Fireside Chat
Bill Maris: How Google Could Crush AI Competitors, Why Small Funds Win, and AI's Atari Stage
- Bill Mares will exercise high selectivity in investing for Section 32, utilizing six funds averaging $400 million in size to target companies such as CrowdStrike, Cohere, and Coinbase.
- Performance expectations are set with a goal for all six Section 32 funds to rank in the top decile based on DPI, leveraging historical data showing funds under $750 million average a 4.76x return compared to 2.42x for funds exceeding $1 billion.
- Specific return targets dictate that a $500 million fund requires $15 billion in exit value for a 3x long-term return, whereas a $7 billion fund would need $210 billion, a figure that historically exceeds total venture-backed M&A and IPO exit value except potentially in the current year.
- Mares identifies a risk of extreme market bifurcation with a handful of winners and numerous losers, predicting that achieving a $100 billion single-fund return requires public market liquidity that may not be supportable for entities like SpaceX or Anthropic.
- Strategic focus for Section 32 excludes large AI models in favor of enabling infrastructure including platforms, physics engines, controllers, and GPUs.
- Mares anticipates the AI gaming sector will resolve issues regarding memory, consistency, and session resets within five years, transitioning the industry from a command-line stage to a consumer-ready "PlayStation 10" equivalent in that timeframe.
- A 100-year forecast predicts the world will change by orders of magnitude due to AI, though non-financial metrics for fund success are deemed impossible to measure and likely to fail.
- Risks in the AI sector include the potential for Google to arbitrarily cut token costs by 80%, causing critical compression for competitors like OpenAI and Anthropic, alongside the possibility of firms using capital as a weapon to burn cash and capture market share similar to an Uber model.
- Public market dynamics may force passive funds and ETFs to acquire companies with bizarre S&P 500 exceptions, potentially creating "bag holders" among retail investors, retirement plans, and 401ks if valuations are deemed overpriced.
- Mares expects to evaluate public market sentiment regarding private company valuations and lockups approximately six months after listing, while noting insufficient data science supports the conclusion that late-stage investment is a trend for massive multi-trillion dollar exits.
- Life sciences investment is a primary area of interest given the sector's large total addressable market, with expectations for massive acceleration if in silico human cell stimulation becomes viable, despite FDA and clinical trial constraints limiting exponential growth.
- Concerns regarding basic research funding in the US due to regulatory and political factors are driving a capital flight to India and China, where recruitment of European and Indian scientists is eroding US neurological reserves.
- Alex Kokoroski notes that deep tech investments and long cycles are becoming more tractable due to accelerating market dynamics.
- Sachs argues that the incentive structure for giant funds writing large checks is broken, as a $5 billion fund returning 1.01x allows GPs to raise subsequent funds and earn more than smaller funds with higher returns.
- Sachs predicts the pendulum will swing away from late-stage strategies where investors write $50 million checks and wait for breakouts, citing data that suggests this approach is not sustainable long-term.
- A risk highlighted by Sachs involves entrepreneurs accepting inflated valuations from giant funds unless they are seasoned enough to recognize pitfalls, with examples showing $100 million valuations potentially rising to $4 billion for a $250 million check.