Inside the AI Sprint, Understanding Anthropic's Strategy | Tomasz Tunguz, Theory Ventures

Tomasz Tunguz, founder of Theory Ventures, characterizes the current AI landscape as an all-out sprint, with infrastructure spending on data centers and GPUs projected to reach $1.2 to $1.4 trillion this year. This rapid growth mirrors historical infrastructure booms like the national highway system, yet the industry faces unique challenges regarding GPU scarcity and power availability. While frontier models converge in capability, the primary competitive advantage shifts toward capturing user workflows and maximizing inference usage. Anthropic’s strategy of commoditizing complements—making workflow tools free to drive inference volume—exemplifies this shift toward high-margin, agentic B2B applications. As AI transitions from a search-based tool to an agentic workflow engine, productivity gains will likely manifest as increased throughput in software engineering and back-office operations, fundamentally reshaping labor markets and corporate efficiency.