Skip to content

AI Decision Support

A cloud vendor just gave an AI lab a path to 5% of its own stock to win a compute deal. Here is what that says about AI's real cost.

On September 24, 2026, Akamai Technologies announced a seven-year, $11.6 billion agreement to supply Anthropic with computing capacity, with room to grow to roughly $20 billion. The direct answer for a business owner: this is not a story about your ChatGPT or Claude subscription changing price tomorrow. It is a story about how AI labs are now paying for the massive computing capacity behind those subscriptions, increasingly through complex, multi-year financial commitments rather than simple usage-based contracts, and what that means for anyone building a business that depends on one of these vendors.

By Fabio Rabelo · Founder, ATLACIS ·

What happened

Akamai Technologies, a cloud and content-delivery company, announced on September 24, 2026 that it signed a seven-year, $11.6 billion contract to supply Anthropic with dedicated cloud computing capacity, specifically for CPU workloads rather than the GPU clusters used to train Claude models. The agreement can expand by another $9 billion, for a total potential commitment of about $20 billion, according to Akamai's own press release and its SEC filing. What makes this deal different from an ordinary cloud contract is how part of it was paid for. Akamai issued Anthropic a warrant, a right to buy stock at a set price, for up to approximately 5% of Akamai's common stock. About 2% of that vests immediately with the current $11.6 billion commitment. The rest vests only if Anthropic commits to buy an additional $3 billion in cloud services at a time, up to the full $9 billion expansion. Akamai says it expects to spend about $5.5 billion building out the infrastructure this deal requires, and it does not expect meaningful revenue from the contract until the second half of 2027. This is the third large, multi-year compute commitment Anthropic has made public since late August 2026. Reuters and Bloomberg confirmed a $45 billion, six-year deal with UK infrastructure company Nscale on August 26 for a West Virginia data center, and a $35 billion deal with Nvidia-backed Lambda on August 31 for a Texas data center. All three deals share the same shape: a young, fast-growing AI lab locking in years of future computing capacity through unusually large, structurally creative contracts.

Why it matters for business owners

Most business owners will never sign a deal like this. But every business that runs on ChatGPT, Claude, or any product built on top of a frontier AI model is, indirectly, a customer of exactly this kind of infrastructure spending. The price you pay per token, the reliability of the service during peak demand, and how quickly new model capability reaches you are all downstream of whether the lab behind your tool of choice has secured enough computing capacity, and on what financial terms. When a vendor pays for capacity partly with equity in itself rather than only cash, that is a signal about how tight the market for computing capacity has become. Akamai was willing to give up a slice of ownership to lock in a customer this large, and Anthropic was willing to accept equity risk in a cloud vendor as partial consideration, rather than simply negotiating a lower cash price. That is not how a business normally buys a commodity input. It is how a business buys something it is worried it might not be able to get enough of otherwise.

What owners should not misunderstand

This is not evidence that AI is about to get more expensive for your business tomorrow, and it is not evidence of financial distress at Anthropic or Akamai. Akamai explicitly said the deal will not affect its 2026 revenue guidance, and the contract is structured as a take-or-pay commitment, meaning Akamai gets paid once it delivers the capacity, regardless of exactly how Anthropic ends up using it. It is also not the same mechanism as other recent AI financing stories. Earlier this year, a chip supplier took a direct equity stake in an AI lab it sells chips to, and a different chip supplier guaranteed a lab's data center debt. This deal runs the other direction: a cloud vendor is giving an AI lab a path into its own stock, tied to how much compute that lab commits to buy. Three different companies, three different financial structures, all solving the same underlying problem: securing enough computing capacity for a business that is growing faster than normal contracts can comfortably finance. Do not read this as proof that any specific model or vendor is unstable. Multiple credible outlets confirm the deal's terms and Anthropic's broader pattern of large compute commitments, but none report any sign that Anthropic cannot meet its obligations.

The operational lesson

The lesson here is not about Anthropic's balance sheet. It is that the AI vendor your business depends on is making long-duration, multi-billion dollar bets on future demand for its own products, financed through arrangements far more complex than a simple monthly subscription. That has two practical implications for any business relying on a frontier AI vendor. First, current pricing is shaped by a temporary, capital-intensive buildout phase, not a settled market. A vendor locking in seven years of capacity today is betting that demand keeps growing fast enough to justify it. If that bet is right, prices likely keep falling as capacity comes online. If it is wrong, or comes online slower than planned, a business that has built critical workflows around today's low prices could see costs shift with little warning. Second, a wave of large, structurally unusual compute deals in a short window (three from Anthropic alone since late August) is itself useful information when evaluating any AI vendor relationship: it tells you the vendor is scaling aggressively and financing that scale creatively, which is different from a vendor whose capacity and pricing have already stabilized.

What a serious business should do next

Do not overreact to this story by rushing to lock in a long-term AI contract of your own, and do not treat it as a reason to distrust any specific vendor. Nothing here suggests a service disruption is imminent. Do treat it as a prompt to ask a basic due-diligence question the AI hype cycle usually skips: is the vendor I am building on financing its growth through ordinary revenue, or through a stack of multi-year capacity commitments that assume continued rapid growth? That is a fair question to ask a vendor directly, or to have someone check on your behalf before you build a core workflow around a single provider. Do separate your own near-term decisions (which model to use for a given task, whether to build on one vendor's API) from your long-term posture (how exposed your business is if that vendor's pricing, capacity, or roadmap shifts). A business with light, easily swapped AI usage has little to worry about here. A business with deep, hard-to-migrate dependence on one vendor's specific model or platform has more reason to understand what is driving that vendor's cost structure. Do not assume owning your own hardware avoids this dynamic entirely. On-premise and private deployment carry their own capital and financing tradeoffs; the point is to make that comparison with real numbers, not with assumptions about which path is more stable.

The Atlacis view

Stories like this tend to get read as finance-page noise that has nothing to do with an ordinary business's AI budget. That is a mistake. Every dollar an AI lab spends locking in computing capacity, and every unusual financing structure it uses to do it, eventually shows up somewhere in what a business pays, how reliable the service is, and how much leverage that business has if it ever needs to switch vendors. Atlacis helps owners look past the model name and the marketing and ask the more useful question: what is this vendor's actual cost structure, how exposed is my business to a shift in it, and would a different mix of cloud, private, or on-premise AI give me more control over that exposure. A deal like this one is not a reason to panic. It is a reason to have that conversation before a vendor's financing decisions become your business's problem.

The short version

  • On September 24, 2026, Akamai announced an $11.6 billion, seven-year cloud computing deal with Anthropic, expandable to roughly $20 billion, confirmed directly through Akamai's press release and SEC filing.
  • Akamai paid for part of the commitment by issuing Anthropic a warrant for up to approximately 5% of Akamai's own stock, vesting in tranches as Anthropic's spending commitment grows.
  • This is the third large, multi-year compute deal Anthropic has signed since late August 2026, alongside a confirmed $45 billion deal with Nscale and a $35 billion deal with Lambda.
  • The deal does not signal an imminent price change for ordinary AI subscriptions, but it does show that today's AI pricing sits on top of an active, capital-intensive capacity buildout, not a settled market.
  • The practical lesson is to ask any AI vendor your business depends on how it is financing its growth, and to weigh that against how deeply your workflows depend on that specific vendor.
Tags:AI vendor riskvendor dependencyAI infrastructureAI costcloud vs on-premiseAI decision support
Keep reading

More from the blog

Nvidia may guarantee $250 billion of OpenAI's debt so OpenAI can build a data center neither company could finance alone. Here is what business owners should know about AI vendors whose finances are tangled together.

The Wall Street Journal reported on July 26, 2026 that Nvidia is in talks to guarantee roughly $250 billion in financing for OpenAI, helping it lease a 10-gigawatt data center in Ohio that OpenAI could not easily borrow for on its own credit. Reuters, CNBC, and other outlets independently confirmed the reported figures, though neither company has confirmed the talks. The deal is not an investment and not a compute lease. It is a chip supplier co-signing its biggest customer's debt, a deeper and less visible form of financial entanglement than either of the two vendor-financing stories covered here in the last two weeks.

AMD just agreed to invest up to $5 billion in Anthropic, the same company it is selling billions of dollars in AI chips to. Here is what business owners should know before treating a vendor's confidence as an independent recommendation.

On July 22, 2026, AMD and Anthropic announced a deal in which Anthropic will deploy up to 2 gigawatts of AMD's AI chips and AMD will invest up to $5 billion in Anthropic in return. Reuters called it the latest in a wave of 'circular deals' now common across the AI industry, where a chip supplier takes a financial stake in the same company buying its hardware. The useful lesson has nothing to do with GPUs. It is that a supplier's public enthusiasm for its own customer is not the same thing as independent proof that a tool, platform, or vendor is right for your business.

Meta may lease a competitor $10 billion in computing power. Here is what business owners should know about AI capacity, not just AI price.

Meta Platforms is in early talks to lease Anthropic up to $10 billion in AI computing power over two years, even though Anthropic competes with Meta on AI models. Anthropic is already spending $50 billion on its own data centers and already leases capacity from SpaceX and TeraWulf. The lesson is not about either company's finances. It is that AI compute capacity is scarce enough that even the best-funded labs in the world cannot buy their way past it, and that changes what a business should ask any AI vendor before depending on it.

Make better AI decisions, starting with one call.

Book a free AI Fit Call. We will tell you what to use, what to avoid, and where to start. No jargon, no pressure.