Can Anthropic and Amazon’s $100B AI Deal Challenge OpenAI

It feels like Wall Street meets Silicon Valley at a blackjack table, except one player keeps winning whether the cards go high or low. That player right now looks a lot like Amazon. Over the past few months, Amazon has quietly lined up two massive AI relationships, one with Anthropic and another tied to OpenAI’s $110 billion funding round. On paper, these look like bold investments in the future of artificial intelligence. Under the hood, they look a lot more like long-term revenue engineering tied to cloud infrastructure. For finance professionals, the real story is not who builds the smartest model. It is who locks in the cash flows.

Is this really an investment, or just prepaid AWS revenue?

Start with the Anthropic deal. Amazon is putting in $5 billion now, with the potential to invest up to $25 billion total when you include prior commitments. In return, Anthropic has agreed to spend more than $100 billion on AWS infrastructure over the next decade. That is not a small detail. That is the deal. Anthropic also secured up to 5 gigawatts of compute capacity, with nearly 1 gigawatt expected to be operational by the end of 2026. That level of infrastructure is massive, closer to industrial-scale power consumption than a typical software deployment. And it comes after Anthropic faced reliability issues due to surging demand for its Claude models, which now generate over $30 billion in annualized revenue.

So, what is Amazon actually buying? It is not control. It remains a minority investor with no board seat. What it is buying is demand. Predictable, long-term, high-margin cloud demand. Think of it like a landlord funding a tenant’s expansion so the tenant signs a ten-year lease. The upfront cash looks like an investment, but the real return comes from rent.

Why is Amazon backing both Anthropic and OpenAI at the same time?

Amazon is committing up to $50 billion as part of OpenAI’s $110 billion funding round, which values the company at around $730 billion. Alongside that, AWS is building out infrastructure partnerships to support OpenAI workloads, including a multi-billion-dollar compute commitment and deeper integration into Amazon’s Bedrock platform. At first glance, this looks like Amazon hedging its bets. Two competing AI leaders, both backed heavily. But the structure tells a cleaner story.

With Anthropic, Amazon locks in a clear $100 billion cloud spend commitment over ten years. With OpenAI, Amazon participates in a massive equity round and secures a strategic infrastructure role, even if the spend-back commitments are less explicitly framed in a single headline number. In both cases, Amazon sits in the same position: the infrastructure provider. It is a bit like owning toll roads on both sides of a busy highway.
It does not matter which lane traffic prefers; you still collect.

Where are these billions actually being deployed?

• First, compute capacity. Anthropic’s 5 gigawatts of secured capacity signals how AI has shifted from software scaling to power and hardware scaling. These models are no longer just code. They are energy-intensive systems that need a sustained compute supply.
• Second, custom silicon. Amazon keeps pushing Trainium and Graviton chips as lower-cost alternatives to Nvidia GPUs. These deals help validate that strategy. If Anthropic runs core workloads on Trainium for the next decade, that is a serious vote of confidence.

• Third, data centers and capex. Amazon expects roughly $200 billion in capital expenditures this year, with a large portion tied to AI infrastructure. Projects like its massive Indiana data center campus and the Rainier compute cluster show where the capital is going: land, power, cooling, and networking.
• Fourth, enterprise integration. AWS customers can now access Anthropic’s Claude platform directly within their existing AWS environments. No new contracts, no extra billing layers.

That matters more than it sounds. Procurement friction drops, adoption rises, and AI usage becomes embedded inside existing enterprise spend. For a CFO, this is not experimental anymore. It starts to look like another recurring technology line item.

Are these projections realistic?

Anthropic’s $100 billion commitment assumes sustained demand growth, continued enterprise adoption, and stable pricing dynamics for compute. At the same time, the cost of computation has historically declined over time, which could compress margins or shift how much capacity is actually needed per unit of output. There is also execution risk. Delivering multiple gigawatts of AI infrastructure is not trivial. Delays in data center buildouts, chip supply constraints, or energy availability could all affect how quickly that capacity becomes usable.

On the OpenAI side, the $110 billion funding round signals confidence, but it also raises expectations. Investors will look for monetization paths that justify that valuation. So, while the projections look bold, they are not guaranteed. They depend on sustained demand, efficient deployment, and continued enterprise willingness to pay for AI capabilities.

What does this mean for accounting and finance professionals?

• First, these deals blur the line between investment and revenue contracts. The commercial substance may matter more than the equity ownership.
• Second, capital intensity is back in the spotlight. AI is not just software margins anymore. It is capex-heavy, infrastructure-driven, and dependent on long-term planning.
• Third, vendor concentration risk increases. When an AI firm relies heavily on a single cloud provider for mission-critical workloads, that relationship becomes a key operational dependency.
• Fourth, enterprise cost structures will shift. AI is moving from experiment to recurring expense.

So, who’s actually winning here?

Amazon is playing a long game that looks pretty solid. By backing both Anthropic and OpenAI, it does not need to predict the ultimate winner in the model race. It just needs both to keep growing and consuming infrastructure. That turns AI demand into recurring cloud revenue, which is far more predictable than betting on a single product. Anthropic gets the compute it desperately needs to scale and avoid outages. OpenAI gets massive capital and broader infrastructure support. Both get what they need. Amazon gets something even better: position. The takeaway is simple. The headlines talk about billion-dollar investments. The real story is long-term, contracted infrastructure revenue. That is not flashy. It is just smart business.