Intel is running two turnarounds simultaneously, either of which would be difficult alone: rebuilding a competitive process technology, and finding a place in AI compute after arriving late.
The accelerator story has been the weaker half. Gaudi 3 never found volume, and Intel has repeatedly reset the roadmap — Crescent Island is now positioned for inference rather than training, which is an honest read of where it can actually win.
The process story matters more, and not only to Intel. It is the only American company still fabricating leading-edge logic on home soil, which turns a corporate recovery into a question of industrial policy.