i
News
News · 2026-09-25

Anthropic adds an $11.6 billion Akamai cloud commitment

@neuronium_ai @neuronium_ai

Anthropic has struck an $11.6 billion cloud deal with Akamai, adding another major computing commitment to a spending run that has reached $517 billion in reported capacity deals in 11 months. The announcement matters less as a standalone supplier win than as evidence of how aggressively AI labs are reserving infrastructure ahead of revenue they expect to earn. A month after committing $45 billion to Nscale, Anthropic is making the same financial bet again: capacity first, commercial proof later.

Cover: Anthropic adds an $11.6 billion Akamai cloud commitment

Anthropic’s commitments keep stacking

The latest deal sits inside a rapidly expanding set of obligations:

$11.6 billion cloud deal with Akamai
$45 billion in computing capacity from Nscale, agreed one month earlier
$517 billion in reported computing-capacity deals over 11 months

Anthropic and OpenAI are reserving capacity on the assumption that future revenue will justify the commitments. That makes the $517 billion figure more than a measure of demand for infrastructure. It is also a bet on how much revenue the companies can generate later.

The risk behind the numbers

Anthropic CEO Dario Amodei warned in December 2025 that the company could go bankrupt if its estimates were even slightly wrong. OpenAI CEO Sam Altman has also recently urged more caution around spending on computing infrastructure, warning that the costs of building it may prove unsustainable.

My read is that Anthropic’s Akamai deal looks less like a single strategic turning point than another escalation in a race where the contracts are arriving faster than the evidence needed to justify them. The total is impressive, but its size does not show how much capacity Anthropic will actually use or how quickly it can turn that capacity into revenue.

That is the question the announcement leaves quiet: what happens if demand grows, but not fast enough to pay for the reservations? The danger is not only that one forecast misses. It is that several large commitments were made on the same optimistic forecast.

Daily AI news

Every day we pick what actually matters in AI and explain it plainly — no hype, no filler. Subscribe if you want to follow where the industry is going.

Only what matters — every day

Follow on X