Vesoma exits stealth with 60+ staff and an ex-DeepMind AI chief, nine months after founding

Vesoma, a humanoid robot company based in Munich and Limassol, Cyprus, came out of stealth on September 22–23, 2026 with more than 60 employees and over 3,500 square metres of labs and offices. The company began work in December 2025, says its first prototype walked within six months, and is now building the first units of its product robot, Vesoma 1.

Why it matters: Europe has few humanoid makers with in-house hardware, controls and AI under one roof. Vesoma has assembled a team in nine months that covers all three, though it has shown a prototype, not a product.

Who founded Vesoma?

Vesoma was co-founded by chief executive Nikolai Ensslen and chief strategy officer Peter Skoromnyi. Ensslen previously co-founded Synapticon, a German maker of motion control and functional safety systems, the drives-and-safety layer that sits inside robot joints. Skoromnyi co-founded Easybrain, a mobile puzzle-games company. Martin Riedmiller, formerly a research director at Google DeepMind and a long-time reinforcement learning researcher, is chief AI officer. Design is led by Marcelo Gutierrez, who previously worked on Tesla Optimus, Apptronik Apollo and Cartwheel’s Yogi, according to Dealroom and Humanoids Daily.

The mix is the story. Most European humanoid efforts come either from AI researchers or from industrial suppliers. Vesoma has leaders from both, plus consumer-software management experience.

What is the Vesoma 1 humanoid?

Vesoma 1 is the company’s planned product humanoid, and its first units are being built now. The company has not given specifications, a price or a delivery date. Humanoids Daily notes that the published images of Vesoma 1 are design visualisations, not finished robots. What has been shown working is “Kyle,” an alpha development platform first revealed on September 14, which walked within six months of the company’s start using a learned locomotion policy.

Initial target markets are manufacturing, logistics and warehousing, according to Dealroom. The company’s own announcement adds care work and other labour-intensive sectors facing staff shortages in Europe as longer-term uses.

How does Vesoma want its robots to learn?

By interacting with the physical world rather than relying mainly on human demonstrations. The company describes self-improving agents grounded in physics-based learning, an approach that fits Riedmiller’s reinforcement learning background. Much of the field currently trains robots on teleoperated demonstrations, which are expensive to collect at scale.

The claim is a direction, not a result. Vesoma has not published manipulation success rates, task results or data showing an advantage over imitation-learning systems. Its one demonstrated capability is walking, which learned policies already handle across the industry.

Will Vesoma robots be used by the military?

No, according to the company. Vesoma says it rules out weapons, military and surveillance uses of its robots, directly or indirectly. That is a stronger statement than most humanoid makers have made, and it arrives as defence ministries, including Japan’s, begin studying humanoids for non-combat roles; see Japan’s SDF humanoid study.

The funding question is open. Sixty staff and 3,500 square metres in two countries imply significant backing, but Vesoma has not named investors or amounts. For comparison, the UK’s Humanoid raised a $152 million Series A at a $1.35 billion valuation in July, with Bosch and Schaeffler taking part; see Europe’s first pure-play humanoid unicorn. Vesoma will need a disclosed round, or a named industrial partner, before its production plans can be judged on the same terms.