Key Points:
- Anthropic signed a $35 billion cloud computing agreement with Nvidia-backed cloud provider Lambda to secure dedicated AI chip capacity.
- The computing infrastructure will operate out of a 700-megawatt data center developed by Hut 8 in Nueces County, Texas.
- Nvidia holds the direct property lease on the Texas facility and will supply advanced graphics processing units for Lambda to install and manage.
- The contract follows Anthropic’s recent $45 billion computing deal with Nscale, expanding its total dedicated GPU commitments past $80 billion.
Artificial intelligence developer Anthropic signed a massive $35 billion cloud computing agreement with Lambda, a specialized graphics processing unit (GPU) cloud provider backed by Nvidia. The multi-year contract secures vast computational power to support the rapid growth of Anthropic’s Claude conversational assistant and next-generation frontier models. The deal represents one of the largest infrastructure agreements in AI history, highlighting the astronomical capital required to build and deploy advanced generative artificial intelligence.
Under the unique operational structure of the agreement, the computing capacity will run out of a state-of-the-art data center located in Nueces County, Texas. Energy infrastructure firm and data center developer Hut 8 is constructing the high-density computing campus. Nvidia directly holds the commercial lease on the Texas site after securing access to the property several weeks ago, creating a coordinated triangular partnership between chipmaker, infrastructure builder, and cloud operator.
Lambda will populate the Texas facility with high-end artificial intelligence accelerators purchased directly from Nvidia, which also maintains an equity stake in Lambda. Lambda will then manage the hardware infrastructure and deliver dedicated cloud compute services directly to Anthropic’s engineering teams. This arrangement allows Lambda to execute a multibillion-dollar enterprise contract without having to finance the real estate purchase directly, while Anthropic locks in guaranteed access to cutting-edge silicon.
The deal addresses persistent hardware bottlenecks that have slowed artificial intelligence developers over the past year. Surging demand from enterprise customers, software developers, and research labs previously caused server capacity constraints for Anthropic, limiting user access during peak hours. By securing dedicated computing infrastructure through specialized cloud partners, Anthropic ensures that enterprise clients can deploy Claude across automated business workflows without latency spikes or unexpected service interruptions.
This $35 billion contract follows another massive computing transaction that Anthropic finalized earlier this month. The AI company signed a $45 billion computing agreement with Nscale, another Nvidia-backed specialized cloud provider, to lease computing capacity from a major facility in West Virginia. Together, these two agreements represent an unprecedented $80 billion infrastructure commitment aimed at locking down GPU supply through the end of the decade.
The transactions underscore Anthropic’s multi-cloud strategy as it competes aggressively against OpenAI and Google. While Anthropic maintains deep technical partnerships with hyperscale giants—securing $30 billion in computing commitments from Microsoft Azure alongside billions in backing from Amazon Web Services and Google Cloud—the company is deliberately expanding its footprint with specialized “neocloud” providers. These dedicated GPU clouds focus exclusively on machine learning clusters, offering optimized hardware networking and faster server deployment.
For Nvidia, the agreement reinforces its central position within the global artificial intelligence economy. With a market valuation exceeding $5 trillion, Nvidia is actively using its balance sheet and venture capital arm to back emerging cloud operators like Lambda and CoreWeave. By securing data center leases and supplying hardware to these fast-growing providers, Nvidia creates guaranteed commercial demand for its advanced microchips while building alternative distribution channels outside traditional Big Tech clouds.
The energy demands of the new computing facility highlight the critical link between artificial intelligence expansion and power generation. The Texas campus developed by Hut 8 features a planned capacity of up to 700 megawatts, requiring dedicated high-voltage grid connections and specialized liquid-cooling systems to manage thousands of densely packed server racks. AI developers now compete as aggressively for electrical substation capacity and power purchase agreements as they do for software talent.
The immense scale of these infrastructure commitments reflects the intensifying race to achieve artificial general intelligence. Training next-generation models with hundreds of billions of parameters requires continuous computing clusters running uninterrupted for months at a time. Industry experts estimate that training future frontier models will cost upwards of $1 billion per training run, making long-term access to reliable, low-cost compute capacity a vital competitive moat for top AI labs.
As Anthropic prepares for a potential public offering, its aggressive infrastructure buildup positions the company to meet soaring enterprise adoption worldwide. By securing tens of billions of dollars in dedicated computing power across Texas, West Virginia, and global hyperscalers, Anthropic ensures it has the physical foundations in place to power the next generation of enterprise automation and intelligent agents.
