Fresh off its breakup with OpenAI, Figure is showcasing its self-developed AI model that the Silicon Valley startup says paves the way for its autonomous humanoids to enter homes at scale.
Figure says the development means its robots can reason like humans.
When Figure founder Brett Adcock announced his decision to end the robotics firm’s partnership with OpenAI, he said it was because their capabilities couldn’t provide the intelligence needed to deploy AI robots at scale.
Two weeks later, Figure just released details on what it calls a breakthrough for embodied intelligence called Helix. Before the announcement, Adcock teased Helix as the most significant AI update in the company’s 3-year history.
The unveiling comes as Figure is reportedly in advanced discussions to raise $1.5 billion in Series C funding, which would value the company at nearly $40 billion. If accurate, the amount is a 1,400 percent increase from its 2024 valuation of $2.6 billion.
According to Figure, Helix is a vision language action model tailored for humanoid robotics control. The startup calls it the first VLA model to enable high-rate control of the entire upper body, including wrists, fingers, torso, and head.
With less than 500 hours of training, robots powered by Helix can pick up virtually any object by voice command. Figure says the model is also the first of its kind to run on two robots simultaneously.
In a demo video, two Figure 02 humanoid robots are shown collaborating as they put away groceries. Natural language prompts drive the actions.
While employing two high-tech robots to slowly put away groceries is overkill, Figure says its breakthrough will accelerate their development so they can become useful in household settings. The company eventually plans to offer its humanoids to the consumer market. They’re aiming for a price point similar to a modest vehicle, similar to other perceived leaders in humanoid robotics like Tesla and 1X Technologies.
Since Helix runs on one neural network for all behaviors, there’s no need for task-specific fine-tuning. It runs entirely on embedded low-power GPUs, making it more feasible for real-world applications.
Figure says the technology is essential to bring AI-powered humanoid robots into homes. Unlike industrial settings, where Figure 02 robots are currently training extensively, households are full of unpredictable and diverse objects.
On the X social media platform, Adcock said 2025 will be a pivotal year for Figure as it ships more robots and tackles home applications.
Figure proposes that combining intelligence from collected robot behavior with the Helix model will trigger exponential growth. They provided this graph to illustrate, though it was criticized by some as being based on vibes rather than actual metrics.
Sunnyvale-based Figure is developing its next-generation robot, presumably called the Figure 03, that’s optimized for manufacturing at scale. Adcock recently said Figure has signed on its second commercial customer, without naming the company. He said he sees a path for shipping up to 100,000 humanoid robots over the next four years.
Figure is seen as perhaps the biggest threat to Tesla, which plans to produce up to 10,000 Optimus humanoid robots this year. While Tesla’s production target is much larger than Figure’s, Elon Musk has admitted it’s highly aspirational.
Figure and Tesla face stiffening competition, with Google investing heavily in Apptronik and tech giants like Apple, OpenAI, and Meta reportedly planning to enter the market.