We are generalists inside a narrow band. Each area has a partner who has worked in it, which is the only reason we claim to understand it.
Storage, compute, observability, and the layers underneath. We look for teams who have run the systems they are now trying to fix, and who can explain the cost curve without a slide.
We are less interested in models than in the work around them: the data access, the evaluation, the last mile into a workflow somebody is paid to perform. That is where most of the value settles.
Payments, risk, and the plumbing of financial operations. The good businesses here tend to look boring for two years and then compound quietly for ten.
Identity, access, and evidence. Buyers are sceptical and slow, which suits founders who understand the threat model better than the vendors already in the room.
Software that moves electrons, materials, or measurement. We back companies with a physical constraint at the centre, because those constraints keep competitors honest.
Care delivery, clinical operations, and research infrastructure. We are patient here. Sales cycles are long and the regulatory floor is real, so we underwrite for it from the first cheque.
Whether inference costs falling by an order of magnitude changes which vertical software businesses are worth starting, or just makes the existing ones cheaper to run.
Whether grid flexibility becomes a software market or stays a hardware one with software attached. We have a view. We are not certain about it.
Whether the compliance burden arriving in Europe creates a durable class of companies or a temporary consulting boom.
Send a short note. We read every one and reply within five working days, whatever the answer.