Fueling founders, building ventures.

A VC bets on startups that already exist. We manufacture them. The studio originates the idea, breaks the tech barriers with an in-house AI team, then recruits founders and funds the company from day 0 — acting as a co-founder, not a spectator. Incubators host. Accelerators polish. VCs write checks. We build.

Two profiles. Founding CEOs: commercially fearless operators who can sell a vision to a Fortune 500 buyer and a first hire in the same afternoon. Founding CTOs: engineers and researchers who’d rather ship deep tech than publish about it. In both cases: people who want to run a company, not join one.

Evidence you make things happen at an unreasonable rate: sales or operating experience in a serious industry, a network you’ve actually used, and the appetite to own a P&L from day one. Domain expertise in our verticals — health, finance, retail, cyber, commerce infrastructure — is a strong plus. A previous startup, even a failed one, counts double.

Depth, not buzzwords. You’ve trained, fine-tuned or deployed real models in production — or built systems that don’t fall over when customers show up. You can lead a small elite team, argue with our AI engineers as equals, and you think of research papers as raw material, not decoration.

No — and that’s the point. Ventures are originated and de-risked in-house: by the time you step in, the pain point is validated, the prototype exists and the first design partners are waiting. If you do bring your own idea and it survives our inception phase, even better. Both doors lead to the same water.

The foundry runs in person — the intensity doesn’t survive video calls. Expect to be with the school in our hubs during the build phases, with the US and our other bridges on the itinerary as the venture goes to market. If a relocation is what stands between you and a company of your own, we’ll help make it a detail, not a barrier.

A $200k SAFE from day 0, a validated prototype, an in-house AI team three days a week, advisors with skin in the game, market access through LP networks, and a Delaware Inc. fast track. And you keep the majority. Full terms are shared during the selection process.

Fifteen months from spark to scale, through four phases: Inception, Prototype, PMF, Acceleration. Ventures with deep R&D can loop the prototyping phase. The team is guided the whole way — senior operators, embedded AI engineers, weekly cadence. Speed is a feature, not a compromise.

An accelerator gives you a curriculum and a t-shirt for a company you already built. We build the company with you: the idea, the tech, the first data, the first customers and the first check all come from inside the foundry. You don’t attend the program — you’re a co-founder in it. And we stay at the table for 15 months, not 12 weeks.

Two batches a year — each new manta with its own ocean. Before the foundry, the same team shipped DeepEcho, Talaty and ToumAI. The method is not theoretical.

Batch #3 is forming now — applications close September 10. No 47-slide deck: the application is five sharp questions and a 1-minute video of you talking. That’s enough for us to know if we should talk. If it’s a yes, expect the first conversation within days, not weeks.

Two doors. Join as an LP — windows open ahead of each batch, from smart-money angels to institutional partners. Or co-invest at demo day through club deals with allied early-stage VCs. Detailed terms and performance travel by voice, not by website: dare@34stud.io.

Mantas have the largest brain-to-body ratio of any fish. They cross oceans, they never stop swimming, and they filter enormous volumes to keep only what feeds them. If you have a better metaphor for a deeptech studio, send it to dare@34stud.io.