Intel’s CPU Comeback Lifts All Boats — Can AMD and ARM Ride the Same AI Wave?

Intel reported Q1 2026 DCAI revenue of $5.1B, up 22% year over year, with overall revenue of $13.6B, up 7%, reinforcing demand for AI infrastructure beyond GPUs. The article highlights a sector rotation toward CPU-centric inference and agentic AI workloads, which helped lift ARM 15% and AMD 14%; AMD also received a Buy upgrade from D.A. Davidson with a $375 target after its record Q4 2025 data center revenue of $5.38B (+34%).

Analysis

This read-through is less about Intel as a standalone beneficiary and more about a regime change in the AI capex stack: compute demand is broadening from accelerator-led training into an installed-base upgrade cycle for CPUs, memory, networking, and software orchestration. That matters because it shifts incremental economics toward vendors with higher attach rates to the server bill of materials rather than purely the GPU leader; in practice, the next leg of AI spend is likely to be more diversified, lower-beta, and longer duration than the first wave.

ARM’s move is the cleanest expression of that second-order effect. If inference and agentic workflows are becoming more distributed, ARM’s royalty model gains leverage through every custom silicon design win without needing to outspend competitors on fabs or inventory. The market is likely underestimating the optionality in data center penetration: even modest share gains in server CPUs can compound into a disproportionately large earnings pool because the business scales off unit growth, not manufacturing capex.

AMD is the more direct beneficiary, but the setup is trickier. The stock is already discounting a multi-quarter catch-up trade, so upside now depends on whether server CPU share gains come with margin resilience and supply normalization, not just headline revenue growth. The key swing factor is whether Intel’s supply constraints persist long enough for AMD to harvest incremental sockets, or whether a faster-than-expected Intel recovery caps the duration of the trade.

The market is still likely underpricing the risk that this is a sequencing shift, not a permanent re-rating of all AI hardware. If inference loads move to more CPU-heavy architectures, the immediate losers are not necessarily NVIDIA outright, but adjacent beneficiaries of GPU scarcity and pricing power; the bigger threat is to the assumption that every AI dollar must flow to accelerators. That creates a window for relative-value trades in semis as the market digests which names capture architecture transition versus which names merely trade with the AI basket.

Key Decisions for Investors

• Long AMD / short NVDA via 1-2 month relative-value trade: express the thesis that CPU mix shift and server share gains can outperform a crowded accelerator trade over the next earnings cycle; target 8-12% spread capture with a stop if Nvidia guidance re-accelerates training capex.
• Add ARM on pullbacks, but only as a 6-12 month compounder: the risk/reward is favorable if data-center royalty penetration continues, yet the valuation leaves little room for an AI capex slowdown; size smaller than AMD because multiple compression is the main risk.
• Trade the Intel confirmation effect with a tactical long INTC vs broad semis for 4-8 weeks: if the market keeps rewarding evidence of CPU demand, Intel can continue to mean-revert on sentiment, but trim quickly if investors refocus on execution and margin dilution.
• Use AMD upside calls rather than outright stock into the next print: a defined-risk upside structure captures continued re-rating from data-center momentum while limiting downside if Intel’s competitive response compresses share expectations.
• Do not chase NVDA on this headline; instead, wait for a better entry after any post-earnings or sentiment-driven pullback: the near-term risk is multiple compression from a broader AI hardware rotation, even if the long-term secular story remains intact.