Nscale has committed an initial $3.5 billion of compute to Figure, with stated intent to exceed $6 billion, for up to 100,000 Nvidia Vera Rubin GPUs targeted for Barstow, Texas, from the second half of 2027. Nscale will become a Figure shareholder and its preferred compute provider, making the arrangement more strat...
Published byEdited with GPT-5.6 TerraImages generated with GPT Image 2
Research answer

Create a landscape editorial hero image for this Studio Global article: What are the terms, scale, technology, timeline, and strategic significance of Nscale’s multi-year partnership with Figure—including its ini. Article summary: Nscale’s Figure agreement is a strategic, multi-year physical-AI compute partnership rather than a conventional cloud lease: Nscale commits an initial $3.5 billion of capacity, intends to expand it beyond $6 billion, wil. Topic tags: general, general web, user generated, news. Style: premium digital editorial illustration, source-backed research mood, clean composition, high detail, modern web publication hero. Use reference image context only for broad subject, composition, and topical grounding; do not copy the exact image. Avoid: logos, brand marks, copyrighted characters, real person likenesses, fake screenshots, UI text, readable text, watermarks, charts w
Nscale’s partnership with humanoid-robotics company Figure is a long-term infrastructure and investment arrangement designed to supply the computing capacity behind Figure’s next generation of Helix AI models and robots. Nscale announced an initial commitment of $3.5 billion of compute, with an intent to scale the relationship to more than $6 billion. The partnership could deploy up to 100,000 Nvidia GPUs based on the Vera Rubin platform. 20
The headline figures are substantial, but the central qualifier matters: the first GPUs are targeted to begin deployment in the second half of 2027 at Nscale’s Barstow, Texas site. This is a planned capacity buildout, not compute that Figure can use immediately. 20
| Element | What has been announced |
|---|---|
| Initial compute commitment | $3.5 billion |
| Potential scale | Intent to grow beyond $6 billion, rather than a disclosed final contracted total |
| Hardware | Nvidia Vera Rubin platform, with potential for up to 100,000 GPUs |
| Initial deployment | Targeted for the second half of 2027 |
| Location | Barstow, Texas |
| Commercial relationship | Nscale becomes a Figure shareholder and preferred compute provider |
| Primary use | Training and supporting Figure’s next-generation Helix models and humanoid-robot deployment |
Nscale and Figure have not disclosed the size of Nscale’s equity stake, detailed capacity-pricing terms, reservation conditions or any take-or-pay provisions. That means the announced $3.5 billion should be understood as an initial compute commitment, while the amount above $6 billion remains a stated intention to expand the partnership. 20
38
The arrangement combines three roles for Nscale: infrastructure supplier, equity investor and Figure’s preferred compute provider. Nscale’s investment in Figure is undisclosed, but the structure aligns the cloud provider more closely with the robot maker’s commercial progress than a conventional customer-vendor agreement would. 20
38
For Figure, preferred-provider status creates a path to a very large dedicated AI-computing deployment. For Nscale, an equity position gives it exposure to Figure’s potential upside while it supplies the compute needed to train and deploy Figure’s models. The exact financial allocation between the cloud commitment and the equity investment has not been made public. 20
38
Nscale says the infrastructure will power the next generation of Figure’s Helix models. Helix is intended to support humanoid robotics, so the compute is relevant not only to model training but also to the broader process of developing and operating physical-AI systems. 20
40
The partnership’s potential scale—up to 100,000 Vera Rubin GPUs—suggests a much larger infrastructure ambition than a single training run. Humanoid robotics can require repeated model training, simulation and evaluation as a developer moves from lab demonstrations to a fleet of deployed machines. Still, the companies have not published a workload split between training, simulation and inference, so the precise use of the capacity is not known. 20
The strategic case is a robotics feedback loop: simulated environments can help train and test robotic policies, deployments can generate real-world operational data, and those results can inform later simulation and retraining. Nvidia describes related physical-AI workflows using simulation and digital twins to train and test robots in realistic environments. 42
For Figure, a large cloud deployment could support this kind of continuous improvement cycle for Helix. The important distinction is that the Nscale and Figure announcement confirms the planned compute partnership and its scale, but does not disclose the exact simulation pipeline or the allocation of GPUs to each stage. 20
The companies are also exploring use of Figure humanoid robots in Nscale’s supply chain. That is significant because it could make Nscale more than Figure’s cloud provider and investor: it could become a potential customer and operational testing environment. 23
24
However, this is exploratory. Neither company has announced a production deployment of Figure robots in Nscale data centers or supply-chain facilities. 23
24
The Figure partnership follows reports of a much larger Nscale agreement involving Anthropic. Bloomberg-reported terms describe Anthropic agreeing to spend about $45 billion over six years for roughly 460 megawatts of computing capacity at Nscale’s Monarch Compute Campus in West Virginia, using Nvidia Vera Rubin systems. 3
6
The Anthropic arrangement should be treated differently from the Figure announcement. Nscale publicly announced the Figure partnership and its core terms. By contrast, the Anthropic contract has been reported by multiple outlets, but detailed terms have not been publicly released by the parties. 20
3
The two arrangements point to different types of AI demand:
The Figure deal is smaller in announced dollar value, but it broadens Nscale’s exposure from frontier language-model workloads into robotics—a segment where demand may extend across model training, simulation and operation of deployed systems.
The partnership’s size does not remove its delivery risks. The initial deployment is targeted for the second half of 2027, so the practical value of the agreement depends on the timely completion and operation of the Barstow site, availability of Vera Rubin hardware and supporting networking and power infrastructure. 20
It also depends on Figure turning computing capacity into commercially useful humanoid robots. The announced agreement establishes the infrastructure ambition and strategic alignment; it does not guarantee that the full GPU deployment will occur or that the planned expansion beyond $6 billion will be reached. 20
Nscale’s Figure partnership shows how physical AI is becoming an infrastructure story as well as a robotics story. A humanoid-robot developer is seeking capacity on a scale associated with major AI-model builders, while the cloud provider is using a mix of dedicated compute, preferred-provider status and equity ownership to deepen the relationship. 20
38
The clearest takeaway is not that 100,000 GPUs are already powering Figure robots. They are not: deployment is targeted to start in late 2027. Instead, the agreement is a sizable forward commitment to the idea that training, simulating and improving humanoid robots could become a major long-duration source of cloud-compute demand. 20
Studio Global AI
This page includes a source-backed answer you can continue inside Studio Global.
Nscale has committed an initial $3.5 billion of compute to Figure, with stated intent to exceed $6 billion, for up to 100,000 Nvidia Vera Rubin GPUs targeted for Barstow, Texas, from the second half of 2027.
Nscale has committed an initial $3.5 billion of compute to Figure, with stated intent to exceed $6 billion, for up to 100,000 Nvidia Vera Rubin GPUs targeted for Barstow, Texas, from the second half of 2027. Nscale will become a Figure shareholder and its preferred compute provider, making the arrangement more strategically intertwined than a standard cloud capacity contract; the size and terms of the equity investment ha...
The deal puts Nscale in both frontier model compute and robotics infrastructure: Figure’s planned physical AI capacity sits alongside a separately reported $45 billion, six year Anthropic agreement at Nscale’s West Vi...
Nscale has committed an initial $3.5 billion of compute to Figure, with stated intent to exceed $6 billion, for up to 100,000 Nvidia Vera Rubin GPUs targeted for Barstow, Texas, from the second half of 2027. Nscale will become a Figure shareholder and its preferred compute provider, making the arrangement more strat...
Published byEdited with GPT-5.6 TerraImages generated with GPT Image 2
Research answer

Create a landscape editorial hero image for this Studio Global article: What are the terms, scale, technology, timeline, and strategic significance of Nscale’s multi-year partnership with Figure—including its ini. Article summary: Nscale’s Figure agreement is a strategic, multi-year physical-AI compute partnership rather than a conventional cloud lease: Nscale commits an initial $3.5 billion of capacity, intends to expand it beyond $6 billion, wil. Topic tags: general, general web, user generated, news. Style: premium digital editorial illustration, source-backed research mood, clean composition, high detail, modern web publication hero. Use reference image context only for broad subject, composition, and topical grounding; do not copy the exact image. Avoid: logos, brand marks, copyrighted characters, real person likenesses, fake screenshots, UI text, readable text, watermarks, charts w
Nscale’s partnership with humanoid-robotics company Figure is a long-term infrastructure and investment arrangement designed to supply the computing capacity behind Figure’s next generation of Helix AI models and robots. Nscale announced an initial commitment of $3.5 billion of compute, with an intent to scale the relationship to more than $6 billion. The partnership could deploy up to 100,000 Nvidia GPUs based on the Vera Rubin platform. 20
The headline figures are substantial, but the central qualifier matters: the first GPUs are targeted to begin deployment in the second half of 2027 at Nscale’s Barstow, Texas site. This is a planned capacity buildout, not compute that Figure can use immediately. 20
| Element | What has been announced |
|---|---|
| Initial compute commitment | $3.5 billion |
| Potential scale | Intent to grow beyond $6 billion, rather than a disclosed final contracted total |
| Hardware | Nvidia Vera Rubin platform, with potential for up to 100,000 GPUs |
| Initial deployment | Targeted for the second half of 2027 |
| Location | Barstow, Texas |
| Commercial relationship | Nscale becomes a Figure shareholder and preferred compute provider |
| Primary use | Training and supporting Figure’s next-generation Helix models and humanoid-robot deployment |
Nscale and Figure have not disclosed the size of Nscale’s equity stake, detailed capacity-pricing terms, reservation conditions or any take-or-pay provisions. That means the announced $3.5 billion should be understood as an initial compute commitment, while the amount above $6 billion remains a stated intention to expand the partnership. 20
38
The arrangement combines three roles for Nscale: infrastructure supplier, equity investor and Figure’s preferred compute provider. Nscale’s investment in Figure is undisclosed, but the structure aligns the cloud provider more closely with the robot maker’s commercial progress than a conventional customer-vendor agreement would. 20
38
For Figure, preferred-provider status creates a path to a very large dedicated AI-computing deployment. For Nscale, an equity position gives it exposure to Figure’s potential upside while it supplies the compute needed to train and deploy Figure’s models. The exact financial allocation between the cloud commitment and the equity investment has not been made public. 20
38
Nscale says the infrastructure will power the next generation of Figure’s Helix models. Helix is intended to support humanoid robotics, so the compute is relevant not only to model training but also to the broader process of developing and operating physical-AI systems. 20
40
The partnership’s potential scale—up to 100,000 Vera Rubin GPUs—suggests a much larger infrastructure ambition than a single training run. Humanoid robotics can require repeated model training, simulation and evaluation as a developer moves from lab demonstrations to a fleet of deployed machines. Still, the companies have not published a workload split between training, simulation and inference, so the precise use of the capacity is not known. 20
The strategic case is a robotics feedback loop: simulated environments can help train and test robotic policies, deployments can generate real-world operational data, and those results can inform later simulation and retraining. Nvidia describes related physical-AI workflows using simulation and digital twins to train and test robots in realistic environments. 42
For Figure, a large cloud deployment could support this kind of continuous improvement cycle for Helix. The important distinction is that the Nscale and Figure announcement confirms the planned compute partnership and its scale, but does not disclose the exact simulation pipeline or the allocation of GPUs to each stage. 20
The companies are also exploring use of Figure humanoid robots in Nscale’s supply chain. That is significant because it could make Nscale more than Figure’s cloud provider and investor: it could become a potential customer and operational testing environment. 23
24
However, this is exploratory. Neither company has announced a production deployment of Figure robots in Nscale data centers or supply-chain facilities. 23
24
The Figure partnership follows reports of a much larger Nscale agreement involving Anthropic. Bloomberg-reported terms describe Anthropic agreeing to spend about $45 billion over six years for roughly 460 megawatts of computing capacity at Nscale’s Monarch Compute Campus in West Virginia, using Nvidia Vera Rubin systems. 3
6
The Anthropic arrangement should be treated differently from the Figure announcement. Nscale publicly announced the Figure partnership and its core terms. By contrast, the Anthropic contract has been reported by multiple outlets, but detailed terms have not been publicly released by the parties. 20
3
The two arrangements point to different types of AI demand:
The Figure deal is smaller in announced dollar value, but it broadens Nscale’s exposure from frontier language-model workloads into robotics—a segment where demand may extend across model training, simulation and operation of deployed systems.
The partnership’s size does not remove its delivery risks. The initial deployment is targeted for the second half of 2027, so the practical value of the agreement depends on the timely completion and operation of the Barstow site, availability of Vera Rubin hardware and supporting networking and power infrastructure. 20
It also depends on Figure turning computing capacity into commercially useful humanoid robots. The announced agreement establishes the infrastructure ambition and strategic alignment; it does not guarantee that the full GPU deployment will occur or that the planned expansion beyond $6 billion will be reached. 20
Nscale’s Figure partnership shows how physical AI is becoming an infrastructure story as well as a robotics story. A humanoid-robot developer is seeking capacity on a scale associated with major AI-model builders, while the cloud provider is using a mix of dedicated compute, preferred-provider status and equity ownership to deepen the relationship. 20
38
The clearest takeaway is not that 100,000 GPUs are already powering Figure robots. They are not: deployment is targeted to start in late 2027. Instead, the agreement is a sizable forward commitment to the idea that training, simulating and improving humanoid robots could become a major long-duration source of cloud-compute demand. 20
Studio Global AI
This page includes a source-backed answer you can continue inside Studio Global.
Nscale has committed an initial $3.5 billion of compute to Figure, with stated intent to exceed $6 billion, for up to 100,000 Nvidia Vera Rubin GPUs targeted for Barstow, Texas, from the second half of 2027.
Nscale has committed an initial $3.5 billion of compute to Figure, with stated intent to exceed $6 billion, for up to 100,000 Nvidia Vera Rubin GPUs targeted for Barstow, Texas, from the second half of 2027. Nscale will become a Figure shareholder and its preferred compute provider, making the arrangement more strategically intertwined than a standard cloud capacity contract; the size and terms of the equity investment ha...
The deal puts Nscale in both frontier model compute and robotics infrastructure: Figure’s planned physical AI capacity sits alongside a separately reported $45 billion, six year Anthropic agreement at Nscale’s West Vi...