Insider Brief
- Genesis AI unveiled a new robotics foundation model called GENE-26.5 alongside a proprietary robotic hand, glove-based data collection system and simulation platform designed to improve how robots learn complex physical tasks from human behavior.
- The company demonstrated the system performing coordinated manipulation tasks including cooking, laboratory procedures, wire harnessing, Rubik’s Cube solving, multi-object sorting and piano playing, while also introducing a robotic hand and tactile glove system intended to capture large-scale human movement and interaction data for robot training.
- The release marks the company’s evolution into a full-stack robotics company spanning AI models, simulation, hardware and data infrastructure, following its $105 million funding round led by Eclipse and Khosla Ventures and recent commercialization hires including former Amazon executive Vivian Sun.
Genesis AI unveiled a new robotics foundation model called GENE-26.5 alongside a proprietary robotic hand and data collection system designed to improve how robots learn complex physical tasks from human behavior.
The system is intended to address the difficulty of collecting large amounts of usable training data for physical tasks, which the company pointed out is one of the biggest limitations in robotics. Genesis AI said its approach combines AI models, simulation software and hardware designed to mirror the movements of a human hand.
“The brain and hand are the two most valuable and complex pieces of robotics, and today we are presenting the industry’s most advanced versions of both,” co-founder and CEO Zhou Xian said in the announcement. “For the first time ever, we’re enabling robots to do what only human hands could, and do it reliably, at scale.”
What is GENE-26.5?
GENE-26.5 is built to help robots perform longer and more complex manipulation tasks across changing environments. The company released a demonstration video showing the system carrying out a range of coordinated activities, including:
- Preparing meals involving multi-step cooking tasks
- Making smoothies using coordinated two-handed movements
- Performing laboratory procedures such as pipetting and liquid transfer
- Organizing and securing wire bundles
- Solving a Rubik’s Cube through continuous hand manipulation
- Sorting multiple objects simultaneously with one hand
- Playing piano sequences requiring rapid finger coordination
Genesis AI said the model relies heavily on a new data collection framework centered around a robotic hand designed to match the structure and movement of a human hand. A connected glove system equipped with tactile sensors allows humans to perform tasks while the system captures movement and interaction data that can later be transferred to robotic systems.
According to the company, the glove system is designed to lower the cost and complexity of robotics data collection compared with traditional teleoperation systems. Genesis AI said it plans to deploy the gloves with partners in real-world workplaces to gather additional training data from everyday human activity.
The company also said its broader training pipeline incorporates egocentric video captured from wearable cameras as well as internet video sources to help robots learn how humans interact with objects and environments.
Alongside the robotics model, Genesis AI introduced a simulation platform intended to reduce the gap between virtual training and real-world robotic performance. The company said the system is designed to allow robots to be trained and tested extensively in simulation before deployment in physical environments.
Evolution of Genesis
In July, Genesis AI emerged from stealth with $105 million in funding to develop a universal robotics foundation model aimed at automating physical labor across industries. The round was co-led by Eclipse and Khosla Ventures, with participation from Bpifrance, HSG, Eric Schmidt and Xavier Niel.
With GENE-26.5, it’s now calling its evolved into a full-stack robotics company and indicated it plans to reveal its first general-purpose robot in the future. In March, Genesis AI appointed former Amazon automated driving strategy executive Vivian Sun as vice president of commercial and strategy to lead commercialization, partnerships and go-to-market efforts.
“At Genesis, we believe winning in robotics requires excellence at every level,” said Theophile Gervet, co-Founder and President. “That’s why we’re obsessed with innovating across the full-stack, from AI to hardware. By controlling every layer, we can build a cohesive system and solve the problem holistically. Our approach gives us a huge competitive advantage by harnessing unprecedented amounts of data as that ultimately defines what foundation models can achieve.”
Image credit: Genesis AI