Anthropic Pledges $200B to Google Cloud in 5-Year Deal

Anthropic Locks In $200 Billion Google Cloud Commitment

Anthopic has committed to spending approximately $200 billion on Google Cloud infrastructure over the next five years, according to a report from The Information. The deal, set to begin in 2027, puts a concrete price tag on the 5-gigawatt server capacity partnership the two companies announced last month — and fundamentally reshapes the economics of the cloud computing industry.

The staggering figure represents more than 40% of Google Cloud's total revenue backlog, which the company disclosed to investors last week had doubled to over $460 billion. That means Google's future cloud revenue is now deeply dependent on a single AI startup.

Key Takeaways

• Anthropic will spend roughly $200 billion on Google Cloud starting in 2027 over a 5-year period
• The deal accounts for over 40% of Google Cloud's $460 billion revenue backlog
• Two AI startups — Anthropic and OpenAI — now represent nearly half of all major cloud providers' combined backlog
• The four largest cloud providers (AWS, Azure, Google Cloud, Oracle) hold a combined $2 trillion in revenue backlog
• Amazon has also signed a separate $100 billion, 10-year server lease deal with Anthropic
• OpenAI added $100 billion in AWS commitments in a single quarter

Two AI Startups Now Prop Up the Entire Cloud Industry

The Anthropic-Google deal is not an isolated event. When you combine the cloud spending commitments of Anthropic and OpenAI, the two companies together account for nearly half of the roughly $2 trillion in revenue backlog held by the four major cloud providers: Amazon Web Services, Microsoft Azure, Google Cloud, and Oracle Cloud.

This concentration is unprecedented in the history of enterprise technology. Never before have two privately held companies wielded so much influence over the revenue trajectories of the world's largest cloud platforms.

The numbers tell a striking story. Amazon reported its Q1 revenue backlog surging 49% year-over-year to $364 billion. Of that increase, OpenAI's new AWS commitments alone contributed over $100 billion — accounting for more than 80% of the quarterly growth. Amazon then doubled down in April by signing a separate 10-year, $100 billion server lease agreement with Anthropic.

Google's Cloud Business Becomes an Anthropic Story

For Google Cloud, the implications are particularly dramatic. With Anthropic representing over 40% of its total backlog, Google's cloud division — long the third-place player behind AWS and Azure — has effectively hitched its growth narrative to a single customer.

This is both a massive opportunity and a significant risk. On the upside, Anthropic's commitment provides Google Cloud with revenue visibility that most enterprise software companies can only dream of. Five years of guaranteed spending at this scale transforms Google Cloud's financial profile and gives it ammunition to invest aggressively in infrastructure expansion.

On the downside, customer concentration at this level is a red flag that investors typically scrutinize heavily. If Anthropic were to face financial difficulties, pivot its infrastructure strategy, or renegotiate terms, the impact on Google Cloud's revenue would be severe. Google's cloud division generated roughly $41 billion in revenue in 2024 — meaning the Anthropic deal alone could represent a substantial multiple of its current annual run rate.

Why AI Companies Are Spending at This Scale

The massive cloud commitments from Anthropic and OpenAI reflect the extraordinary computational demands of training and serving frontier AI models. Each new generation of large language models requires exponentially more compute power, and the inference costs of running these models at scale for hundreds of millions of users are equally enormous.

Several factors are driving this spending surge:

• Model training costs are escalating rapidly, with next-generation models expected to cost billions of dollars in compute alone
• Inference demand is growing as AI products gain mainstream adoption and enterprise customers deploy AI agents
• Competitive pressure forces companies to secure compute capacity years in advance to avoid being bottlenecked
• Custom silicon — like Google's TPU chips — offers cost advantages over general-purpose GPUs for certain workloads
• Long-term contracts lock in favorable pricing and guarantee capacity in a market where GPU and server availability remains constrained

For Anthropic specifically, Google's Tensor Processing Units (TPUs) have been a core part of its training infrastructure since the company's founding. Anthropic was started by former OpenAI executives, including CEO Dario Ahassi and President Daniela Amodei, and the company has maintained a close technical relationship with Google from its earliest days.

The New Power Dynamic in Cloud Computing

This wave of AI-driven cloud commitments is creating a new and unusual power dynamic in the technology industry. Traditionally, cloud providers held the leverage — they owned the infrastructure, set the pricing, and could afford to lose any single customer. Now, with two AI companies representing such an outsized share of future revenue, the balance of power has shifted.

Anthopic and OpenAI are effectively playing cloud providers against each other. Anthropic maintains significant relationships with both Google Cloud and AWS, while OpenAI has expanded from its original Microsoft Azure partnership to also commit heavily to AWS and Oracle Cloud. This multi-cloud strategy gives the AI companies negotiating leverage and infrastructure redundancy.

The cloud providers, meanwhile, are competing fiercely to win these deals — offering not just compute capacity but also equity investments, co-marketing arrangements, and technical collaboration. Google has invested billions in Anthropic. Amazon has invested up to $8 billion. Microsoft's cumulative investment in OpenAI exceeds $13 billion.

This creates an interesting circular dynamic: cloud providers invest in AI companies, which then commit to spending those funds (and much more) back on cloud services. The AI companies get capital and compute; the cloud providers get long-term revenue commitments and the prestige of hosting the world's most advanced AI systems.

What This Means for the Broader Market

For enterprise customers and developers, the deepening ties between AI companies and cloud providers have several practical implications:

• Pricing stability: Long-term deals suggest cloud compute pricing for AI workloads may stabilize, as providers gain revenue certainty
• Infrastructure availability: Massive buildouts driven by these commitments will eventually increase overall capacity, potentially benefiting smaller customers
• Platform lock-in risks: As AI companies align more closely with specific cloud providers, their APIs and models may become more tightly integrated with those platforms
• Innovation acceleration: The guaranteed revenue streams give cloud providers confidence to invest in next-generation chips, networking, and data center technologies

For investors, the concentration of cloud backlog in two AI companies raises important questions about sustainability. These commitments assume that Anthropic and OpenAI will generate enough revenue from their AI products to justify spending hundreds of billions on infrastructure. If the AI market fails to grow as projected, or if a new paradigm reduces compute requirements, these deals could become problematic for all parties involved.

Looking Ahead: A $2 Trillion Bet on AI's Future

The $200 billion Anthropic-Google deal is perhaps the clearest signal yet that the AI industry is operating on a timeline and at a scale that has no historical precedent in technology. The combined cloud commitments from just two companies now approach $1 trillion across multiple providers — a figure that would have seemed absurd even 2 years ago.

Several key questions will determine whether these bets pay off. Can Anthropic and OpenAI convert their technological advantages into durable, profitable businesses? Will enterprise AI adoption accelerate fast enough to justify this infrastructure spending? And what happens if a technological breakthrough — such as more efficient model architectures or dramatically cheaper inference — reduces the need for raw compute?

The deal also raises antitrust considerations. Regulators in both the US and Europe are already scrutinizing the close financial ties between major cloud providers and AI companies. A commitment of this magnitude — effectively making Google Cloud financially dependent on Anthropic, and vice versa — could attract additional regulatory attention.

For now, the message from both Anthropic and Google is clear: they are betting big on each other, and on the assumption that the demand for AI compute will only grow. The 2027 start date gives both companies time to prepare, but it also means the real test of this partnership is still years away. In the fast-moving world of AI, a lot can change between now and then.