Approach
We invest where technical depth creates durable advantage.
We are a small, hands-on ownership and operating platform. We do not run a large book of names or chase momentum. We take a limited number of positions we can understand at the level of the code, the economics, and the hardware — and we stay involved.
The operating model
Four principles, applied consistently.
These are not slogans. They describe how a decision actually gets made here — what we say yes to, what we walk away from, and what we do after the wire clears.
Diligence is technical first
Before economics, before narrative, we ask whether the thing works and whether it can keep working under load. We read the protocol, the contracts, and the code. We model the unit economics of compute and mining down to energy and hardware. A convincing story with no technical floor is a pass.
This is why we tend to concentrate in a few domains rather than spread across many — we can only hold real conviction where we can do the engineering ourselves.
Capital is deployed selectively
We would rather do a handful of things well than many things adequately. Each position has to earn its place against the others — on merit, not on optionality. When something no longer clears that bar, we are willing to say so and move on.
We back builders who value the same discipline. If speed is your only edge, we are probably not the right partner.
Ownership means operating
Where we hold and operate infrastructure — compute, mining, and the systems around them — we treat it as an operating responsibility, not a line on a page. Uptime, efficiency, and cost are engineering problems we own alongside the teams doing the work.
Being close to operations is also how our diligence stays honest: it is hard to fool yourself about a system you have to run.
The horizon is long, the end is planned
Durable advantage takes time to build and time to prove. We hold through cycles when the thesis holds. We also plan for the ending from the beginning — a venture that is sold, matured, or wound down is a normal outcome, not a failure, and clarity about that keeps everyone honest.
Our record reflects this: some positions are held and operating today; several have reached a clean exit.
If this is how you build, talk to us.
We are most useful to teams working on hard technical problems in compute, networks, and markets — and least useful to everyone else. A short, specific note is the best start.