The Infrastructure Layer of the AI Economy: Where We're Investing

Every major technology wave produces the same investment dynamic. Early capital flows to the most visible applications — the companies building products that consumers and enterprises can see and touch. Later capital, and often the most durable returns, flows to the infrastructure layer that makes those applications possible.

The AI economy is following this pattern. The application layer is attracting enormous capital — AI-powered SaaS, vertical AI solutions, generative AI consumer products. But the infrastructure layer — the systems, platforms, and services that the entire AI economy depends on — is where we believe the most durable value will be created.

We define the AI infrastructure layer broadly. It includes the obvious components: compute, networking, and storage. But it also includes the less obvious ones: data infrastructure, model serving and deployment platforms, AI observability and governance tools, and the security and compliance infrastructure that enterprise AI adoption requires.

Our investment thesis in this space is built around three observations. First, the demand for AI infrastructure is structural and growing — every organization that deploys AI needs it, and the complexity of enterprise AI deployments is increasing, not decreasing. Second, the switching costs in infrastructure are high — organizations that build their AI stack on a particular platform tend to stay there. Third, the competitive dynamics in infrastructure favor technical depth over distribution, which plays to the strengths of the founder profiles we back.

The specific areas where we are most actively investing include: enterprise data infrastructure (the systems that make organizational data AI-ready), AI deployment and orchestration platforms (the tools that manage the complexity of running AI in production), and AI governance and compliance infrastructure (the systems that allow enterprises to deploy AI within regulatory constraints).

We are deliberately cautious about the application layer — not because we don't believe in the opportunity, but because we believe the competitive dynamics are more uncertain and the moats are harder to build. The infrastructure layer offers a more predictable path to durable competitive advantage, and that is where our capital and attention are focused.