Broadcom and AMD Face Valuation Pressure Amid AI Capex Slowdown Fears, Yet Challenger Thesis Holds

Broadcom and AMD stocks have declined sharply on concerns about AI infrastructure spending deceleration, but analyst support and ecosystem partnerships suggest the challenger-chip thesis remains intact despite near-term sentiment headwinds.

What changed

Broadcom (AVGO) has fallen 3% to $381 as of mid-July 2026, extending a broader semiconductor selloff driven by AI capital expenditure slowdown concerns. Marvell Technology dropped 8% to $189 on the same day, with the broader chip complex under pressure. AMD (AMD) has also declined modestly, down 1.3% over the past 30 days to $500.94 as of the market snapshot date.

Despite the near-term weakness, Wall Street analysts continue to see upside: one Wall Street analyst flagged Broadcom as a "core AI winner," and the average of analyst price targets indicates potential upside of 32.8% from current levels. Broadcom also secured a new partnership milestone: the Cloud Native Computing Foundation (CNCF) elevated Broadcom to Platinum Member status to advance AI-ready cloud-native infrastructure, signaling continued institutional adoption of its custom silicon platform.

AMD's ecosystem strategy remains active, with the company appointing Alan Smith as a new Corporate Fellow, though this appointment carries limited direct bearing on the AI silicon thesis.

Why it matters

The selloff reflects a real near-term risk to the challenger thesis: if hyperscalers are decelerating AI capex spending, demand for custom silicon alternatives to Nvidia may contract alongside GPU orders. The fact that Broadcom and AMD stocks are down despite strong analyst support suggests the market is pricing in a near-term demand pause rather than a fundamental loss of confidence in their competitive positioning.

However, the CNCF partnership elevation is material to the thesis. The CNCF is the steward of Kubernetes and cloud-native standards; its decision to elevate Broadcom to Platinum status indicates that Broadcom's custom silicon is being integrated into the open-source cloud infrastructure layer that hyperscalers depend on. This is distinct from one-off customer wins—it signals that Broadcom's silicon is becoming embedded in the infrastructure standards that cloud providers use to build AI platforms. If hyperscaler capex resumes, Broadcom's silicon will be pre-positioned in the architectural layer, reducing switching costs and competitive friction.

The analyst support (32.8% upside, "core AI winner" designation) reflects confidence that the current valuation weakness is cyclical rather than structural. This matters because the parent thesis hinges on Broadcom and AMD becoming sustained alternatives to Nvidia, not one-quarter wonders. If analysts are willing to defend the thesis through a capex slowdown, it suggests they believe the custom-silicon business model is durable even in a softer demand environment.

AMD's appointment of a Corporate Fellow is largely ceremonial and does not directly advance the challenger thesis, though it may signal internal organizational confidence.

Opposing sources and risks

Multiple sources contradict or weaken the thesis with fairly high certainty:

Nvidia's revenue dominance remains stark. One analyst noted that Nvidia's AI hardware revenue is nearly eight times larger than AMD's, despite AMD shares surging 150% year-to-date while Nvidia has barely moved. This highlights the scale gap between the incumbent and the challengers; even if AMD gains share, Nvidia's installed base and revenue moat remain formidable.

Nvidia is expanding beyond GPUs into CPUs and other silicon. Nvidia's "Vera" CPU is expected to generate $20 billion in sales by the end of fiscal 2026, and Perplexity has announced plans to use Nvidia's new CPU. This suggests Nvidia is not standing still; it is diversifying into the same custom-silicon and CPU markets where Broadcom and AMD are competing. If Nvidia can leverage its GPU dominance to cross-sell CPUs and other silicon, the challenger thesis faces a more entrenched competitor than a pure-GPU vendor.

Broadcom faces EU antitrust scrutiny. Five cloud business groups have urged EU interim measures against Broadcom, and the company faces an antitrust review over VMware licensing changes. If regulatory action constrains Broadcom's ability to bundle or cross-sell silicon to cloud providers, it could weaken its competitive position relative to Nvidia, which faces less direct regulatory pressure in the EU.

Valuation concerns persist despite strong fundamentals. Erste Group downgraded Broadcom to Hold despite an expected $29.4 billion in quarterly revenue and AI strength, citing valuation. This suggests that even bullish analysts believe Broadcom's stock price has run ahead of fundamentals, creating downside risk if growth disappoints.

What to watch

Hyperscaler capex guidance in Q2 2026 earnings. Microsoft, Amazon, and Google will report capex plans in the coming weeks. If they signal sustained or accelerating AI infrastructure spending, the capex slowdown fears will ease and Broadcom/AMD stocks should recover. If they signal a pause, the thesis faces a near-term headwind.

Broadcom's custom silicon revenue growth rate. The parent thesis hinges on Broadcom's custom ASIC business growing "faster than almost anyone expected." The next earnings report should quantify this growth rate and provide forward guidance. A slowdown in custom silicon growth (vs. GPU growth) would weaken the thesis.

AMD's AI GPU market share gains. AMD's MI300 and MI325 GPUs are competing with Nvidia's H100/H200 for hyperscaler orders. Quarterly data on AMD's AI GPU shipments and revenue will indicate whether AMD is winning share or losing ground to Nvidia.

EU antitrust outcome for Broadcom. The regulatory review could take months, but any interim measures or formal charges would materially weaken Broadcom's competitive position and could force divestitures or licensing restrictions.

Nvidia's CPU and custom-silicon revenue contribution. Nvidia's guidance on Vera CPU and other non-GPU silicon will clarify whether Nvidia is successfully defending its moat by expanding into the custom-silicon market. If Nvidia's non-GPU silicon grows faster than Broadcom's or AMD's, the challenger thesis loses credibility.

Related Arbora context

This thesis is distinct from but complementary to the AI infrastructure and data center build-out thesis, which focuses on the overall capex cycle and Nvidia's GPU dominance. The challenger-chip thesis assumes that hyperscalers will diversify away from Nvidia even within a strong capex environment; if the capex cycle itself contracts, both theses face headwinds.

The CPU renaissance and advanced process node competition thesis also overlaps: Broadcom and AMD are competing in custom silicon and CPUs, which are part of the broader CPU renaissance driven by AI workloads. However, the challenger-chip thesis is narrower—it focuses specifically on Broadcom and AMD as alternatives to Nvidia in AI compute, whereas the CPU thesis includes Intel and ARM as well.

The Micron memory chip supercycle thesis is complementary: if Broadcom and AMD gain share in AI accelerators, demand for high-bandwidth memory will rise, benefiting Micron. Conversely, if the capex slowdown persists, Micron will also face headwinds.

Sources

This article is research notes and should not be construed as financial advice.