AMD earnings forecast momentum points to a broader AI buildout, with server CPUs in the spotlight

Advanced Micro Devices’ (NASDAQ: AMD) stock has been benefiting from a wave of estimate upgrades, a pattern that suggests analysts see more upside to the company’s AI exposure. In recent weeks, the forward-looking earnings picture has strengthened, with estimates for the upcoming fiscal year showing many more upward revisions than downward ones.

A widely cited screen of analyst estimate changes shows 42 upward EPS revisions and 3 downward revisions over the past three months for AMD’s next fiscal year. On the revenue side, the same tally points to 43 upward revisions and 2 downward revisions, reinforcing the idea that the upgrade momentum is not limited to profit forecasts alone.

The catalyst behind the market’s evolving AI framing is not just that customers want GPUs for accelerated workloads, but that CPUs for servers are increasingly part of the same AI systems. A recent report cited by the coverage links the upgrade activity to demand for AI-optimized server CPUs, and it argues that AI demand is widening to “agentic” uses, broader cloud buildouts, and enterprise modernization rather than being confined to the most visible accelerator buying cycles.

In that context, Wells Fargo’s commentary, as relayed in the coverage, ties AMD’s server CPU outlook to a large growth projection for 2026. The report said Wells Fargo raised its AMD price target to $615 from $505 while keeping an Overweight rating, attributing the move to expected server CPU revenue growth. It specifically projected AMD server CPU revenue to grow 68% in 2026 to $16 billion.

The coverage also underscores a key product distinction that helps explain AMD’s “CPU in the AI stack” thesis. AMD designs EPYC server CPUs, which are used in enterprise and data-center servers, while its Instinct GPUs are designed for high-performance parallel compute workloads. The argument presented is that AI system demand can translate into CPU purchasing alongside accelerator purchasing, and that this can make the earnings revision story feel more durable.

Beyond the immediate estimate changes, the coverage points to additional leverage AMD has in data-center platforms and rack-scale AI systems, implying that AMD’s role can expand at the system level rather than staying purely at the component level. That theme matters for a semiconductor supplier because the economics of selling into rack-scale deployments can differ from selling chips alone, and analysts typically look for evidence that platform share is improving.

Still, not all the specifics behind the upgrades are fully disclosed in the brief market coverage. The post does not provide a full breakdown of contract wins, backlog, or customer concentration, and it does not clarify what portion of the projected server CPU growth is driven by new designs versus incremental adoption within existing deployments.

For investors and industry watchers, the next signposts will likely be AMD’s own disclosures around data-center and AI-related revenue drivers, as well as any commentary on CPU attach rates to AI accelerators and rack-scale systems. With the estimate revision momentum skewing upward, the question going forward is whether reported results and forward guidance can validate that broader CPU-driven AI demand is broad and sustained, not just a short-term optimism cycle.