Cathryn Chen on AI, Valuations, and What Actually Lasts

Key Concepts

• Infrastructure-First Investing: Prioritizing long-term, foundational assets (space, data centers, nuclear energy) over speculative software trends.
• AI Integration: The perspective that AI is not a standalone vertical but a mandatory horizontal layer for all future-proof businesses.
• Market Volatility: The observation that AI-driven innovation cycles are accelerating, leading to faster company growth and failure rates.
• Pre-seed Valuation Inflation: The trend of early-stage startups raising capital at extreme valuations (e.g., $4 billion).

Investment Philosophy and Market Observations

The speaker highlights the current volatility in Silicon Valley, noting that the rapid pace of AI development has created an environment where companies rise and fall in significantly shorter cycles than in previous decades. A specific point of concern is the trend of pre-seed companies raising capital at valuations as high as $4 billion. The speaker intentionally avoids investing in these high-valuation, speculative entities, viewing them as potentially unsustainable in the current fast-paced climate.

The Shift Toward Foundational Infrastructure

Instead of chasing AI-specific software applications, the speaker advocates for an investment strategy focused on "true infrastructure technology." The core argument is that while AI will impact every sector, the most resilient investments are those that provide the physical and energy-based backbone for the digital economy. The three primary pillars identified are:

• Space: Infrastructure related to satellite technology and orbital operations.
• Data Centers: The physical facilities required to house the massive compute power necessary for AI training and inference.
• Nuclear Energy: The power generation capacity required to meet the surging electricity demands of modern data centers.

AI as a Mandatory Utility

The speaker rejects the notion of AI as a separate investment category. Instead, they argue that AI is an essential tool for any company aiming to remain competitive. The perspective is that AI integration is no longer optional; it is a baseline requirement for operational efficiency.

• Key Statement: "If anyone here who's not using Claude, I'm sorry, you're already behind."
• Supporting Evidence: The speaker suggests that the speed of technological advancement is such that failure to adopt current LLMs (Large Language Models) like Claude results in an immediate competitive disadvantage.

Synthesis and Conclusion

The overarching takeaway is a shift in venture capital strategy from "hype-driven" software rounds to "utility-driven" infrastructure. By focusing on the physical requirements of the AI revolution—energy, compute, and connectivity—investors can mitigate the risks associated with the rapid boom-and-bust cycles of AI software startups. The speaker concludes that the future belongs to companies that treat AI as an integrated utility rather than a standalone product, supported by a robust, long-term physical infrastructure.