Akamai Technologies — AI supply-chain exposure

Akamai grew up as a content-delivery network — the distributed edge that caches web content close to users so pages load fast. It is now repositioning that footprint into a compute platform, using its Linode acquisition to place GPU capacity near where requests originate. The intent is inference serving at the edge: running already-trained models to answer queries with low latency, rather than doing the heavy training itself. That puts Akamai in the chain as a producer of inference capacity.

The structural bet is that a meaningful share of inference wants to run close to users — for latency, cost and data locality — instead of only in a handful of centralised hyperscale regions. Akamai's advantage is a geographic distribution and set of peering relationships that are slow to replicate. Its exposure is that it is a challenger to far larger cloud incumbents in AI compute, so the thesis rests on edge inference becoming a distinct, durable slice of the market.