Genesis AI Goes Full Stack with Humanoid Robotic Hands and New Model
The Khosla-backed startup is moving beyond software to build its own dexterity-focused hardware and data-collection gloves.
Primary source: TechCrunch AI. Full source links and update notes are below.
Fast summary
Start here
- Genesis AI unveiled the GENE-26.5 model alongside proprietary robotic hands designed to mimic human size and shape.
- The company is adopting a full-stack approach to reduce the embodiment gap between AI training and physical performance.
- A new sensor-loaded glove allows the company to collect high-fidelity human movement data from workers in labs and factories.
What happened
Genesis AI, a robotics startup supported by Khosla Ventures, has transitioned to a full-stack development model. The company showcased its first foundational model, GENE-26.5, in coordination with custom-designed robotic hands. Developed in partnership with Wuji Tech, the hardware features a human-like form factor intended to bridge the gap between AI simulations and real-world physical manipulation.
What's new in this update
The company has introduced a lightweight, sensor-loaded data collection glove designed to be worn by human workers. Unlike traditional, bulky motion-capture devices, this glove is intended to be used during daily tasks in industries such as pharmaceuticals and manufacturing. By capturing 'egocentric video data' and precise hand movements, Genesis AI aims to generate the massive datasets required to train robots in complex, long-horizon tasks like lab work or cooking.
Key details
The GENE-26.5 model was demonstrated performing a variety of tasks, including cracking eggs, slicing tomatoes, playing the piano, and solving a Rubik's cube. CEO Zhou Xian stated that while model intelligence is the primary goal, controlling the hardware is necessary to achieve high-fidelity interaction. The startup's hardware strategy relies on five-fingered hands rather than the two-finger grippers common in the industry, which Genesis claims allows for more natural data collection from human demonstrations.
Background and context
Genesis AI entered the market with a $105 million seed round aimed at building foundational AI for robotics. It operates in an increasingly competitive sector alongside firms like Physical Intelligence and Skild AI. The company's leadership includes CEO Zhou Xian and President Théophile Gervet, a former research scientist at Mistral AI. Their strategy focuses on solving the 'embodiment gap,' the technical challenge of translating digital intelligence into effective physical movement.
What to watch next
Genesis AI is currently in discussions with potential industrial customers to deploy its data-collection gloves. A critical factor for the company's growth will be the willingness of human workers to wear devices that train the very models intended to automate their roles. The company expects frequent iterations of its GENE model, utilizing a proprietary simulation environment to speed up the evaluation and training process.
Why it matters
The transition to full-stack robotics suggests that foundational models alone are insufficient for complex physical tasks, requiring specialized hardware to achieve human-level dexterity.
Read next
Follow this story through the topic hub, more ai coverage, and the latest updates.
Weekly briefing
Get the week's key developments in one concise email.
Get a fast catch-up on the biggest stories, the context behind them, and the links worth your time.
Cadence
Weekly, for a quick catch-up
Coverage
AI, business, world, security, sports
Format
Clear takeaways and useful context
Request the briefing
Leave your email to open a prepared request and get on the list for the weekly briefing.
Author
See who assembled this story and follow more of their work.
Sources and methodology