declines, that's obviously fewer tuition dollars.
The other thing that we have going on is that, over the last couple of years, it's been more difficult for international students to come over to the States, get appropriate visas, stay, and so on. Those are full-pay students, obviously. That compresses higher ed financial books in a different way.
Collectively, there's a lot of competition for domestic students these days. If you look at this incoming class—I guess it would be the class of 2030—there are a lot of schools across the country that did not meet their targets for the incoming student class. What that means, of course, is that the revenue has to come from somewhere else.
In this time and space, and realistically for the next 10 or 15 years, the distributions coming off endowment funds are going to be increasingly important. We manage with that in mind.
We've always been good at the downside. Historically, our office—in the first quarter of 2026, I think the S&P was down 4%, and we were flat. If you go back over time and look at the fourth quarter of 2018, the first quarter of 2016, and 2012, our office in general tends to outperform on the downside.
What we've gone back and looked at is how we could get better at the upside. We started this about 5 years ago, knowing that this high school student issue was going to be a problem. We've reorganized things over the last 5 years to make sure that we're doing better on the right side of the distribution.
Is it possible to do both?
I've had finance faculty laugh at me when I say what we're trying to do, so I'll let you be the judge of that. Effectively, what we're doing is running a value-centric, high-quality book, particularly on the equity side, because it's really, really hard to control the equity beta. You could buy puts, but that's a money-losing effort over long periods of time. We try to do it thematically through factor allocations.
That means, of course, that to the upside, when you're in a momentum-driven market or a growth-led market, you're going to trail. The issue is that the market is up 70% of the time. If you're going to trail to the upside, that's going to be problematic.
What we've tried to do over the last 3 to 5 years is increasingly solve that with convexity. We've tried to do that in a manner such that we're not paying a theta bill on a normalized basis.
I have to ask: before we move to theta, what do you mean by “increasingly solve that with convexity”?
What we've done is that, typically, higher ed outsources the investment of the endowment to a whole set of different managers. By nature, we have a bunch of investments in a number of commingled funds.
Commingled funds, by definition, mean that there's 1 GP managing the money, and then there are 100 or 1,000 LPs receiving the returns on that money. The issue with commingled funds is that, at any given point in time, you're receiving the average risk-return profile that the manager is providing in order to keep all of those LPs satisfied.
At any given point in time, Baylor's risk-return needs might vary from what the average LP in that fund would desire. What we've tried to do is go directly to the GP and say, “This commingled thing isn't totally working for us. We need to optimize our risk-return profile better. If we give you a bunch of money, would you run the same strategy but do it just for us?” Then we have a look or a call into what's going into the portfolio.
Let me give an example. Say we have NVIDIA in the portfolio. Then the next marginal manager wants to add NVIDIA to their portfolio. The GP doesn't know that. We can look at our portfolio and say, “We've got plenty of NVIDIA. We don't need more NVIDIA.”
Or conversely, let's say we're value, high quality. The next marginal manager wants to add NVIDIA, and we say, “We actually don't have any of that. We'd take the NVIDIA that you're offering, but why don't you make it 3 times as big? That's what we need to back into a more appropriate, more optimized risk-return profile for our portfolio.”
It's worked exceedingly well over the last 2 or 3 years.
2. Liquidity Comes Before Manager Selection
Can I ask, just taking a step up, when you think about portfolio construction today, you have a blank canvas. How do you at Baylor think about portfolio construction today? What does that blend look like—publics, privates, credit, debt, venture, PE?
It's interesting. We spend a lot of time talking about this. In fact, I think we probably spend more time talking about this than we do manager selection, which is unique in the space.
Presently, we're around 45% to 47% private and 53% to 55% public. I think it's really important that, if you're starting with a blank sheet and you're going to do privates, you really need to nail down the private side first. The private side is going to suck your liquidity and hogtie your ability to allocate between managers or between strategies.
And so you really need to figure out what sort of liquidity environment you can live with on the private side and determine what that allocation is going to be. Then I think you need to box it, set it aside, and say, “This is what’s going to be operating here.” The reason you have to do that is because this can’t change, right?
I mean, you can do secondaries and tweak it at the margin, but it’s really, really hard to move a private book around.
What would your answer be for what sort of liquidity profile you thought you needed when you were considering this?
Our allocation range around privates is 35% to 55%, which means that we want 55% to be the case when we have a denominator issue, right? When equities have gone down and the public side is smaller than it usually is, because of this particular difficulty in the market, the private side isn’t going to increase so much that we’re going to be forced into selling. The number one thing to avoid is fraud, and the number two thing to avoid is forced selling. That’s a disaster.
We target 45%. If the public markets race ahead, then it puts some downward pressure on that. If we get into a financial crisis or something like that, it would put upward pressure on that. For example, in the last bit of 2022, when tech slid a bunch, or if you go back a couple of years prior to the pandemic, I think our private side got to around 51% or 52%, but it wasn’t so much that it constrained our ability to allocate, and it certainly wasn’t enough that we got into a forced-selling situation.
When you think about the 45% that we have as the ideal, drilling one layer lower, how do you think about splitting that up between venture, private equity, and every other private investment that we can do?
We’ve had a different perspective on this over the last 5 or 6 years that really came out of what I was talking about before, when we knew that the school was going to have issues as they related to enrollment, right? It’s not just a Baylor thing, but every school. Demographics can be a slow-moving train wreck, but the benefit of the slow movement is that we can sit back, look 5 years out, and know what’s going to happen.
We started this 5 or 6 years ago, shortly after the pandemic, and basically said the single reason that privates exist is to make money, period, end of story. Anything in the private book that isn’t going to lend itself to excess returns—we need to create money to create more distributions for the school, which is going to have enrollment concerns. If you’re not going to keep up with the highest returns that we can generate out of the private book, we’ve moved on from that. A lot of the real-asset stuff in our book is winding down and not being renewed.
Mm-hmm.
To get back to your question, today we’re focusing on VC, expansion and growth equity capital, and buyout. That’s about it, right? If we’re going to lock up money, we want the highest returns.
How do you think about trying to get into the big names—the Sequoias, the Benchmarks, the Founders Fund, you name those big brands—versus trying to find the young upstart, the little boutique provider that could do a 10x?
I will say that we’re coming along a little bit later to the party than some of the Ivy League or Stanford, or what have you, as it relates to the VC brand names that you’re talking about. It hasn’t been for lack of trying; it’s just that when you knock on the door, they kind of don’t answer, right? So we’ve had to try to figure that out differently.
What I will say, though, is that the ladies in our office have had absolutely exceptional returns out of the expansion and growth equity category. We’ve actually had some questions like, “Should we just allocate more dollars to that sector of the market?” At the margin, we have, but I would say we still do VC. It’s still probably in newer, upstart-y names.
David, do you like VC?
I do. I’m a little perplexed by the length of some of these funds. I’ve got to be honest: it’s not clear to me that the GP incentives are aligned with the math that runs endowments.
What does that mean?
Let’s just use an example. Historically, they were 10- or 12-year funds. Now they’re 15- or 18-year funds, much to the chagrin of all LPs. The issue that you run into is that you get your money back in 15 or 18 years, and let’s just say it was a phenomenal experience and you’re up 15x. You’re like, “That’s fantastic.”
But the issue is that it happened over 15 to 18 years. Simple math would suggest that if you were in a growth equity fund that was 6 years in weighted average life and you were up 3x, then you redeployed into another growth equity fund that was up 3x in 6 years, and then you did it again, over the course of 18 years you’d be up 27x, which is better than 15x by a factor of two.
I understand why people want to hang on to their winners, but the compounding of capital—and I’m trying to create the largest pile of money for students. Students can’t pay their tuition with returns; they have to pay with dollars. I’m expressly interested in creating the largest pile of money, and the largest pile of money is governed by simple compounding math.
What we’re really after is the velocity of capital, not just returns on capital. Whenever the velocity of capital starts to asymptotically approach wherever it’s going to be, we want to be out and move on to the next thing. In other words, it’s really, really hard to do 3x in 6 years, right?
Yeah.
It’s easier when you have winners. The company’s going okay. It actually looks better in your marketing if you’re up 6x instead of 3x. If people held onto it for another 5 years and got a double, they’d be up 6x instead of 3x. That suggests that the next fund will be raised, et cetera, et cetera.
But I actually don’t care about any of that. That’s a business decision. That’s related to the business, and I’m not optimizing for the best business for the GP. I’m trying to optimize for the biggest pile of money for our students. I understand that there’s a little bit of a disconnect there, but the math issue does drive me nuts.
Can I ask you a blunt question, then? I love this interview because it’s completely not in my interest as a venture investor—and as someone who interviews venture investors.
No, no, this is why I love it. I have the best job in the world. Given the requirements on velocity of cash and the value of compounding, which I very clearly see, do you not have an internal question of, “Why do VC at all if we can do growth equity or mid-market and get the 3x in 6 years?” I get you, David. I’m not doing that for you, and neither are the best firms.
That is a question that gets batted around a lot in our office. There is something to be said about laddering returns, right? It’s okay to allocate money to some manager and say, “Those returns are going to show up 6, 7, or 10 years from now. These other returns are going to show up 3 to 5 years from now. Then, on my side, those returns are going to show up 1 to 3 years from now.”
We do think about it that way, but I would say that there’s a rule in our office that you’re not allowed to talk about returns without also talking about time. It’s very common on the private side to just say, “Well, you’re up 2x, 3x, 5x,” whatever. But that tells you nothing. If you’re up 5x over 30 years, that’s horrible. If you’re up 5x in 5 months, that’s amazing. I guess that’s SpaceX.
Is venture then just a pure diversification play for you?
3. Venture Diversifies The Endowment
It is for us. It could be the case that somebody allocates to something that really takes off and goes quite well. For example, we happen to have about 2.5% of the endowment in Anthropic. We have no exposure to SpaceX. We’ve had no exposure to OpenAI, but about 2.5% of the endowment is in Anthropic.
Well done.
That’s not us, right? That’s managers.
David, for goodness’ sake, will you please learn from your managers? Lesson number one of venture capital: even if it was not you, you take credit and say, “Thank you so much.” I remember that one.
That’s not really how we roll at Baylor, but understood.
Can I ask you—it’s a really difficult question, and I’m not saying with Anthropic here, but I’m saying with positions that go public? With positions that go public, obviously Anthropic will be one. How do you think about actively managing it as the holder versus the common response I hear, which is, “That’s not our job. We just liquidate the minute that we get it because we don’t know about this asset”? How do you think about it?
It depends on what we think about the name, and it also depends on the size of the position once it is public. We’ve sold shares before. We’ve also hedged shares before. We’ve also let shares run before. It depends on what we’re expecting, what the profile of the portfolio looks like, and the position and the risk associated with it.
I was talking to Sean before this show, who you mentioned we should chat to. He's brilliant, Sean Barrett. He said that you think more like Charlie Munger than anyone he's ever met. How did that resonate?
Only because we're in the middle of the country, I think.
He said that when software was getting killed early in 2026, you went deep on the situation, wanting to understand every bit of research, and then piled in. Can you talk to me about your process there, what you saw that others didn't, and how you thought about that? I'm fascinated, given that.
4. Human Behavior Challenges AI Hype
I would say that if we had an edge, we're pretty good on human behavior. I don't dispute any of these things. I'm not an engineer. Much of the stuff that comes out of Silicon Valley is over my head, but I do know how people think, and I do know how people make decisions.
It was pretty easy in this case. Software is dead, it's all going to zero, somebody's going to vibe-code this, and whatever. I have friends that run three 500-person private family businesses, and it's easy enough to pick up the phone and call them. We're like, "Hey, say your son-in-law vibe-codes something and you're going to tear out your CRM." And they're like, "Not in a million years." It's not their job.
I have a good friend who runs a vertically integrated potpourri business. He knows everything there is to know about that, but he is not going to tear out key, important parts of what makes his business run behind the scenes on some unproven thing. I think it was the CEO of Salesforce who said six or eight months ago that the best AI was going to be was 93% right, which is phenomenal and might be better than a lot of people. But the issue with software is that it has to be 100% right. If you need your books to match up, that's not going to happen.
As we've thought about it more, I actually think that in some of these vertical industries, software is going to be the delivery mechanism for AI. In other words, for my friend who's in a niche business and very, very good at what they do—I think they're the only vertically integrated potpourri maker in the world—the trust that's been built up with the software providers is going to translate into, "Hey, could you add AI bits for me on the back of this software?" Of course, the SaaS companies aren't stupid. It's not like they're sitting there thinking, "Hey, we're worth $20 or $50 billion. We should let this go to zero."
What's interesting to me is that you analyze this situation and then decide to act on it. That's very rare for an institution to do normally.
Sean and others have told me that, but I don't actually understand it. Software at that point was on sale to the tune of 50% or 60% from October 2025. If the thesis is that software is going away, it's down 50% or 60%, and you call businesses and they say that's not true, you're like, "I'll own that."
I get you. But it's throwing the baby out with the bathwater. The trouble is, I'm not sure what's the baby and I'm not sure what's the bathwater. With the greatest of respect, I live in technology—
And that's why we have managers like Sean, right? He's the expert. I'm like, "I'm going to give you more money, but I want you to go through your list with me and tell me all the things that are least likely to be disrupted by AI, and then own those."
I'm making a decision based on human behavior and how I know people make decisions, and I'm allocating based on that. But I'm relying on the manager to be an expert in their individual field and give me the correct perspective on what's going on on the ground.
What's so interesting is that most people just delegate to managers and go, "You're the experts." You delegate to them, great, and then you go, "I'm also going to operate where I have decisions myself, and I'm going to interject in those markets." It's different.
I kind of think that's our job, right? My seat is an allocator seat. My job is to allocate.
Yeah.
To go back to the Buffett or Charlie example, they also are allocators, and they're deciding who gets the incremental dollars. Do they send it to Burlington Northern or do they send it to their energy company? Depending on what the outlook is, what the CapEx requirements are, et cetera, they get budgets submitted to them, and they may or may not allocate more of their cash pile to those companies.
Quite a lot of LPs that I speak to say, "I get the liquidity challenge of venture, and I get the time lags of venture being difficult. But I learn a lot from what happens in my venture portfolios in terms of AI penetration, new technologies, and adoption cycles." Is your venture portfolio a learning academy for you or not?
Not for me. I would say it goes the other way. I actually learn a lot from the public-side managers.
What I find is that there's a lot of this spun-up, "Oh my gosh, we're going to have autonomous cars in three years," as people were saying in 2016. Yeah, right. All the regulatory stuff that you have to go through so that you don't kill somebody. We're 10 years on, and what do we have, like 50,000 cars on the road? Please.
I get the mental imagination that you can say, "We could put something on the moon and we could mine the moon," and whatever. Yeah, okay. Get back to me in 30 years.
Okay, but you're not worried, then, about the casinoization of public markets?
No. The public markets are the big leagues. There are millions of people making decisions on dollars every single day for every single company.
You know how things get valued on the private side. Of course you do. Three people get in a room and say, "Hey, I think the value is X." And they're like, "I'll fund it at that." Great, and that resets the whole price—
But I think public markets, in many respects, are as irrational as private markets are. You saw that. You saw that with SpaceX.
They can be irrational because they are governed by people. The difference is that there are tens of millions of people trading on that information, whereas on the private side, there's like three.
And those tens of millions of people decided that Elon Musk is a premium in himself, and that SpaceX should be a $1.8 trillion business.
Yeah. That doesn't mean that they're right. It just means that it incorporates all available information, which does not happen on the private side.
What you're saying is that the sheer scale of people voting in this buying decision means that it's a more legitimate price than on the private side, just so I understand.
Correct. I don't think there's any question about that. I literally have been in these conversations where three guys get together and are like, "Hey, I think it should be this." Like, on what? And they're like, "Well, I'll give you $50 million at that price." Okay, fine. But that's—
On the fact that I tried the product and I liked it—
Exactly.
—David. Why are you asking me such intellectual questions?
Exactly. Right.
Do you trust the prices coming back from your managers? I didn't mean that badly, but we all have our books and our portfolios, for people listening, and we mark them in different ways and explain them to everyone.
5. Conservative Marks Guide Better Decisions
We do. That's one thing that the ladies have done an extremely good job of. Recall again that I'm coming from the public side. When you run trading books, everything has to be priced every day, ostensibly so you make better decisions.
If you have things mismarked, then psychology works against you. If you say that this is worth $30 million and it should be worth $10 million, and somebody offers you $20 million, then because you would ostensibly take a loss from $30 million to $20 million, you're liable not to take that, even though it's a premium to the actual value.
Pricing is just a way to make sure that you are psychologically aligned to the reality of the market. One of the things that we really try to do is make sure that our managers are not pushing valuations. We want valuations to be conservative rather than aggressive.
You can see that in our return data in the 6 to 9 months prior to something being taken out. I think the average gain on that is 60% to 90%, and I think from a market perspective, it's more like 30% to 50%. That would suggest that our marks, our managers' marks, tend to be more conservative than others.
I sit on top of this thing, and I have to vouch for the valuations that we have as it relates to talking to the regents or administration. I feel pretty comfortable that on the private side, our marks are actually more sane than average.
As venture eats more and more of the world, with your OpenAI, your Anthropic, and your SpaceX, and your biggest companies in the world all being venture-backed companies, do you feel that you need more in venture, more in tech? Does it change how you view the world? Does the mindset change?
No, I feel pretty comfortable with where we stand. I think our biggest allocation is in growth equity on the private side, and we feel pretty comfortable with our capability and the manager set that we have there.
Why do you like growth equity? Is it because of the return timeline profile?
The return timeline, and there are also fewer zeros, which gets to the value piece. It’s simple math: if there are fewer zeros, everything else doesn’t have to cover for the things that don’t work, which helps get you to the outcome. I think their book is annualizing at around 30% on the growth equity side, so that obviously meets our 8% or 9% bogey. I don’t even know if I’ve ever had that question before.
How do you think about mulligan vintages across venture and PE, with “mulligan” meaning not-very-good vintages? A lot of people are talking about 2021 and 2022 for venture—
Mm-hmm.
—and PE being very bad vintages. We all went crazy. It was COVID. Mea culpa. And you’ve got Thoma Bravo now. Obviously, you had Medallia, which was quite a well-known return, and the keys situation that just—
Right.
—
I think that just comes with the territory, right? Basically, what we do is say, “This is the amount that’s going to be in privates.” Then we say, “We’re going to allocate to PE expansion capital and VC, and we’re going to do it in these sectors.” Then I let the ladies have at it, and they come up with a portfolio. The portfolio overall has an expected return hurdle that they need to clear. If they’re not clearing it, then that’s a problem. If they are clearing it, then that works great.
What are the annual liquidity requirements? Obviously, as an endowment, you mentioned paying for tuition, which is really important. What are the annual requirements in terms of liquidity for you?
It’s on a couple of fronts. Obviously, on the distribution side, that’s something we can’t get around, right? That’s about 5% on an annual basis. That dollar amount keeps going up, which we want it to, right? That’s the thing that pays for scholarships and professorships, et cetera.
On the subjective side—that’s the objective side of the liquidity equation—there’s the question of what capital you need to have around to allocate to the next thing that’s going to go up 20% or 30%. We talk to our newer analysts about this, and we say, “What do you think the odds are that we find something to be up 20% sometime in the next 4 years? Anything, anywhere?” They say, “Wow, really high.” We say, “Great. So then cash is worth 5% a year, apart from what you’re going to earn on cash.”
If cash is earning 3.5%, plus 5% in opportunity cost, cash is worth 8.5%. If we find things to do that are north of that, then we do them. If there’s a period in the market, such as 2017, 2018, or 2019, where we’re not finding things to do in that ballpark, then we let cash get larger. We came into the pandemic with around 15% or 16% in cash because we were looking around and saying, “I don’t see something to do.” Our cash balance is indicative of what we’re seeing to do to make money.
It’s very difficult to keep your head when everyone else is losing theirs. It’s a brilliant Rudyard Kipling poem, but it’s very difficult to do. When momentum and excitement kick in—
Mm-hmm.
—it takes one disciplined mind.
Interestingly, in this period—in the 2017, 2018, 2019 cycle—we weren’t finding other things to do. This time, we actually are finding stuff to do, so we’ve kept our cash balances pretty low because we keep finding 20% or 30% annualized things to do. Our cash balances end up being a function of what the environment is.
I think one learns a lot from their mistakes if they’re reflective. When you look at allocation decisions, what is an allocation mistake that comes to mind first, and how do you reflect on it and learn from it?
I can’t come up with a specific example right off the top of my head, but I will say this: whenever you’re trading, you’re going to lose money. Sometimes you’re going to lose a lot of money, and sometimes you’re going to lose a lot of money for a long period of time.
Basically, everyone goes through it. Everyone walks into the seat and thinks, “That’s not going to happen to me. This seems pretty easy.” Invariably, you get kicked in the shins and then hit over the head by a 2-by-4. The takeaway from that is that I never want to be all in. Things can always get worse.
When we were allocating to software in February and March of this year, we weren’t drawing a line in the sand and saying, “Every available dollar is going to software.” It’s down 50% or 60%. Who’s to say it’s not going to be down 70% or 80%? We set it up so that we’re methodically and mechanically allocating into difficult markets. The reason we do that is to try to take the emotion and psychology out of it.
How do you literally do that—methodically allocate into markets?
I’ll give you a perspective on the overall markets.
Yeah.
We basically say that if the market is down 0% to 10%, we don’t care. We’re an infinite-life portfolio; 0% to 10% is normal stuff.
The way I approach it with young analysts is, “If something’s on sale for 10%, do you rush out to the store to buy it?” They say, “No, not really.” I say, “What about 20%?” They say, “I’d think about it, maybe.” “30%?” “Yeah, probably.” “40%?” “For sure.”
We think about declines in the market in 10-percentage-point increments, and we have liquidity set up in such a way that we could allocate every 10 percentage points down. We don’t really worry about 0% to 10%.
How do you think about catching a falling knife? Let’s make this real. I’ve done that before. I’ve looked at Wix or monday.com, which were down impressively large amounts. I love the founders, but I determined that I couldn’t distinguish the baby from the bathwater and did nothing.
Right.
But they had another 10%, 20%, or 30% to drop.
That’s why we do it methodically and mechanically. We’re never drawing a line in the sand and saying, “Down 20%? I’m all in.” We’re saying, “Down 20%, maybe I’m 20% in. Down 30%, I’m another 20% in. Down 40%, I’m another 20% in.”
We do it that way, and the reality is that we never get all the way invested before it rebounds. You could say that we leave money on the table. That’s true. But the benefit is that we’re never in the situation where we’re saying, “Oh, my gosh, I love this so much, and it’s down, and I just can’t have any more of it.”
That’s the scenario we’re trying to avoid. It comes from perspective, history, and experience—from having trading scars all over your body because you thought you were right, thought you knew where it was going to go, put a whole bunch of money to work, and then it went lower. It’s a terrible place to be.
It is. When you’re holding a stock and it’s just down and you’re not in a good place, how do you determine the balance between “it’s going to come back, I was right, and I’m going to stick to my beliefs” versus “forget it, I just need to sell”? The utility value of cash, even if it’s a loss, is that it can be recycled. How do you think about that?
A lot of that is in the hands of the managers, of course, because we’re not trading individual stocks. But what I do find is that we spend a lot of time working with managers, making sure that their psychology and emotions are in the correct place.
For example, interacting with Sean—you brought up the software space—in the first part of this year, I was probably on the phone with Sean every day for 4 weeks. We were talking through individual names. I was relaying what I was hearing in the market, and he was relaying what he was hearing in the market. We were sending each other articles, quotes, and news stories at all hours of the day.
I constantly ask him, “Okay, you have this name, but if it goes down another 20%, what are you going to do?” Or, “You have this name and another name. Which one do you feel better about, or which has the better risk-adjusted opportunity set here?” Then I’d push him to be more concentrated. That’s actually what the portfolio ends up doing.
To your point, that’s what ends up happening in most cases in real-life downdrafts: portfolios end up getting more concentrated.
Does that make you nervous?
No. We own everything under the sun. So does every endowment portfolio, right? We own everything from sunscreen to helium to technology. We own all sorts of consumer packaged goods that you would see in the mall. We own all sorts of business-to-business software and technology companies that I’ve never even heard of before. We own real estate development projects. We own everything.
It’s always funny to me when people compare an endowment portfolio to the S&P 500 or something like that. We’re infinitely more diverse than the S&P 500. It’s not even close. So no, if we get a little bit more concentrated at the margin, that doesn’t remotely change anything for us.
What do you see your endowment CIO cohort do that you think is nuts or wild?
There’s something that we do that not a lot of schools our size do, and that is we almost hire exclusively from undergraduate ranks.
Now, to be clear, the caveat there is schools or endowments our size. So we're about 2.7 billion. Fourteen months ago, we were 2.2 billion. A couple of years before that, we were 1.4 billion, right? So we're in that sort of 1.3 to 3 billion range.
I've figured out why a lot of people don't do it, so it was something that I missed. The point is, if you hire undergrads, and given where we are, our office is located in Waco, about 100 miles from Dallas and 100 miles from Austin, right in the middle between the two, it's pretty difficult for us to hire a mid-career professional and get them to stay for a long period of time.
It would be really difficult to pull somebody from Los Angeles or New York to Waco and say, “I need you to be here for 10 years.” So what we've done to try to solve that is hire from the undergraduate ranks. They clearly have chosen the school; by definition, they've chosen the area. They've been around, and we actually screen pretty hard for that when we're hiring people.
The issue is that when you do that for the next 5 or 6 years, you're spending a lot of time pouring into that person and helping them level up. During that period of time, while they're leveling up, it's all still on your shoulders. I totally forgot that part. I got the idea that we'd have a stable investment team and these people wouldn't go anywhere, but I forgot the part where, for the next 5 or 6 years, you're going to be wearing all sorts of hats during that time.
So do you think your colleagues are nuts, then, for not hiring internally?
I think “nuts” is not the word that I would use. I would say that they are accepting alternative risks. The upside for me is that I have a stable team. I've worked with Renee for almost 16 years. The next person that we hired, Jen, has been here 11 years, and you can go down the line.
That actually accrues, and it's pretty evident across the industry: longevity begets returns. So I'm benefited on the stability front. The negative for me is the upfront burden of all that time. There's a period of time where I have to carry the team.
On the flip side, if you hire mid-career professionals, you don't have that upfront cost of having to carry the team because they're more plug-and-play. But you take on the risk of turnover and potentially poorer returns.
David, do you think the incentive structure for LPs is broken? Let's be specific about LPs, or endowment fund investors. If you look at funds of funds, if I crush it for my fund of funds, they obviously have carry and will do very well from that. With traditional endowment fund investing, if I do really well for you, it doesn't necessarily translate to a huge paycheck. Do we have the wrong incentive mechanism?
I don't think it's the wrong incentive mechanism. I think that it requires people in the space to be very missional. I wake up every morning motivated by sending some sophomore in high school to Baylor who hasn't even thought about college yet, or some 7th- or 8th-grader who doesn't know if they're going to go to college and is thinking about baseball scores from the prior night.
Me getting out of bed in the morning, going to work, and wanting to crush it is entirely due to that. Everyone likes to be able to get their wife something nice, redo the kitchen in their house, or go on trips, but that is not the motivating factor for either myself or the people on my team.
Do you worry about the impact of AI on education?
I worry about the impact of AI on human thinking. There are a number of studies out—I don't know about their veracity—but they're coming out of MIT and other places like that, suggesting that students who are using AI for everything they do actually show less brain function.
This isn't really a surprise. You see the same thing if you just sit in a chair all day: your muscle atrophies. So I do have concerns about the effect of AI on actual human logical thinking. That's an innately human trait. Animals don't think; humans think. But if you abdicate your responsibility for thinking, it's not clear that humans do that either.
So I have concerns about that. I think education can figure it out and use it beneficially. I think it's more of a human-discipline problem.
For a lot of my friends who are CIOs of other endowments, sometimes larger ones, they've been hit with the endowment fund tax, which is really hitting larger organizations. How do you think about that? How do you advise them?
I would love to be in their position. We are not, because our endowment per student is too small to be subject to that. But I promise you, if I went to the president of Baylor—I’ve actually had this conversation with Linda—and said, “There's good news and there's bad news. The bad news is that we're going to have to pay an endowment tax. The good news is that our endowment is 3 times bigger than when I last talked to you,” she'd be like, “Yeah, and?”
So I would love to have to pay the endowment tax because the endowment was bigger.
Do you play a game of comparison? What is it—comparison is the thief of joy. I think you said earlier that in down times, you're obviously brilliant, and in up times, it's more challenging. I think you play a defensive game. Full-year 2025, Baylor returned 9.4%.
Mm-hmm.
Dartmouth was lowest at 10.8%.
Mm-hmm.
Do you do the comparative side-by-side, or do you row your own race?
We do both, which I think is the right way to do it. Every school has a different set of priorities and needs. Baylor's current priority is to get the endowment higher on a per-student basis.
What you're referring to in terms of last year was disappointing on a relative basis, but there were 2 things going on. One was that we had increased the allocation to private investments, in terms of annual commitment amount, by about 60% to 70% in 2020, 2021, and following years. So returns from the private side have been dealing with a second J-curve, if you will.
Mm-hmm.
The fund of one, and another asset class that we'd allocated to, had been flat and were starting to inflect up. So this fiscal year was the first year that we weren't dealing with the J-curve impact on the private side, and the first year that we got returns from both the fund-of-one category and this other category.
We feel very, very good about it. Our newest analyst was like, “So basically, you guys tried to change the engine while the car was moving.” And, yes, that's 100% what we were trying to do. We were trying to put a new, bigger engine in the car while it was still going down the highway, and we did it.
We had a little bit of a lag last year. I think we'll be at 18.5% to 19% this year without any SpaceX or Cerberus or anything like that. Structurally, I think the next couple of years look pretty good from a tailwind perspective.
Can I ask you—we chatted before about a friend of mine who you're going to be working with—how do you think about position sizing in the positions that you do decide to engage with?
6. Position Size Must Move The Needle
Yes, this is an interesting one, and you're talking specifically about the private side.
Yes, on the private side.
We've spent a lot of time on this because of what I talked about earlier: we own everything under the sun. One of the things that we figured out is that we have this relationship with a GP, and they send us a little note saying, “So-and-so company got sold. It was a 7X return.” I'm like, “Okay, great. What does that mean to us?” And they're like, “Well, we'll get back about $400,000.” And I'm like, “What? Who cares?” It means nothing to the overall endowment.
So one of the things that we've changed in sizing is that we start with how much money we want to have in each underlying company. In other words, if the company's going to be up 5X, we want that 5X to matter to the overall fund.
Basically, what we're doing in expansion and buyout—venture is a little bit different because it has a bigger company set—is that we're saying we want $3 million to be in each underlying company. So if they have 10 companies on their platform, that means we'll allocate $30 million.
I get you totally. Another way that I think about it—and you can tell me if I'm wrong, which very possibly could be the case; I'm a low-IQ individual, after all—is that it's a $200 million fund, and you commit $20 million to it with the theory that if they say we have 10% ownership in every company, great. If we're 10% of their fund, our exposure is 1% per company.
Yes, we think about it in terms of dollars. We say, “How many companies are you going to have—8, 10, 12?” And then we want $2.5 million to $3 million in each company. Obviously, it's up to the manager, et cetera. We're not dictating that, but we're just doing the math from a dollars perspective.
So if you have 10 companies, we want $3 million in each company. If it was up 5X, we'd get $15 million back. That matters. That's enough to matter.
Do you want your manager to do what they said they would do or to play the game on the field? Ventures change more in a year than I've seen in a decade, and actually playing the game on the field, as Bill Gurley says, is the job of a venture investor. That may be different from what I said to you I'd do.
We always want managers to do what they said they were going to do. My example is always this: I view my job as the general manager on a baseball team. I'm going to hire a third baseman, a shortstop, a second baseman, a first baseman, et cetera, for various reasons, depending on your fielding percentage, your batting average, and so on. But if I walk out on the field and I have 2 people on second base, someone's getting fired, and it's probably the third baseman who switched to playing second because I have people set up on the field to play particular roles for particular reasons.
If you're a third baseman and you think you can play second base better than my second baseman, then you should come talk to me. But if I ever walk out on the field and I have 2 second basemen and no third baseman, the third baseman is getting fired, full stop. I don't care what your returns are. So again, we started off by talking about this: We spend much more time on asset allocation and why things are where they are than we do on individual managers.
Okay. This is so interesting for me. The markets have changed in the time that I've raised from you. I've moved with those markets, and I've done that well, and I'm showing you great numbers. I'm making you money. But my position on—
If she doesn't fit what we're trying to accomplish, we won't re-up.
That's so interesting. So you would rather I stayed on second base and did worse financially than moved to third base, where you've already got someone else?
I would like you to have a conversation with me before you change your stripes.
What would you say in that conversation?
I'd be like, "Why do you think that you should be able to do this when we have no data to suggest that you're good at it?" Let's move out of VC. Let me do something in public equity that's easier. There are managers who are like, "We don't know how to time allocations into and out of cash. We're just going to be fully invested because we don't know if the market's going to go up, down, or sideways. We're good at picking stocks, so we're going to keep basically zero cash. That's what we do."
And then there are other managers who are like, "We actually use cash as an allocation methodology, and cash will be from 0% to 15%, depending on what we see to do, whatever." Both of those track records are subject to comparison with benchmarks. We don't change the benchmark depending on whether somebody holds cash or doesn't.
If somebody is like, "We're fully invested all the time," and then I wake up some morning and they have 10% in cash, yeah, they're getting fired. I don't want to be the guinea pig. I don't want to be the person they're like, "Hey, you have a new idea now, and now you're more of a global macro equity manager, and you think that you can time the markets when you have no prior experience or data to suggest that you can?" Yeah, no, you're fired.
I get that example. I think venture is more nuanced. It's kind of closer, you know—
It always is, right? I'm not talking about lines in the sand around artificially generated category limitations. That's just an artifact that people made up. I'm talking about: A, we're going to invest with a manager who is investing in companies where product-market fit has already been determined, and then that manager is like, "Yeah, that doesn't work anymore. We're just going to invest in 2 guys in a garage, and we don't know if they'll come up with something or not."
Yeah.
Right? Those are 2 very different approaches. So, yeah, the switch between those is a no. If you want to go from B to late A, who cares? That's the same thing.
So we're actually completely aligned. It's interesting—I thought we were misaligned. I 100% agree. I think your example there is kind of like—I always say pre- and post-data, which is: You either have nothing and we're selling Walt Disney. "Tell me a story." Some people are great at that.
Right.
And you should bet on them for being great at that.
Right.
Or you're Jerry Maguire: "Show me the money," which is the post-data, and some people are great at that.
Right.
And so I totally get that, and I think you're absolutely aligned there. Can I ask you a tough one? In venture, it's kind of assumed—and it's the unwritten rule—that you commit for 3 funds. And you should, because that's the duration required to determine quality in a manager. Do you think that's kind of bullshit, coming from a more macro perspective where you see different asset classes?
I don't know. I think we've kind of done that. The issue with 1 fund, even 2 funds, is you almost don't have enough data to make a decision, right?
You don't.
And so, yeah, that kind of makes sense because you don't have data to prove it otherwise. We tend to be very good when there's data to be analyzed, and we tend to be less good at, "You guys have a vision. I got a dog. Give us some money." That's really hard for us. People are good at different things.
If I were to say you had unlimited money today, unlimited constraints, and you had the Harvard balance sheet, what would you do differently?
I don't know that I would do anything differently. I think it gets a lot harder, for sure, at that size and scope. Hats off to Narv and what his team is trying to do. That's really, really hard. I've actually talked to other CIOs about this because I want to be prepared for down the road. At what point do you have to change how you invest? That's very top of mind for us, and that's something that a lot of allocators work on, think through, and struggle with.
How do you answer that?
I've talked to the Notre Dame folks, and they're at $20 billion. They're kind of like, "We actually thought that we would run into this at $10 billion, at $15 billion. We actually haven't." I wonder if there's a place between where Notre Dame is and where Harvard or UTIMCO is where you actually do have to change how you invest, or you can't invest in the same manner.
Because at some point—and I think that some of the Ivy Leagues are running into this—it kind of doesn't matter how good benchmark returns are. When you have $40 billion or $60 billion and you can allocate $20 million to a fund, even if you're up a real lot, it doesn't move the needle as much as it used to.
When you put that into perspective, if you have a $20 million check in a fund, and I'm sure a benchmark—we can use that with a multiple here—$20 million, and you do a 50x, say it's another eBay fund—
Mm-hmm.
—which would be amazing. I mean, Jesus, amazing. 50x a fund. That would return $1 billion. And so, to your point of materiality to a fund, yeah, if you're a $40 billion or $50 billion endowment—
That's 2%, really.
Well, are you going to send me a Christmas card thanking me? Come on. Give me—
Yeah. I mean, $1 billion is great, but you see the point.
I do.
I think that's a little bit of what the a16z kind of thing is tapping into, right? Just the—
Yeah, the platforms win. Just don't do those checks. Just give me—
Right.
—$300 million.
Exactly. And so that's what I mean: At certain sizes, maybe you have to play the game a little bit differently.
Do you like the large venture platforms? Or are you like, "Nah, I don't like the post-$1 billion funds. We like small"?
I think it just gets harder. I think it gets harder to have a return figure over the requisite period of time that actually pays you for the risk that you're taking. I get how they do it. I get why they do it, but I think it is the law of large numbers, right? It's just harder.
I've loved this conversation. It's been very unusual. That is true.
Normally, everyone in my world—and this is the most idealistic, AI-pilled venture investor—is just like, "We're not going to have jobs in a year."
Yeah, that's literally not true.
Everyone's going to be replaced.
7. AI Infrastructure Hits Permitting Limits
I mean, you are seeing the pushback on AI at the data center level because the data centers are being built in my neck of the woods, not in Silicon Valley.
And you don't want it.
I do because we're invested in it.
Right.
But you're seeing this sort of pushback nationally; you're seeing it actually internationally. The most valuable thing for a data center used to be power. If we go back 5 or 6 years, it used to just be land, and then it was powered land, and now it's actually permitted powered land.
And the reason is because people are kind of fed up with it, and they're just like, "Not in my backyard." And so you are getting sort of this pushback on data centers.
I'm sorry, I don't understand this. Why? They're utilizing land. They're bringing jobs; they're bringing construction.
Power prices go up, and water prices, particularly in arid regions like Texas or Arizona, are a big issue. So, if the hyperscalers solve the water thing, that would go a long way toward the average person being more accepting of it, but then the power thing still exists. We know that power dispatch is still supply-constrained, and people's power prices are going to go up until that gets solved over the next 5, 6, 7 years.
David, what do you think happens here? As you said, you're an investor, and it's a fascinating perspective you have. What happens here? I'm very naive.
What happens with data centers?
Yeah, do we see continued protest and pushback?
Yeah, I think we do. We're seeing it in real time in our book. The data centers that have power and permits are becoming more valuable. Literally, we have this situation in our book where our data center sites are up 50% from where they were 6 months ago.
Can I ask what percentage of data centers you think will fail to get up and running despite having been built? This could be permitting, power, or whatever you want.
I'm not an expert in this in terms of the total number that have power, the total number that have permits, et cetera, so I'm not going to be able to give you an answer that's going to satisfy the question. But I will say that enough aren't happening that the power companies are coming to those who do have permits and saying, "We can get you power sooner than we thought." That's literally happening.
And just so I understand, the bottleneck on those that aren't is permitting.
Mm-hmm.
A pushback from—
Yeah.
Locals. What's the one thing?
Yeah, it's permitting.
Permitting.
Yeah. It is now, and that's something that didn't exist 6 months ago.
And just so I understand, again, I'm dumb as rocks: Why is it so difficult to get permits for these?
Because the permitting boards are governed by the citizenry, and the citizenry is putting signs up in everybody's front lawn saying, "We don't want this," right? So if those people want to get reelected, then they've got to say no.
Do you not worry that this doesn't happen in China, where it's just a free-for-all?
I don't think that it does happen in China because they don't particularly care. They just build it where they need it. But it's a bigger issue. Honestly, it's a huge, massive issue in the UK. It's by far a much bigger issue in the UK than it is in the US on the permitting front.
What are you talking about? We're not allowed to go outside or move a bin, let alone build a data center. Are you—
That's my point.
Are you—
That's my point. So we have a permitted data center site in the UK, and it's worth a lot of money simply because we have a permit.
Are you bullish on Europe, given what you just said there?
No.
Because of permitting? Because of—
Yeah, because of all of it: because of the defense structure of it, because of Russia, because of being behind on AI, because, because, because. Yeah.
Would that prevent you from allocating toward European managers?
No, we have allocated to long-short managers in Europe precisely because I think there are going to be some companies that win and some companies that lose. But I will also say that some of our bigger macro hedges are on European indices.
David, I could talk to you all day. I'd love to do a quick-fire round.
Okay.
I say a short statement, and you give me your immediate thoughts.
This could be highly dangerous for me.
Oh, don't worry. We've gone to Chinese permits, so trust me, the quick-fire will be like a piece of cake. What have you changed your mind on in the last 12 months?
Software was one, so we leaned into it pretty hard. We also took energy off around the same time, sort of with the advent of the US-Iran war and the Strait of Hormuz situation. When crude went north of $100, we took a lot of our energy length off. We did add to private equity sponsors around March or April. We don't really like private credit, but we do like the private equity sponsors, and so we've allocated more in that direction.
What asset class do you think is overhyped today?
Private credit, because it's easy.
Why don't you like private credit? I'm not in it; I don't understand it.
I'm not in it, and I don't understand it either.
But is it just shit returns? I remember I had a girlfriend who did private credit, and she told me it was crap returns. I listened, and I was like, "Yeah, you're right. It is crap returns."
Well, I think effectively what is happening is that you have credit exposure in companies that looks and acts a lot like equity to the downside, but you don't have upside equity returns.
Mm-hmm.
So I think the risk-reward profile is kind of off, right? We prefer equity to that.
Which other endowment fund do you most respect and admire because of its build-out, and why?
Brown, without question.
Why them?
I just have a ton of respect for Jane and the team that they have built there. It's also the case that their returns are better than ours, at least over the last 10 years. I think our returns might be better than theirs over the last 5 years, but we've got a lot of wood to chop to catch up to where they're at.
They are what I would describe as real investors. They'll do things that take a lot of courage. I'm not saying that they're riskier, but they're thinking through the risk-return profile of things. They've just done an extraordinary job. It's not just because I know Jane. We talk and chat and whatever; I just have the utmost respect for that team.
Which fund are you not in that you would most like to be in? We mentioned some of the big names.
Probably Benchmark.
That'd be the same for me. Yeah.
A ton of respect there.
Final one for you. What are you most excited about in the next few years?
I do think that biotech is going to be even more impactful over the next 10 years than it has been over the last 10 or 20 years. So we're spending more time on that. In fact, later this week I'm headed to a biotech conference, and then again in October. Biotech is something that we're actually spending a lot of time on. We certainly have a lot of biotech exposure, but we're wondering if we should have even more.
It seemingly is less correlated with markets; certainly, the science is less correlated with markets. What scientists are doing these days in actually solving diseases, as opposed to simply treating symptoms, is extraordinary. So biotech is certainly high on the list and something that we're spending a bunch of time on.
Aside from that, from a personal perspective, I'm really excited to see our team and our office build out over the next 3 years. As I've done this, I think there is really a major inflection point that happens when you're going from $1 billion to $5 billion, and we're right in the middle of that.
We're dealing with all of the issues around how do you grow a team? What systems do you set up so that when you're at $5 or $10 billion, you can actually keep track of everything? How do you systematize things so that this is a self-perpetuating office, et cetera, but retain the creativity to continue to do the new things that you've done in the past to get here?
There's a lot of decision-making that has to go on between $1 billion and $5 billion, and I didn't really appreciate that until being in the middle of it over these last couple of years. We're halfway through it, but I think in the next 2 to 3 years we'll get out to the other side and be off and running. At a personal level, that would be tops for me.
David, I've so enjoyed this. I'm very grateful to you for putting up with my varying questions and my naivety in certain cases, but I've loved it, so thank you so much for joining me.
No worries. We're down here in central Texas trying to do a good job, so thanks for having us.