The upside case depends on converting AI design wins into repeatable, production-standard rack deployments with improving software portability. If that happens, a higher-quality revenue mix supports durable compounding into 2030.

AMD can compound equity value through 2030 by scaling from “credible #2 silicon” into a repeatable rack-scale AI platform (CPU+GPU+networking+software), monetizing multi-sourcing and sovereign/enterprise demand while adding higher-margin software/services attach that reduces pure chip-cycle dependence.

Compute becomes the bottleneck; AMD is one of few scaled suppliers of heterogeneous compute and can win as the strategic second platform where buyers want optionality from a single-vendor stack.

By 2030, the bull case is not “more PCs,” it’s AMD proving it can deliver a full, production-standard AI rack alternative at scale: silicon cadence (Instinct + EPYC), cluster-level reference designs, and enough software maturity and tooling that large buyers can multi-source without paying a large productivity tax. If that conversion happens, revenue mix shifts structurally toward data center (higher ASPs, longer programs, stickier platform decisions) and AMD earns a more durable premium than a typical cyclical CPU vendor. The additive opportunities (OPEX pods, enterprise marketplace/clearing layer, chiplet/IP licensing, security/verification attach) matter mainly as mix and predictability improvements—less as standalone giants—because they stabilize demand signals and improve customer time-to-scale. Against comparables, AMD’s multiple plausibly remains below the category leader but above “mid-cycle semis” if it becomes the default second platform for large-scale AI infrastructure procurement.

The swing factor is platform conversion: turning design wins into repeatable, supportable, high-utilization production AI racks without margin leakage. The second-order risks are policy whiplash (exports/tariffs/fees), supply-chain bottlenecks (HBM/advanced packaging), and buyers’ increasing leverage via in-house accelerators and multi-vendor procurement.

The core thesis is consistent, but it is now anchored more explicitly on management’s long-term data center and rack-scale roadmap targets and recent proof points in platform partnerships.

Confidence improved modestly due to clearer long-term targets and stronger recent execution, supporting a slightly higher 2030 premium multiple assumption.

$239.97