Anthropic has signed a cloud-computing deal worth $35 billion with Lambda, a cloud provider backed by Nvidia, to bring new Nvidia GPU capacity online at a Texas data center, according to people familiar with the matter.
The deal, first reported by the Wall Street Journal and confirmed by Bloomberg and Reuters, will bring online Nvidia capacity to meet growing demand for Anthropic Claude AI. The agreement represents the largest single computing commitment by an AI company and ranks among the biggest cloud infrastructure deals in technology history.
Lambda, headquartered in San Francisco, operates cloud GPU clusters primarily used for AI training and inference. Nvidia invested in the company as part of its broader strategy to expand the supply of AI computing infrastructure beyond the hyperscale cloud providers like Amazon, Google, and Microsoft.
Scale of the Deal
At $35 billion, the Lambda pact exceeds the computing commitments most AI companies have made with individual cloud providers. Google and Anthropic announced a separate deal worth tens of billions of dollars in late 2025 focused on TPU capacity. Amazon has invested billions in Anthropic through equity stakes and cloud credits on AWS. Microsoft OpenAI partnership includes extensive Azure computing arrangements, though specific dollar figures have not always been disclosed.
The Lambda deal is notable for routing through a mid-size cloud provider rather than a hyperscaler. It reflects a growing trend in AI infrastructure where companies seek diversified computing sources to avoid concentration risk on any single platform. Lambda operates data centers in multiple US locations and has been expanding capacity aggressively since its Series B funding round.
Industry estimates peg the cost of a 1-gigawatt data center at around $50 billion, with roughly $35 billion of that typically allocated to chips. The Lambda deal suggests Anthropic is securing a substantial portion of new GPU capacity at or near this benchmark cost structure.
Why It Matters for the AI Race
Compute has become the primary bottleneck in the AI industry. Training frontier models requires thousands of high-end GPUs running for weeks or months. The availability of computing capacity increasingly determines which companies can train the largest models and serve the most users.
Anthropic launched Claude Fable 5.1 last week, its latest flagship model that outperforms predecessors in coding and reasoning while cutting cache read costs by 25%. The model has driven a surge in API usage as enterprise customers adopt it for coding assistants, customer service automation, and data analysis. That usage growth is what drives the need for additional computing capacity.
The company is now one of the fastest-growing AI startups by revenue. Anthropic annualized revenue run rate has reportedly exceeded $10 billion, driven by enterprise adoption of Claude across industries including financial services, healthcare, and software development. The Lambda computing deal positions the company to scale that growth without depending entirely on AWS capacity.
Nvidia Role
Nvidia backing of Lambda puts the chipmaker in an unusual position as both a supplier and an investor in the infrastructure chain. Nvidia sells the GPUs that Lambda operates, but it also invested in Lambda to ensure sufficient cloud capacity exists for its chips to run AI workloads.
The arrangement benefits Nvidia by expanding the market for its high-end accelerators. Rather than selling GPUs only to the three hyperscale providers, Nvidia can channel additional sales through Lambda and other independent cloud providers. This reduces the bargaining power of any single cloud customer and increases overall demand for Nvidia hardware.
Nvidia stock traded at $217.44 on Monday, down 1.51% as broader market concerns about rising interest rates and geopolitical uncertainty weighed on tech shares. The Lambda deal announcement did not produce a significant after-hours move, suggesting the market viewed it as confirmatory rather than incremental.
The Texas Connection
The data center at the center of the deal is located in Texas, which has become a hub for AI infrastructure buildout due to its relatively low energy costs and business-friendly regulatory environment. The state has attracted billions in data center investment from companies including Tesla, Oracle, and various crypto mining operations that have pivoted to AI compute.
Texas power grid challenges during extreme weather events remain a concern for data center operators, but the AI industry has generally accepted the risk in exchange for lower electricity costs compared to markets like Northern Virginia or the Pacific Northwest.
The Lambda facility will deploy Nvidia latest-generation GPUs optimized for AI training and inference workloads. Specific GPU models were not disclosed in the deal announcement, but Lambda current catalog includes H100 and upcoming Blackwell-series accelerators.
Competitive Landscape
The deal intensifies competition in the AI infrastructure market. CoreWeave, another Nvidia-backed GPU cloud provider, went public earlier in 2026 at a valuation exceeding $30 billion. Together, Lambda and CoreWeave represent a new class of AI-native cloud providers that compete with the hyperscale giants on speed, flexibility, and pricing for GPU-intensive workloads.
For Anthropic, the Lambda deal provides a hedge against over-reliance on any single computing partner. The company already operates on Google Cloud and AWS, and adding Lambda as a third major computing partner gives it more negotiating leverage and operational resilience.
The deal also signals that AI companies are willing to commit enormous sums to secure computing capacity in advance, locking in supply before competitors can acquire it. In a market where GPU availability remains constrained, these pre-commitments have become a strategic weapon in the race to build and deploy the most capable AI systems.
discussion