Akamai’s $11.6B Anthropic Deal: A B2B Positioning Lesson
Anthropic committed $11.6B to Akamai’s cloud. Neither company has said why. What the deal terms suggest, who else was in the running, and the lesson for B2B.
On 24 September 2026, Akamai announced that Anthropic had committed $11.6 billion over seven years to Akamai Cloud, the largest deal in Akamai’s history. Neither company has said why Anthropic chose Akamai. The deal terms say a lot anyway, and they carry a positioning lesson for any B2B company that sells against giants.
Akamai is best known as the content delivery and web security company. It won this deal as a distributed cloud for the kind of computing AI agents need. That shift did not happen in a week, and the way Akamai made it is the part worth studying.
What did Akamai and Anthropic announce?
Akamai filed the details with the US Securities and Exchange Commission (Akamai 8-K, September 2026):
- Anthropic committed $11.6 billion over seven years to Akamai Cloud’s distributed infrastructure and software, “to support CPU workload growth at scale”.
- Anthropic can expand the commitment by up to $9 billion within the term, taking the total to roughly $20 billion.
- Akamai issued Anthropic a warrant for up to about 5% of Akamai’s common stock at $111.33 a share. About 2% vests with the initial commitment and about 1% more for each additional $3 billion of spend.
- Akamai CEO Tom Leighton said Anthropic “chose Akamai’s capabilities for building and operating AI infrastructure at scale.”
The filing contains no quote from Anthropic. TechCrunch reported that Akamai “didn’t say what Anthropic will use them for”, and that Akamai shares rose as much as 17% after hours (TechCrunch).
Why would Anthropic buy CPU capacity from Akamai?
Neither company has explained the choice, so what follows is our reading of the terms. Four things stand out.
The deal is for CPUs. Most AI infrastructure spending goes on GPUs for training and running models. TechCrunch called CPUs “a less-hyped corner of AI infrastructure” and noted that demand for them grows as AI agents run code, browse the web and use tools. A buyer that needs a large amount of general-purpose compute has reason to look beyond the providers whose capital is going into GPUs.
Akamai is spread out. Its filing describes “a vastly distributed network spanning thousands of points of presence”, built over decades of delivering and securing web traffic. Work that touches the open web benefits from being close to it, and from a provider whose core business is keeping that traffic fast and secure.
Akamai does not build frontier models. Anthropic’s largest compute partners, including Amazon and Google, also build their own. A supplier that competes with none of your products is a simpler partner, and one more credible supplier strengthens every other negotiation.
Anthropic now shares Akamai’s upside. The warrant gives Anthropic a stake in the supplier it is paying. If the deal lifts Akamai’s value, as the 17% move suggested, part of that gain comes back to the buyer. It is an unusual sweetener in a cloud contract.
Who else could Anthropic have chosen?
No bidding process has been disclosed, so this is our assessment of the realistic alternatives, what each one brings, and where Akamai probably had the edge.
| Alternative | Their edge | Where Akamai likely had the edge |
|---|---|---|
| AWS | Anthropic’s primary cloud partner, enormous capacity, its own AI chips | A supplier that does not build competing models, and CPU capacity from a provider whose headline investment is elsewhere |
| Google Cloud | Deep existing ties, TPUs, a large compute commitment already in place | Neutrality: Google builds Gemini, a direct competitor |
| Microsoft Azure | Enterprise reach and an existing strategic partnership | Neutrality again, plus a network built around the open web |
| Oracle Cloud | An aggressive bidder for large AI capacity contracts | A distributed footprint and decades of web security |
| Cloudflare | The closest like-for-like: edge compute, security, a developer platform | General-purpose cloud capacity at this scale, through Linode (our reading) |
| GPU clouds such as CoreWeave | Specialists in AI infrastructure | This deal is for CPUs, outside their core |
| Budget clouds such as DigitalOcean, Vultr or Hetzner | Low unit cost for plain CPU capacity | Global reach, security and a seven-year partner at $11.6 billion scale |
| Building its own data centres | Full control of cost and hardware | Speed: capacity that already exists, in thousands of locations |
Read across the right-hand column and a pattern appears. Akamai rarely wins on size or price. It wins on a combination no single rival offers: distribution, security, neutrality and an equity stake. That combination is a position.
How might positioning have helped Akamai win?
Positioning decides which deals a company is even considered for. Here is how Akamai’s choices likely put it on this shortlist. This is our reading; Akamai has not described its sales process.
It chose a category where it could lead. Competing as “another cloud” puts Akamai against the three largest companies in computing. “Distributed cloud for AI” is a category where its oldest asset, a network in thousands of places, becomes the headline. The Inference Cloud launch in 2025 named that category publicly, a year before this deal.
Its story was continuous. Content delivery, then security, then Linode compute, then inference at the edge, then agents. Each step extended the last, so a buyer could believe the next one. A pivot that contradicts a company’s history needs far more proof.
Its proof was checkable. “Thousands of points of presence” and a web security business built over decades are facts a buyer’s engineers can test. A claim to be “AI-native” gives them nothing to test.
It used what it is not. Akamai builds no frontier models and competes with none of Anthropic’s products. For a buyer whose other suppliers are also rivals, that absence is an advantage, and Akamai’s positioning let the buyer see it.
It put its own value on the line. A warrant tied to spend tells the buyer the supplier expects the relationship to grow and is willing to share the upside. The deal structure itself became part of the pitch.
How did Akamai earn a place on the shortlist?
Akamai spent four years changing what buyers could consider it for. The sequence matters more than any single announcement:
- March 2022: Akamai completed its acquisition of Linode, a developer-focused cloud provider, for about $900 million (Akamai). A security and delivery company gained general-purpose compute.
- October 2025: Akamai launched Akamai Inference Cloud, described as “expanding inference from core data centers to the edge of the internet” (Akamai 8-K, October 2025).
- September 2026: Anthropic’s commitment, framed around distributed CPU workloads.
Each step moved Akamai from “the company that delivers websites” towards “the distributed cloud for AI”. By the time a buyer needed large, spread-out, general-purpose compute, Akamai was a credible name for it. A company is only chosen for things buyers already believe it can do. That belief is built at the category entry points where buyers first think of a supplier.
What can B2B companies learn from the Akamai deal?
Five lessons carry over to any company selling against larger competitors.
- Pick the contest where the giants are weakest. Akamai competed on CPUs and distribution, where the hyperscalers’ capital is least focused.
- Name your real alternatives. The shortlist for this deal was a handful of providers. Positioning starts with the options your buyer is weighing.
- Make your difference something a buyer can check. “Thousands of points of presence” and decades of web security are facts a buyer can verify. “Innovative” gives them nothing to check. Our note on why explanation is not positioning goes further.
- Change the category before you need it. The Linode acquisition came four years before the deal it helped win. Buyers had time to update what they believed about Akamai. Our B2B repositioning guide sets out that sequence.
- Use neutrality as a position. Sometimes the strongest argument is what you are not: not a competitor to your customer, not tied to one of their rivals.
We cover the method behind these lessons in B2B brand positioning, explained and how to map your competitive alternatives. For what happens when the category shift is announced before the business changes, see the positioning mistakes we see most.
Final thoughts
Akamai won a contest it had spent years setting up, for a job where its oldest strengths suddenly mattered. Most B2B companies will never sign an $11.6 billion contract, but the move is the same at any size: find the job your largest competitors are least built for, and become the obvious name for it before the buyer arrives.
For the brand side of that work, read why AI search gets your company wrong and what a brand promise is.
If you are repositioning against bigger competitors, talk to us about brand positioning.
Frequently Asked Questions
Swathi Mohan
Swathi writes sharp, smart copy, sometimes poetic. Quick on her feet, she has a knack for making people feel heard.