Buy this semiconductor stock that could soon overtake Intel in CPU market, Raymond James says

Raymond James upgraded AMD to Strong Buy from Outperform and raised its price target to $641 from $565, implying ~40% upside from Monday’s close. The firm argues AMD’s datacenter positioning and market-share gains (Q2 x86 share >30% vs Intel’s 69.7%) could help it overtake Intel in CPU market share by 2027, with AI-driven CPU growth (including agentic AI) as the “principal new growth engine.” AMD shares rose more than 2% after the upgrade, with the stock up 113% in 2026 but down ~2% over the past three months.

Analysis

The real implication is not that AMD is becoming "better" in a generic sense; it is that CPU share gains in data-center workloads are turning from a product-cycle story into a durable earnings-mix story. That matters because CPUs sit upstream of AI infrastructure economics: if more inference, orchestration, and agentic workloads are CPU-heavy than the market assumes, AMD gets paid twice — higher unit share and better datacenter mix — while peers built around GPU-only narratives may see some budget reallocation at the margin.

The market is likely underestimating the second-order effect on Intel's pricing power. Once share losses become visible in enterprise procurement cycles, Intel usually has to defend with incentives, which can preserve unit volume but compress gross margin and delay the inflection investors are waiting for. The biggest near-term risk for AMD is that the stock has already rerated into near-perfection; any evidence of slower datacenter share gains, delayed product ramps, or softer AI capex could trigger multiple compression even if fundamentals remain sound.

Time horizon matters: over the next 1-3 months this is mostly a sentiment and relative-performance trade, not a clean fundamental reset. The more important catalyst path is through 2027 procurement cycles and 2026 guidance revisions, where server ASPs, socket share, and datacenter operating margin determine whether this becomes a structural winner or just a strong cyclical name. The thesis is falsified if Intel stabilizes share, AMD's datacenter growth decelerates for two consecutive quarters, or management starts speaking about pricing pressure rather than mix expansion.

Contrarian view: consensus is assuming CPU TAM expansion will be broadly monetized, but the market may be overpaying for that optionality after a 100%+ run. The better expression may be relative value rather than outright beta — AMD can keep winning while the stock still underperforms if expectations outrun execution. In other words, the business case is constructive, but the entry point is not obviously compelling without a pullback or a fresh confirmation print.

Key Decisions for Investors

• Maintain/Initiate a relative-value long AMD / short INTC pair over the next 1-3 months. The thesis is that share gains and mix improvement should keep AMD's earnings revisions upward while INTC remains constrained by pricing defense and slower credibility recovery. Risk: if Intel's next guide shows stable server share or margin discipline, the pair can mean-revert quickly.
• Use pullbacks in AMD to add exposure rather than chasing strength. Best entry is on post-rally consolidation or after an earnings volatility event; the setup is favorable only if datacenter revenue and gross margin continue to inflect. Falsifier: two consecutive quarters of decelerating data-center growth or weaker-than-expected forward guidance.
• For investors who want upside with less multiple risk, prefer an AMD call spread 3-6 months out rather than stock. This captures the rerating if the market extends the share-gain narrative while limiting drawdown if the stock stays "perfectly priced."
• Watch INTC as a hedge against the AI-capex crowding-out narrative. If AMD continues taking socket share, the downside is not just lost revenue; it is potential gross margin compression from defensive pricing. A decisive improvement in INTC server commentary would be the cleanest signal to reduce AMD relative longs.