SAN FRANCISCO — Anthropic is making another multibillion-dollar bet on the future of artificial intelligence, signing a reported $35 billion cloud-computing agreement with Nvidia-backed provider Lambda as demand for computing power behind Claude and other AI software continues to accelerate.
The agreement would bring additional Nvidia-powered capacity online at a data center under development in Nueces County, Texas, according to people familiar with the matter. The facility is being developed by Hut 8 and is expected to provide roughly 350 megawatts of capacity.
The deal underscores a fundamental shift in the technology industry: AI software is no longer constrained only by algorithms and engineering talent. Increasingly, its growth depends on access to enormous amounts of electricity, data-center space, networking equipment and advanced chips.
For Anthropic, the agreement represents another major step in securing the infrastructure it needs to scale Claude and its AI coding product, Claude Code.
A $35 Billion Bet on AI Compute
The Lambda agreement is one of several enormous infrastructure commitments Anthropic has made as it prepares for continued growth.
Reuters reported that the $35 billion deal is designed to bring Nvidia capacity online to meet increasing demand for Anthropic’s AI products. Neither Anthropic, Nvidia, Lambda nor Hut 8 immediately commented publicly on the agreement.
The arrangement is significant not simply because of its dollar value.
It demonstrates how AI companies are increasingly signing long-term agreements for computing capacity before that capacity is fully built.
That strategy allows companies such as Anthropic to secure infrastructure ahead of expected demand rather than waiting until data centers and advanced processors become scarce.
Texas Becomes a Major AI Infrastructure Hub
The data center connected to the Lambda agreement is being developed in Nueces County, Texas.
Hut 8, which previously operated heavily in cryptocurrency mining, has increasingly moved toward the data-center market as demand for AI computing has surged.
The company’s transition reflects a broader transformation across the technology infrastructure industry.
Facilities originally designed or positioned for cryptocurrency mining can offer valuable assets for AI development, including access to electricity, land and existing infrastructure.
AI companies and their infrastructure partners are now looking for locations capable of supporting massive power requirements.
Nvidia’s Role Is Expanding Beyond Chips
The deal also highlights Nvidia’s increasingly influential position in the AI infrastructure ecosystem.
Nvidia is best known for producing the graphics-processing units that power AI training and inference.
But the company’s role has expanded as the AI boom has created demand for entire computing ecosystems.
In the Lambda arrangement, Nvidia-backed financing and infrastructure relationships are helping cloud providers bring additional capacity online.
The company is also reportedly involved in the Texas data center through a lease arrangement with Hut 8.
That means Nvidia can benefit from AI growth not only when customers purchase its processors, but also as additional infrastructure is built to deploy those processors.
Anthropic’s Computing Needs Are Growing Rapidly
Modern AI models require enormous amounts of computing power.
Training a frontier model can involve thousands or even hundreds of thousands of specialized processors operating simultaneously.
Once a model is released, inference creates another major demand.
Every time a user asks Claude a question, generates code or requests an analysis, computing resources are required to produce the response.
As enterprise adoption grows, that demand can become enormous.
Claude Code adds another dimension because AI-assisted software development can involve sustained, complex interactions between users and models.
Claude Code Is Driving Enterprise Demand
Anthropic has increasingly positioned Claude as a tool for businesses rather than simply a consumer chatbot.
Claude Code is designed to help developers write, modify, test and understand software.
That means enterprise users can potentially generate large numbers of model requests during a single development session.
The more deeply AI becomes integrated into business workflows, the greater the demand for reliable compute capacity.
Anthropic’s infrastructure commitments indicate that the company expects that demand to continue rising.
The $35 Billion Deal Is Not Anthropic’s Only Major Commitment
The Lambda agreement follows a series of other major infrastructure deals.
Anthropic recently agreed to spend $45 billion to rent AI computing capacity from Nscale’s planned data-center campus in West Virginia, according to reporting from the Financial Times. That agreement is expected to provide access to hundreds of megawatts of Nvidia’s next-generation processors.
Anthropic has also expanded its relationship with Amazon.
In April, Anthropic announced an agreement with Amazon that would secure up to 5 gigawatts of new computing capacity for training and deploying Claude.
The company said it would commit more than $100 billion over the next decade to AWS technologies under the expanded collaboration.
Taken together, those agreements demonstrate the scale of the infrastructure race surrounding frontier AI.
AI Companies Are Signing Capacity Before They Need It
One of the biggest changes in the technology business is the emergence of long-term compute contracting.
In previous software cycles, companies could often scale servers gradually as customers arrived.
AI is different.
Training and operating advanced models require specialized processors and large amounts of power.
Building a new data center can take years.
That creates a mismatch between rapidly growing demand and slowly expanding supply.
Companies therefore have an incentive to secure capacity years in advance.
The Electricity Problem
Behind every AI model is an energy requirement.
A large data center can consume as much electricity as a small city.
As AI companies expand, they increasingly compete not only for Nvidia processors but also for access to power grids.
That is why Texas, West Virginia and other regions with large amounts of available power have become increasingly important to the AI industry.
The next phase of AI competition may therefore depend as much on energy infrastructure as on software innovation.
Data Centers Are Becoming Strategic Assets
The shift is changing the way technology companies think about infrastructure.
A data center is no longer simply a place to store servers.
For AI companies, it is a strategic asset that determines how quickly models can be trained, how many customers can be served and how much revenue can be generated.
The location of a data center also matters.
Power availability, cooling capacity, transmission infrastructure, land costs and permitting timelines can all influence whether an AI project can move forward.
The Rise of the “Neocloud”
The Lambda agreement also highlights the growth of specialized cloud providers sometimes referred to as neoclouds.
Unlike traditional hyperscalers such as Amazon Web Services, Microsoft Azure and Google Cloud, these providers often focus heavily on AI infrastructure.
They can offer customers access to large quantities of specialized processors without requiring the customer to build its own data centers.
For AI companies, neoclouds provide another way to secure capacity.
For investors, they represent a rapidly growing part of the AI infrastructure market.
Nvidia’s Financing Strategy Is Drawing Attention
Nvidia’s relationship with AI infrastructure providers has attracted increasing scrutiny.
The company has invested in or backed several businesses that build or provide AI computing infrastructure.
That creates a potentially powerful feedback loop.
Nvidia supports infrastructure providers.
Those providers purchase Nvidia chips.
AI companies rent the resulting computing capacity.
The increased use of AI then creates additional demand for Nvidia hardware.
Recent reporting has described the structure around the Lambda deal as another example of Nvidia’s expanding role in financing the AI ecosystem.
Investors Are Watching the Economics
The enormous size of AI infrastructure contracts raises an important question: will demand remain strong enough to justify the commitments?
AI companies are betting that enterprise adoption will continue growing rapidly.
Cloud providers are betting that customers will keep paying for access to specialized compute.
Data-center operators are investing billions of dollars based on long-term demand expectations.
Chipmakers are expanding production based on those same expectations.
The entire ecosystem is increasingly interconnected.
The Risk of Overbuilding
The AI infrastructure boom also creates the possibility of overbuilding.
If AI demand grows more slowly than expected, companies could end up paying for capacity they do not immediately need.
That could pressure cloud providers and data-center operators.
It could also create downward pressure on computing prices.
For now, however, the industry’s dominant concern remains insufficient capacity rather than excess capacity.
Anthropic Is Preparing for an AI Growth Race
Anthropic’s infrastructure spending also reflects its position in an increasingly competitive AI market.
The company is competing with OpenAI, Google and other major developers for enterprise customers and developer adoption.
Access to computing power can directly influence how quickly a company can train new models and expand existing services.
If Anthropic cannot secure enough capacity, its competitors could gain an advantage.
The $35 billion Lambda agreement is therefore partly a defensive move.
It ensures Anthropic has a larger infrastructure base available as demand grows.
Custom Chips Could Change the Equation
Anthropic is also exploring alternatives to complete reliance on Nvidia hardware.
Reuters recently reported that the company explored acquiring AI-chip startup MatX for approximately $7 billion before discussions shifted toward a possible partnership.
Anthropic has also expanded its use of Google’s custom Tensor Processing Units through its broader relationship with Google.
The strategy reflects a wider industry trend.
AI companies increasingly want multiple sources of compute so they are not completely dependent on a single chip supplier.
Why Diversification Matters
Nvidia remains the dominant supplier of AI accelerators.
But demand for its hardware has been so strong that major AI companies have sought alternative processors.
Google has its TPUs.
Amazon has Trainium.
Microsoft is developing its own AI accelerators.
Meta has been working on custom silicon.
Anthropic’s strategy appears to be moving in the same direction.
The goal is not necessarily to abandon Nvidia.
It is to ensure that the company has enough computing options to support rapid growth.
Amazon Remains a Major Infrastructure Partner
Anthropic’s relationship with Amazon is particularly important.
The companies said in April that more than 100,000 customers were already running Claude through Amazon Bedrock.
Anthropic also said it was using more than one million Trainium2 chips to train and serve Claude.
The expanded agreement with Amazon gives Anthropic another major source of computing capacity while strengthening Amazon’s position in the AI infrastructure market.
That diversification could become increasingly valuable as AI demand grows.
Google Is Also Part of the Compute Strategy
Google has emerged as another major Anthropic infrastructure partner.
The company provides Anthropic with access to its custom TPUs, giving Claude an alternative to Nvidia-based systems.
That relationship is strategically important because Google can provide both computing hardware and cloud infrastructure.
For Anthropic, it creates another path to scaling AI models.
For Google, Anthropic is a major customer for its custom AI chips.
The Infrastructure Race Is Reshaping the Software Industry
The AI boom is changing the traditional economics of software.
Software companies historically benefited from relatively low marginal costs.
Once a software product was built, serving another customer could be comparatively inexpensive.
AI changes that equation.
Every model interaction requires computing resources.
More users can mean significantly higher infrastructure expenses.
That means AI software companies need to balance rapid adoption with the cost of serving every request.
AI Pricing Could Become More Important
As infrastructure expenses rise, AI companies will increasingly have to determine how much customers are willing to pay for computing-intensive products.
Subscription models may not always be enough.
Companies may introduce usage-based pricing, enterprise capacity contracts or premium tiers for high-compute applications.
Claude Code and other AI development tools could accelerate that trend because sophisticated coding tasks can consume substantially more compute than simple conversational requests.
The AI Infrastructure Boom Is Becoming a Financing Story
Another important feature of the current cycle is the amount of capital required to build AI infrastructure.
Data centers cost billions.
Specialized chips cost billions.
Power infrastructure requires major investments.
Companies are increasingly turning to private equity, banks and other financial institutions to fund expansion.
That creates new links between AI companies and financial markets.
The more capital flows into AI infrastructure, the more investors will demand evidence that future AI revenues can support those commitments.
Anthropic’s IPO Adds Another Layer
Anthropic is preparing for a potential public offering, according to recent reports.
That could make its infrastructure spending an important issue for future investors.
Investors will want to understand how quickly revenue is growing, how much the company spends on computing and whether long-term cloud contracts will generate sufficient returns.
Large infrastructure commitments can support growth, but they also create fixed obligations.
The economics will depend heavily on future demand.
What It Means for the U.S. AI Economy
The Lambda agreement is another signal that AI infrastructure is becoming a major part of the U.S. economy.
Data centers create construction activity, demand for electricity and networking equipment, and jobs in areas where facilities are built.
But they also create challenges.
Communities must deal with increased electricity consumption, water requirements, land use and infrastructure costs.
The economic benefits therefore come with a growing debate over how AI infrastructure should be developed.
Texas Is Positioned for More Growth
Texas has become one of America’s most important locations for data-center development.
The state offers abundant land, major energy resources and a growing technology ecosystem.
The Anthropic-linked project in Nueces County adds another major AI investment to that broader trend.
If AI demand continues to expand, Texas could see even more data centers, power projects and semiconductor-related investments.
The Next Bottleneck May Be Power
The semiconductor shortage was one of the defining constraints of the early AI boom.
Now another bottleneck is becoming increasingly visible: electricity.
Even if enough processors are available, data centers cannot operate without reliable power.
That means utilities, transmission operators and energy developers are becoming increasingly important players in the AI economy.
AI companies may ultimately compete for electricity as aggressively as they compete for chips.
The Bigger Picture
Anthropic’s $35 billion Lambda agreement is more than another giant technology contract.
It is a snapshot of how quickly AI is changing the infrastructure underlying the software industry.
AI companies are signing multibillion-dollar contracts.
Cloud providers are building specialized capacity.
Chip companies are expanding their financing role.
Data-center developers are converting former cryptocurrency infrastructure into AI facilities.
And energy companies are preparing for unprecedented demand from computing.
The entire technology stack is being rebuilt around artificial intelligence.
The Bottom Line
Anthropic’s reported $35 billion cloud-computing agreement with Nvidia-backed Lambda highlights the extraordinary infrastructure requirements behind the next generation of AI software.
The deal involves a Texas data center under development by Hut 8 and is expected to provide roughly 350 megawatts of capacity. The computing resources are intended to support growing demand for Anthropic’s Claude AI products, including Claude Code.
The agreement comes shortly after Anthropic announced or entered into other enormous computing commitments, including a deal for up to 5 gigawatts of new capacity with Amazon and a reported $45 billion agreement with Nscale for additional AI computing resources.
The scale of those investments shows that the AI software race is increasingly becoming an infrastructure race.
The companies that build the best AI models will still need something just as fundamental: enough chips, electricity and data-center capacity to run them.
For Anthropic, securing that capacity is becoming a strategic priority.
For Nvidia and cloud providers, the boom creates a massive new market.
And for investors, the question is becoming increasingly important: can the revenue generated by AI software grow fast enough to justify the extraordinary cost of building the infrastructure required to power it?
For now, the industry’s answer appears to be yes.
The billions of dollars flowing into new data centers suggest technology companies are preparing for an AI market that could be dramatically larger than today’s.
The real test will come when all that infrastructure is online—and the industry has to prove that the demand for AI software is large enough to keep it running profitably.
Source angle: Reuters reporting on Anthropic’s Lambda agreement, Anthropic’s official Amazon compute announcement, and reporting on Anthropic’s broader AI infrastructure expansion.
