Backing the Foundation ofEnterprise AI.

A bad decision by a human is one failure. A bad decision by an agent operating at machine speed across distributed systems is catastrophic. As models improve, they get better at handling unstructured data on their own. Structured data is a different problem. Incumbent data layers were designed for human, not agent, consumption — and LLMs still need purpose-built infrastructure to integrate and manage them reliably.

We are entering an era where AI agents don't just assist — they replace entire job functions. Agents can now research, draft, negotiate, code, schedule, and close loops across systems without human intervention. This creates massive demand for new infrastructure: orchestration layers that coordinate multi-agent workflows, identity and permissions frameworks for non-human actors, observability for autonomous decision chains, and guardrails that ensure agents operate within policy. The companies that build the connective tissue between agents and enterprise systems — the authentication, the memory, the tool-use protocols — will capture outsized value as work itself gets automated at machine speed.

Attack cost approaches zero. Adding intelligence to legacy architectures improves efficiency, not resilience. AI has dramatically expanded the attack surface while also creating new defensive advantages. The rise of autonomous attackers is forcing a rethink of how security is built, deployed, and defended — creating opportunity for a new generation of cyber platforms. As the cost to create plummets, the world is going to be awash in danger and low trust — owning your inputs, and guaranteeing safety and security is something people will pay for.

Many of the most attractive opportunities are emerging within legacy service businesses, where AI-native models can deliver software-like scalability with service-level outcomes. Generative AI unlocks the rest of the organization: speech, emails, drawings, contracts, sensor feeds — everything outside rows and columns. When you mine that unstructured world, you touch far more value and your addressable market expands dramatically. Lower engineering costs mean founders can tackle neighboring problems without burning the house down. AI widens the revenue pool, slashes service cost, commoditizes shallow features, and cheapens expansion.