Better Buy After the Semiconductor Sell-Off: Marvell or Broadcom?

Broadcom reported Q2 fiscal 2026 revenue of $22.2B, up 48% year over year, with AI semiconductor revenue surging 143% to $10.8B and management guiding AI chip sales to $16.0B in Q3 and over $100B in fiscal 2027. Marvell also posted strong growth, with Q1 fiscal 2027 revenue up 28% to $2.4B and guidance for about 40% full-year growth, but its valuation near 90x earnings looks rich versus Broadcom's roughly 64x. The article argues Broadcom is the better buy after last week’s selloff, despite both stocks remaining highly sensitive to AI spending expectations.

Analysis

The market is starting to discriminate between “AI exposure” and “AI monetization quality.” Broadcom’s setup is stronger because the earnings base is already large enough that incremental AI mix can translate into visible free-cash-flow expansion, while Marvell still needs flawless execution across a narrower customer set to justify its multiple. In other words, AVGO has moved from narrative to self-funding growth; MRVL is still priced like a perfection story.

A second-order effect is that Broadcom’s strength reinforces the idea that the real bottleneck in AI infrastructure is shifting from GPUs to everything around them: custom accelerators, interconnect, and data-center networking. That is good for NVDA on the ecosystem side, but it also increases competitive pressure on smaller ASIC and networking vendors as hyperscalers consolidate spend into fewer suppliers with proven scale, qualification, and supply-chain resilience. The winners are likely the vendors that can bundle silicon, software, and long-term capacity commitments; the losers are subscale pure-plays that depend on one or two design wins.

The contrarian read is that the selloff may not fully reflect how sensitive both names are to capex digestion. The market is extrapolating current booking momentum several quarters forward, but AI infrastructure spend tends to come in waves; if one hyperscaler pauses deployment, Marvell’s growth narrative can compress quickly, and Broadcom’s valuation multiple is still exposed to any sign that 2027 revenue guidance is too aggressive. The key tell over the next 1-2 quarters is not headline growth, but whether AI revenue converts into sustained margin and FCF expansion without a step-up in working capital or customer concentration risk.

Key Decisions for Investors

• Long AVGO / short MRVL for 1-3 months: favors the name with faster growth, stronger cash conversion, and a lower multiple; thesis breaks if MRVL re-accelerates beyond current guidance while AVGO guidance stalls.
• Sell MRVL downside puts or use a put spread into the next earnings cycle: volatility is elevated and the valuation leaves little room for a design-win miss; target 2-1 or better risk/reward.
• Add to NVDA on pullbacks as the ecosystem beneficiary, but size modestly: strongest second-order winner if AI networking and custom silicon spend remain durable over the next 6-12 months.
• For conservative exposure, prefer AVGO common over MRVL calls: AVGO offers better downside support via FCF, while MRVL needs multiple expansion to outperform from here.
• Monitor hyperscaler capex commentary over the next 1-2 quarters; if spending flattens, reduce both longs quickly, with MRVL the first trim due to higher customer concentration risk.