Build.
We only build where we have an unfair advantage. Vertical AI in industries we know cold — domain research, product, GTM all in-house. No tourist bets.

A vertical AI venture studio
We build, invest, and advise AI companies that go deep, not broad.
Origin
The peaks
nobody is climbing.
The name comes from Gangkhar Puensum. 7,570 metres, in Bhutan. The highest unclimbed peak on earth — forbidden to climb since 1994. A peak everyone else has ruled out. That’s the whole thesis: the most defensible AI companies of the next decade won’t have the biggest markets. They’ll have the deepest ones.
Routes
Three things we do. One thing we promise.
We only build where we have an unfair advantage. Vertical AI in industries we know cold — domain research, product, GTM all in-house. No tourist bets.
A small number of conviction bets per year, in founders already deep in their industry. Early stage. Few enough that every check matters.
Hands-on guidance for founders who want a second operator in the room. Product, pricing, GTM, hiring, fundraising, the hard calls. We show up for the work, not the board meeting.
No fund cycle, no five-year clock. GP75 is patient capital with operator roots. We back companies we can spend a decade on, and exit when the company is ready, not when a fund is.
Portfolio
Companies we build, invest in, and advise.
The next route
We partner with a small number of exceptional founders building deep in a single industry. If you’re climbing something serious, we want to hear about it. Capital, conviction, or a co-climber.