Figure Bets on Data, Then Compute: Index Launch and Nscale’s 100K-GPU Deal

The next bottleneck for building a general-purpose humanoid robot isn’t the model — it’s data and compute. The physical-world data robots need doesn’t exist on the internet, and collecting it at scale from the real world takes as much raw compute as it does data.

Figure AI, the humanoid robotics startup that raised a $1 billion Series C at a $39 billion valuation last September — after reportedly negotiating a $1.5 billion round earlier that year — made two moves in ten days that each target one of those bottlenecks.

Index: Coming Out of Stealth With a Human-Data Network

On August 25, Figure came out of stealth with Index, a data-collection app it had been quietly building for four months. The premise: general-purpose robots need video of humans doing real tasks in the real world, not more internet text and images — and buying that data from vendors couldn’t hit the throughput, diversity, or quality bar Figure’s Helix model requires, so the company built its own pipeline.

While in stealth, Index crossed 264,000 downloads across 108 countries and more than 44,000 weekly active users, who have uploaded over 16 million videos. The app now processes 30 minutes of video every second — the equivalent of 4.9 years of human work uploaded daily. Figure has paid out $15 million to the “Creators” contributing that data so far, and plans to spend more than $1 billion on data and compute over the next 12 months to scale collection 100x.

Per 1,000 hours collected, Index contains 373 unique tasks, 1,146 unique manipulated objects, and 116 unique environments — diversity the company says comes directly from its Creators, who record everyday chores at home as well as work inside logistics centers, restaurants, factories, and offices. Submissions run through a five-stage pipeline: automated filtering, fraud review, deduplication, task-quota rebalancing, and hierarchical captioning.

The Nscale Deal: Up to 100,000 GPUs

A week later, on September 3, Figure announced a strategic partnership with GPU cloud provider Nscale to deploy up to 100,000 GPUs on NVIDIA’s next-generation Vera Rubin platform, with initial deployment targeted for the second half of 2027 in Barstow, Texas. The deal represents an initial compute commitment of $3.5 billion, with intent to scale to over $6 billion. Nscale is also making a strategic investment in Figure as part of the agreement — the amount wasn’t disclosed — and the two companies will explore scaling Nscale’s supply chain to support humanoid production.

Notably, Figure said Index is now generating 35 minutes of data every second — up from the 30-minutes-per-second figure it cited a week earlier — while framing compute as the next constraint: “data alone cannot solve this problem.”

“More Capable With More Data and Compute”

Figure founder and CEO Brett Adcock said Helix, like every learned system, gets more capable with more data and compute, and that the Nscale partnership provides the compute runway needed to make that vision real. NVIDIA founder and CEO Jensen Huang called humanoid robots a major new industry extending physical AI into a world built for people, describing a “robotics flywheel” — training Figure’s models on Vera Rubin through Nscale’s AI cloud, validating them in NVIDIA Isaac Sim, and deploying them on NVIDIA GPUs inside Figure’s robots. Nscale founder and CEO Josh Payne said that after strong growth in inference and agentic AI, Figure is pushing AI’s frontier into the physical world, and that Nscale is proud to help build the future of AI robotics together.

Where Figure Sits Competitively

Figure competes with a growing field of general-purpose humanoid startups, including Apptronik, 1X Technologies, and Physical Intelligence, each pursuing its own data and model strategy. Nscale, meanwhile, sits in the increasingly crowded “neocloud” GPU-cloud market alongside players competing to supply compute for both AI and robotics workloads.