Investing
I look for a clear outcome and a path into production. I invest personally, usually from pre-seed through Series A, in AI software and infrastructure companies. I occasionally invest later when I believe a company can define its category.
01 · My investment thesis
A strong AI demo is easy to admire and hard to deploy. The companies I look for make it reliable inside a real workflow, with clear ownership, controls, and measurable outcomes.
The best companies I see start with one expensive problem, solve it end to end, and earn the right to own more of the workflow.
Models will improve and become widely available. The lasting advantage will sit in the workflow, proprietary context, evaluation loop, permissions, or distribution.
The place where valuable work gets done, together with the proprietary knowledge required to perform it well.
The feedback, permissions, and controls that measure outcomes, improve performance, and let agents act safely.
The advantage that turns initial adoption into durable expansion.
02 · Where I invest
These are the four areas where I spend most of my time.
Products that redesign how work gets done instead of adding an AI feature to an existing workflow, especially where agents can own meaningful outcomes.
Especially: Software creation · Enterprise knowledge · Talent · Adaptive interfaces
The infrastructure capable agents need to understand context, access systems, transact, improve, and act safely in the real world.
Especially: Identity + payments · Permissions · Security · Evals + observability
Selected investments
Exceptional teams expanding what AI can understand and discover, especially closed-loop systems that generate new knowledge and evaluate outcomes objectively.
Especially: Multimodal intelligence · Autonomous research · Scientific experimentation
Selected investments
Selective investments in technical infrastructure that makes previously invisible operations measurable, actionable, and deeply embedded in how customers operate.
Especially: Operational intelligence · Systems of record · Physical operations
03 · How I help
I help founders sharpen the initial product, structure early deployments, and turn hands-on customer work into something repeatable.
Where appropriate, I'm also happy to make targeted introductions to relevant operators, customers, and investors.
Pressure-test the wedge, identify where AI creates real leverage, and decide what should stay human-led versus be delegated to agents.
Work through evaluations, reliability, permissions, security, integrations, workflow design, and the controls required to ship with confidence.
Translate technical capability into a clear use case, ROI case, and enterprise narrative; sharpen the ICP, structure pilots, and navigate long sales cycles, procurement, security reviews, and expansion.
Bring a founder’s perspective to product, hiring, fundraising, and the shift from early customer work to a repeatable product and company.
Interested in partnering?
I'd love to hear what you're building, why now, what you've learned from users, and the details of the round.
Share what you're buildingI invest personally. The views here are my own and do not represent OpenAI.
Connect on LinkedIn