Figure's Nscale Compute Deal Faces Infrastructure Hurdles

Recap

Source: Humanoids Daily, report of Sept. 30, 2026

According to the report, Figure’s multibillion-dollar compute agreement with Nscale is part of a broader infrastructure expansion that remains unfinished in terms of financing and construction. The article, citing new reporting from The Information, clarifies that while contracting for future GPU capacity is one step, actually building, powering, and equipping the data centers is a separate challenge. The report explicitly states that it does not establish that Figure’s specific allocation has been delayed.

The infrastructure is intended to support the training of Helix, the AI system behind Figure’s humanoids. Figure plans to access up to 100,000 GPUs on NVIDIA’s Vera Rubin platform through Nscale, with initial deployment in Barstow targeted for the second half of 2027. Nscale is also making a strategic investment in Figure, and the companies plan to explore using humanoids in Nscale’s supply chain. Brett Adcock, Figure’s CEO, has argued that scaling robot intelligence will eventually require hundreds of billions of dollars, a thesis that helps explain the company’s appetite for such a large compute agreement.

The Information’s analysis of Nscale’s IPO plans describes a company whose contracted business is far larger than its current operations. Nscale currently operates five data centers delivering roughly half a gigawatt of capacity, with another 12 contracted to bring the total to 10 gigawatts. The publication reports $24 billion in technology-equipment purchase commitments payable this year and next, along with substantial borrowing since June, including a recently announced $3.36 billion convertible-note raise. These figures describe Nscale’s broader expansion rather than the specific funding status of Figure’s project.

Context

Figure has previously described data and compute as primary constraints on the development of its Helix AI system. The company’s ambition to train robots at a scale comparable to, or exceeding, large language models requires massive computational resources. Nscale, a major player in the GPU cloud market, has attracted significant backing, including a $2 billion Series C round at a $14.6 billion valuation with participation from NVIDIA. However, the gap between contracted capacity and operational reality is a common challenge in the data center industry, where lead times for construction, power, and equipment are often lengthy.

Robot's take

The significance of this report lies in highlighting the physical and financial dependencies behind AI training promises. While Figure’s agreement secures future access to GPUs, the actual delivery depends on Nscale’s ability to finance and construct the necessary infrastructure on schedule. It is not yet clear whether Figure’s specific allocation is fully financed or if there are alternative capacity arrangements in place. For readers following Figure’s progress, the key milestones to watch are financing and construction progress at Barstow, power availability, and delivery against the initial second-half-of-2027 target. The report makes these dependencies more visible but does not indicate that Figure’s compute deal has failed.