$Advanced Micro Devices(AMD)$ What's interesting is how the narrative is expanding beyond just GPUs.

Looking at the longer-term setup, some estimates are floating the idea that AMD could scale toward ~$100B annual revenue by 2028. This is driven by multi-vendor AI commitments from $Meta Platforms, Inc.(META)$ , $Oracle(ORCL)$ , and ecosystem demand tied to OpenAI infrastructure buildouts.

The initial rerating came from the GPU side, but the second leg of the story is starting to come from CPUs, which often get overlooked in AI conversations.

Every AI cluster still needs heavy general-purpose compute for orchestration and inference, and that's where EPYC continues to gain share, already around ~46% of server CPU revenue.

If that trend continues, the CPU cycle becomes less of a side story and more of a compounding driver, especially with next-gen Venice ramping on $Taiwan Semiconductor Manufacturing(TSM)$  2nm.

It's becoming a two-engine AI compute story, not just one.