Watts, Wafers, and the Future of AI Infra | Gavin Baker

Key Takeaways

• The AI sector experienced an unprecedented surge in growth, with companies like Anthropic adding billions in ARR within a single month, dwarfing the decade-long growth of established SaaS giants.
• The market experienced a unique economic and technological environment in March and April, presenting opportunities for investors who could distinguish between temporary drawdowns and fundamental business invalidation.
• The compute demand for AI reasoning models during inference is significantly higher than for non-reasoning models, as evidenced by rising GPU prices and decreased availability.
• Geopolitical events, like the Strait of Hormuz closure, unexpectedly boosted America's relative manufacturing competitiveness by lowering energy input costs compared to other regions.
• Despite the extraordinary AI boom, tech valuations became relatively attractive compared to the broader market, offering a rare chance to invest in a transformative trend at historically low multiples.
• Anthropic demonstrates superior capital efficiency compared to OpenAI, burning significantly less capital to achieve similar revenue scales due to lower cost-per-token and potentially higher intelligence density.
• The intelligence of AI models, like Claude, can be measured by token generation efficiency, suggesting that Anthropic's approach may lead to better performance and higher potential revenue.
• Despite massive inbound interest, AI companies like Anthropic and OpenAI are prudent in managing their capital raises due to the inherent uncertainties and capital-intensive nature of the AI future.