Avoid investing directly in pure-play Large Language Models (LLMs) like OpenAI or Anthropic, as incumbents like Google (GOOGL) are positioned to crush their margins by aggressively cutting token costs. Instead, shift your focus to the "picks and shovels" of the AI revolution, specifically companies providing GPUs, controllers, and platform infrastructure that solve current hardware and latency limitations. For private equity and venture exposure, prioritize small funds (under $750 million) over mega-funds, as smaller vehicles historically represent 95% of top-decile performers. In the life sciences sector, look for Computational Biology and Bio-IT firms that use computer-simulated modeling to bypass traditional, slow-moving clinical trial bottlenecks. Be extremely cautious with upcoming high-profile AI IPOs, as much of the value creation is occurring in private markets, leaving retail investors at risk of buying overpriced assets at the peak.

Artificial Intelligence (AI) Infrastructure & Platforms

• The current state of AI is compared to the "Atari command line" or "Zork" stage—brittle, lacking memory, and inconsistent.
• A massive shift to the "PlayStation 10" stage of AI is expected within the next five years.
• Investment Focus: Avoid investing in the large language models (LLMs) themselves. Instead, focus on the "machinery" and "physics engines" that enable ambient computing.
  • Key areas: Controllers, physics engines, GPUs, and platform infrastructure.
• Competitive Landscape: Google has a "war chest" that could allow them to crush competitors like OpenAI and Anthropic by cutting token costs (e.g., by 80%).

Takeaways

• Avoid LLM "Model Wars": The business models of pure-play AI companies are vulnerable to price wars initiated by incumbents like Google.
• Bet on the "Pick and Shovels": Look for companies building the underlying infrastructure that solves AI's current limitations (lack of session memory, consistency, and high latency).
• Timeline: Expect an order-of-magnitude change in the next 5 years, significantly faster than previous technological cycles.

Venture Capital Strategy & Fund Size

• Small Funds Outperform: Data shows that funds smaller than $750 million represent 95% of top-decile performers.
• The Math of Returns: A $500M fund needs $15B in exit value to return 3x; a $7B fund needs $210B. The latter often exceeds the total annual venture-backed M&A and IPO value.
• Incentive Misalignment: Large funds often prioritize management fees over performance. They may artificially inflate company valuations (e.g., valuing a $100M startup at $4B) just to deploy large amounts of capital.
• DPI (Distributed to Paid-In Capital): This is the only metric that truly matters in venture; paper gains are irrelevant until a liquidity event occurs.

Takeaways

• Investor Selectivity: For those looking at private equity or venture exposure, smaller, more concentrated funds typically offer better risk-adjusted returns than "mega-funds."
• Valuation Warning: High valuations for early-stage AI companies are often a byproduct of fund size pressure rather than intrinsic value.

Computational Biology & Life Sciences

• Shift from Traditional Biotech: There is a move away from traditional therapeutics that require long, expensive human clinical trials.
• Computational Biology: High interest in "in silico" (computer-simulated) human cell modeling to accelerate drug discovery.
• Longevity: Once considered "fringe," longevity science is becoming a mainstream investment sector (e.g., New Limit).
• Geopolitical Risk: The U.S. is losing "brain trust" to China and India due to restrictive H1B visa policies and a perceived "anti-science" sentiment.

Takeaways

• Focus on "Bio-IT": Look for investment opportunities where biology intersects with high-performance computing.
• Regulatory Lag: Even with AI, the FDA process remains a bottleneck; biology will not move as "exponentially" as software due to safety requirements.

Public Markets & IPOs

• The "Bag Holder" Risk: Companies are staying private longer, capturing the majority of value creation for elite private investors.
• 401(k) Risk: There is a concern that by the time these multi-billion dollar AI companies go public, they will be "overpriced products" forced onto retail investors and retirement accounts.
• Bimodal Returns: Venture returns are extremely bimodal—a few massive winners (e.g., SpaceX, Coinbase, CrowdStrike) and many losers.

Takeaways

• Retail Caution: Be wary of high-profile AI IPOs in the coming years. If the "value creation curve" has already peaked in the private markets, public investors may face limited upside or significant downside.
• Monitor Liquidity: Watch for changes in "lock-up" periods and how public markets value future cash flows for companies like Anthropic or SpaceX.

Specific Companies Mentioned

• Google (GOOGL): Positioned to use "capital as a weapon" by lowering token prices to gain market share.
• Coinbase (COIN): Cited as a successful investment from Section 32's smaller fund model.
• CrowdStrike (CRWD): Noted as a top-performing investment.
• Cohere: Mentioned as a strategic AI investment.
• New Limit: A company focused on longevity (backed by Brian Armstrong).
• OpenAI / Anthropic: Identified as being at risk if incumbents (Google) decide to compress margins.