business 5 min read

Why Anthropic Bet on Akamai Over AWS and Azure

Anthropic's $11.6 billion, seven-year cloud deal with Akamai signals a seismic shift in AI infrastructure procurement. The company chose a CDN veteran over the hyperscalers — and the stock market took note.

  • Cloud Computing
  • AI Infrastructure
  • Anthropic
  • Akamai
  • Enterprise Deals

The Deal That Redrew the Map

Akamai just signed the largest single-customer contract in its history: a seven-year, $11.6 billion cloud infrastructure agreement with Anthropic. Under the right conditions, that number can climb to $20 billion. The stock surged 26 percent after the announcement. But the real story isn’t the price tag — it’s who won and who lost in the room where it happened.

Anthropic chose Akamai. Not Amazon Web Services. Not Microsoft Azure. A company most English-language readers know as a content delivery network, not a cloud heavyweight. That choice is a signal, and it’s one of the clearest yet about where the AI infrastructure race is heading.

Why Akamai, Not the Hyperscalers?

The source material makes no secret of the rationale: Akamai’s technical capability earned Anthropic’s trust. Tom Leighton, Akamai’s co-founder, put it plainly in the press release, saying the company was honored that Anthropic — described as “leading the AI revolution” — selected Akamai to build and operate its massive infrastructure.

But let’s read between the lines. AWS and Azure have scale that Akamai simply cannot match today. If Anthropic needed raw capacity above all else, those were the obvious answers. The fact that Anthropic went elsewhere suggests compute availability isn’t the only constraint — or perhaps it isn’t even the primary one.

There are a few plausible explanations.

First, Akamai may have offered a pricing structure or contractual flexibility that the hyperscalers couldn’t match for a deal this size. A $11.6 billion commitment over seven years locks in revenue visibility for Akamai, but it also means Anthropic is betting that Akamai can deliver reliably at scale — something a CDN operator has never been asked to do before.

Second, Akamai’s existing enterprise relationships with the world’s largest companies may have given it a信任 advantage. The press release references Akamai’s track record serving Fortune 500 clients as a differentiator. For an AI company navigating increasing regulatory scrutiny, partnering with an established enterprise infrastructure provider rather than a cloud platform that also competes in other markets may carry strategic weight.

Third, and perhaps most importantly, the deal structure itself reveals a deeper dynamic. Akamai issued warrants to Anthropic — 7.7 million shares of Series B preferred stock, worth roughly 5 percent of Akamai’s common shares outstanding, with a strike price of $111.33. Two percent vests immediately. Another 3 percent vests if Anthropic commits an additional $9 billion in cloud services over the contract period. This isn’t just a vendor relationship. It’s a financial alignment.

The Compute Constraint Nobody Talks About Enough

Every AI company outside the United States watches the infrastructure race with anxiety. The bottleneck isn’t just GPUs — it’s the entire stack: power, cooling, bandwidth, rack space, and the contractual certainty that your compute will actually be there when you need it.

Anthropic’s earlier contract with Akamai was reportedly worth $1.8 billion. In four months, that expanded to $11.6 billion. That’s a six-and-a-half-fold increase. No hyperscaler would necessarily welcome a customer renegotiating a deal of that magnitude in such a short window — unless they had no viable alternative.

The implication is stark: Anthropic has options, and it’s exercising them. The company is diversifying its infrastructure bets across providers rather than locking into a single cloud platform. That’s smart risk management in a market where supply constraints are real and worsening.

For other AI companies — especially those without Anthropic’s valuation or funding — the lesson is less comforting. If Anthropic can pull Akamai away from AWS and Azure for a deal this large, what does that say about the negotiating power of everyone else in the queue?

What the Numbers Actually Say

Analysts project Akamai’s revenue to grow from $4.49 billion in 2026 to $5.08 billion in 2027 and $5.66 billion in 2028. EPS should climb from $2.75 to $3.13 to $4.14 over the same period. The 2028 EPS jump of 32 percent year-over-year would reflect both revenue growth and operational leverage from the Anthropic deal.

These are modest figures for Akamai’s traditional business. The Anthropic contract changes the denominator. A $11.6 billion commitment spread over seven years represents roughly $1.66 billion in annual revenue — more than Akamai’s entire 2026 revenue projection. Even the base case of the deal alone would nearly double the company’s top line within the contract period.

The expansion clause to $20 billion adds another layer. If Anthropic commits the additional $9 billion — which vests in increments tied to $3 billion in new cloud service purchases — Akamai’s revenue could approach $2.86 billion annually from this single customer. That’s a transformation.

Who Wins, Who Loses, What Happens Next

Akamai wins immediately. Revenue visibility, stock appreciation, and a strategic repositioning from CDN provider to AI infrastructure partner. The company is now the largest single-customer play in its own history, and the market is rewarding it accordingly.

Anthropic wins too, though the calculus is more complex. Securing compute access outside the hyperscaler duopoly gives it leverage and optionality. But relying on a company whose core competency has historically been edge delivery, not massive data center operations, carries execution risk. Whether Akamai can scale to meet Anthropic’s demands at the required reliability standards remains unproven.

AWS and Azure lose — at least in this deal. The loss of an anchor customer like Anthropic to a non-hyperscaler sets a precedent. If other AI companies follow suit, the pricing power and lock-in effects that have defined cloud economics start to erode.

The broader AI industry loses competitiveness if compute procurement becomes this concentrated and opaque. When the biggest deals are negotiated in private with bespoke terms, smaller players face even steeper barriers.

Akamai is hosting a conference call at 5:30 p.m. Eastern time on the 24th. Investors will be listening for details on delivery timelines, capacity plans, and whether other AI companies are already in talks. The contract’s expansion clause creates a natural cadence for follow-on announcements — every $3 billion in additional commitments unlocks more equity vesting, which means more reasons for both companies to publicize progress.

The AI infrastructure race is no longer just about who builds the most chips. It’s about who can secure the contracts, and Akamai just proved that the old rules don’t apply anymore.