Interview, Fireside Chat
Miles Dieffenbach: Inside Carnegie Mellon’s $4BN Endowment & The Math Behind DPI, TVPI, Illiquidity
- Venture capitalists are expected to take companies public immediately, with public markets anticipated to reprice risk to allow mature firms like Circle and CoreWeave to trade at healthy multiples, while 2022–2024 is projected to see more IPO capital than 2002–2004 despite the asset class being ten times larger.
- Median IRR for mature venture funds over 10–15 years (1998–present) is estimated at 8% net, with top quartile returns reaching 15%, though 70% of investors will likely lack access to top decile managers.
- The asset allocation model differs by firm size, where multi-stage large-check firms rely 70% on access and 30% on stock picking, whereas small early-stage funds depend 80% on picking and 20% on access.
- Multi-stage funds face significant scaling hurdles, requiring a $140 billion enterprise value investment base and $800 billion in exit market cap to achieve a 4x net return, a target considered staggering against 2021's $850 billion exit total.
- Endowments targeting 8–10% long-term returns face headwinds from a potential 8% tax range, forcing leverage draws to 4.5–5% to counteract 3% inflation, while family offices and LPs face churn due to liquidity crunches and partnership dissolutions caused by broken cap tables and founder transitions.
- The AI sector is predicted to experience a cycle similar to past technological bubbles (railroads, internet), with a likely bubble pop within a 10-year timeframe for 10% GDP productivity growth to materialize, rather than 3–5 years.
- Specific technology risks include NVIDIA facing a potential 40% earnings drop and valuation contraction from 38x to 24x earnings forward if revenue falls 20–30%, alongside a risk of hyperscalers investing a trillion dollars in CapEx (2024–2027) facing pain if returns do not materialize in a decade.
- Liquidity events from 2026, specifically deals like Wiz (Q1 2026) and Figma, are expected to provide market inflows, though a surge of LPs returning in 2027 is not immediate, and "multiple years of really good liquidity" are required to normalize the market.
- Strategic acquisition activity is constrained by a potential 12-month regulatory review window that could render companies obsolete, although a green light on the Wiz deal is expected to restart big splashes by MAG7 companies, which historically preferred strategic M&A over buybacks.
- Current valuations remain challenged, with non-AI deals still priced rich, early-stage venture identified as the riskiest asset class due to lack of entry profitability, and 2021-era returns from $400 million Series A funds assumed to be unattainable for current massive funds like MASA and SoftBank.
- Independent AI companies like OpenAI and Anthropic face high uncertainty regarding their status in five years, with risks of capital market disruption causing them to fail if they lack self-funding control, as demonstrated by OpenAI burning $5–10 billion annually.
- Market structure trends indicate 70% of US deals are AI-focused, creating alignment issues for Chinese funds excluded from USD AI investments, while multi-stage firms are encroaching on the seed market, treating specialized seed funds as competitive shrapnel.
- Historical context notes that 2021 was a "bubble" and "best exit year" with $850 billion in value, but most assets from that peak remain significantly down, with 2021 IPO prices still suppressed, and major funds failing to raise $1–2 billion despite their size.
- Fee structures in venture capital are expected to undergo evolution similar to hedge fund changes seen in 2005–2007 and 2010–2013, while 99.9% of general partners claiming their fund size is "perfect" and static are described as misrepresenting their fundraising intent.
- SpaceX is uniquely positioned with Starlink at escape velocity requiring no cash, whereas other companies like Google and Meta historically maintained 30–40% operating margin gaps at IPO to retain control, a buffer many current companies lack.
- Private market capital historically remained cheaper than public capital for an extended period, driving companies to stay private, but this dynamic is shifting as the speaker anticipates a return to public market pricing that reflects different risk assessments.