Figure AI has just unveiled Helix, a groundbreaking Vision-Language-Action (VLA) robot that’s designed to eventually be deployed in the home. It looks seriously impressive in its official teaser video.
The announcement comes as the company’s valuation recently rocketed to $40 billion. A pending $1.5 billion funding round led by Align Ventures and Parkway Venture Capital suggests the big players are taking notice.
Two Brains Are Better Than One
The system operates through a unique “two-brain” approach:
- System 1: A faster 80M parameter transformer that controls precise movements at 200Hz for real-time reactions
- System 2: A sophisticated 7B parameter vision-language model that handles high-level understanding and reasoning, processing information at 7-9Hz
These systems communicate through a “latent vector” that translates understanding into action
On Linkedin Brett Adcock, Figure’s founder and CEO said:
To bring robots into homes, we need a step change in capabilities. Helix can generalise to virtually any household item… Like a human, Helix understands speech, reasons through problems, and can grasp any object – all without needing training or code.
Human-esque Fluidity
The technical achievements of Helix and the subsequent movement of its robots are very impressive. The system provides comprehensive upper body control so precise that if it were wearing clothes you’d be forgiven for mistaking it for a person (albeit a fairly slow-moving one).
It can coordinate 35 different movement points simultaneously, but what’s truly revolutionary is how these robots work together. In demonstrations, two Helix-powered robots coordinate to unpack groceries, passing items between them and organising them in drawers and refrigerators – all with objects they’ve never seen before.
You can simply tell them “hand the cereal box to the other robot” or “put the milk in the fridge,” and they figure out the details themselves. This is achieved without any special training for teamwork – both robots run on identical software, much like how two humans can naturally work together without specific coordination training.
The system was trained on just 500 hours of human demonstrations – a fraction of what traditional robots need to learn even simple tasks. This efficiency, combined with its ability to run on standard hardware, suggests Figure has cracked a fundamental challenge in robotics: teaching machines to adapt to new situations as easily as humans do.
Money Talks, Robots Listen
While most startups are happy to hit unicorn status, Figure AI has somehow managed to become something altogether bigger. Investment from Microsoft, OpenAI, Nvidia, and Jeff Bezos (who apparently knows a thing or two about automation) has helped fuel the company’s meteoric rise from a mere $2.6 billion valuation last year to almost $40 billion now. They’re planning to deliver 100,000 robots over the next four years too.
These humanoids have been stated as being designed for the homes of the future, but there might be business applications too. Manufacturing floors could see robots doing precision work better than a Swiss watchmaker. Warehouses might finally have someone who doesn’t complain about inventory counts. Healthcare could get assistance that never needs a coffee break, and retail stores might actually keep their shelves stocked!
The Future of Robots
Morgan Stanley analysts note that “this is not a ‘winner take-all’ market.” The future apparently has room for multiple robot citizens, which is somewhat reassuring.
As the technology continues to evolve, we’re left wondering what’s next. Will these robots start giving TED talks? Write their own press releases? For now, Helix’s combination of flexibility, precision, and ease of use suggests we’re entering an era where science fiction is starting to look remarkably unimaginative.
