Rotating Data
Home/Deal Room/Figure commits $3.5B to Nscale AI compute

Cloud capacity · Evidence brief

Figure commits $3.5B to Nscale AI compute

Figure + Nscale + NVIDIA

Primary sourcePress release only2026-09-03
Deal structureAI cloud compute commitment
GeographyBarstow, Texas, United States
Infrastructure layersCloud capacity + chips / compute
Disclosed scale$3.5B initial compute commitment; up to 100,000 Vera Rubin GPUs

Why it matters

The strategic read

Links physical-AI model training demand to a large next-generation GPU cloud deployment. The initial commitment is distinct from the intent to expand above $6B, and GPU delivery remains a future deployment target.

Source-supported terms

What was disclosed

  1. 01

    Initial $3.5 billion compute commitment, with intent to scale to over $6 billion; the higher figure is not the initial contracted amount.

  2. 02

    Up to 100,000 NVIDIA Vera Rubin GPUs, targeted for initial deployment starting in the second half of 2027 in Barstow, Texas.

  3. 03

    Compute supports training Figure’s Helix models; Nscale is also making a strategic investment in Figure, with no investment amount disclosed.

  4. 04

    The parties will explore humanoids in Nscale’s supply chain; this is not a disclosed robot purchase commitment.

Disclosure boundary

What is known—and what is not

Issuer-announced initial compute commitment. Full contract, pricing, duration and termination terms are not public. Initial deployments are targeted for H2 2027, not currently operational capacity. Nscale investment amount and the conditions for expansion above $6B are undisclosed.