Stone Venture leads GENISOM AI into industrial robotics

Industrial robotics is what Stone Venture is buying a ticket into, not a lab demo. UAE-based Stone Venture, also written 磊石资本, led a Series B for China’s GENISOM AI that the company announced around 20 September 2026. Chinese coverage put the cheque at several hundred million yuan. Dollar write-ups called it tens of millions. Exact size and valuation were not published. The use of proceeds is hardware, an embodied control stack and GSD, GENISOM’s self-driving navigation layer, so machines stop touring a site and start finishing a job. That is the punch. A Gulf fund is not collecting another humanoid clip. It is underwriting a Beijing factory that already claims volume. GENISOM, 智身科技, was founded in late 2023. By June 2026 it said cumulative production of embodied robots had passed 15,000 units, with monthly output more than ten times the prior-year average. Those are company figures. They are not audited shipments, accepted robots or hours on a paid route. Treat 15,000 as a production claim until a customer names a fleet. The syndicate mixes money and parts. Beside Stone Venture, financial names cited include Hongshan Capital, Yueke Finance, Wuzhong Rongyue and Wuzhong Financial Holding. Industrial money includes Neusoft Group, Highpower Technology, Yuanjian Capital and Riying Electronics. Convertible preferred stock is how at least one capital-markets note described the paper. Stone has already used the same bridge on other Chinese machines: EngineAI, LimX Dynamics and Volant Aerotech sit in the same thesis of Gulf capital into mainland hardware before a listing. What GENISOM sells is closer to a dog than a warehouse arm. The M1 is an industrial quadruped: about 30 kg continuous walking payload, IP67, in-house CHAMP actuators, inspection and emergency work. The L1 and wheeled L1-W cover lighter patrol. A humanoid line exists in the catalogue. Deployments listed by the company span power, petrochemicals, fire, security, emergency, education and more than 500 ecosystem partners. ICRA 2026 in Vienna was the first big foreign floor. A “Gravity” programme would open 1,000 machines and about 15 million dollars of resources to developers. That programme is a plan, not a booked channel. The MENA angle is the site, not the cap table. Oil, gas, utilities and perimeter security in the Gulf look like the Chinese yards GENISOM already walks. A Dubai or Abu Dhabi lead can open doors that a Suzhou cheque cannot. The September note did not name a UAE offtaker, a pilot megawatt or a delivery quarter. Lead plus ambition is not a purchase order. If Stone’s value is a first paid patrol on a GCC plant, the Series B is a corridor. If the robots stay in Jiangsu while the fund takes a board seat, it is financial tourism with better photos than a humanoid keynote. Caveats should sit on the unit count. Production is not revenue. Payload specs from a conference booth are not a reliability curve in 50-degree heat. “Task completion” is the hard product. Navigation that only maps a corridor still needs a human with a clipboard. Export controls, after-sales parts and who certifies a quadruped next to a flare stack will decide whether industrial robotics from Beijing becomes a Gulf SKU. If GENISOM publishes a named Middle East customer and a second production number that matches invoices, Stone’s global-expansion line holds. If the next update is another capacity multiple, the B round bought a factory story. Count hours on task in the region, not units leaving a line in China.