Better Buy After the Chip Sell-Off: Nvidia or AMD?

The article argues Nvidia is the better buy after the semiconductor sell-off, citing 85% year-over-year revenue growth to $81.6 billion, 75% adjusted gross margin, and a forward P/E of 23 versus AMD's 74. AMD also posted solid results, with revenue up 38% to $10.3 billion and data center revenue up 57%, but its growth and margins still lag Nvidia's. The piece is fundamentally positive on Nvidia relative to AMD, though it is primarily valuation and comparison commentary rather than fresh company-specific news.

Analysis

The key market signal is not that semis are weak, but that the market is beginning to discriminate between monetized platform power and “AI exposure” as a generic label. NVDA’s combination of accelerating growth, unusually high margins, and capital returns means it can absorb a multiple reset better than peers; that creates a relative-value anchor for the entire AI hardware complex. By contrast, AMD’s problem is not just valuation in isolation, but that its valuation already prices in a multi-year execution path that requires share gains, sustained customer concentration, and margin expansion all at once.

Second-order, this likely pressures the AI supply chain in two ways. First, buyers will push harder on non-NVIDIA vendors for price concessions and financing terms, which can help win sockets but compress long-run economics for challengers. Second, the market may start rewarding names that sit one layer above pure accelerator exposure—interconnect, packaging, networking, power infrastructure—because those businesses can capture AI spend without requiring the same level of platform dominance or valuation perfection.

The contrarian read is that NVDA is not “cheap” in an absolute sense; it is simply less expensive than the market’s fear premium on AMD. If AI capex decelerates even modestly over the next 1-2 quarters, AMD’s high multiple can de-rate much faster than NVDA’s because the former is priced for flawless adoption while the latter already reflects a more realistic growth trajectory. The real risk to the bullish NVDA view is not competition from AMD, but a broader digestion phase in hyperscaler spending or further export restrictions that temporarily cap revenue upside for 1-2 quarters.

Key Decisions for Investors

• Long NVDA / short AMD as a 3-6 month relative-value pair: NVDA has the better earnings durability and a much lower multiple; target 15-25% spread capture if AI hardware multiples compress further.
• Sell AMD upside volatility into strength, preferably via call spreads 2-4 months out; the stock is priced for near-perfect execution, so upside is more sensitive to any guidance miss than NVDA.
• Buy NVDA on further weakness in tranches rather than chasing intraday rebounds; use a 2-3 week accumulation window because the stock can still de-rate with sector flows even if fundamentals remain intact.
• Add to AI infrastructure beneficiaries with lower single-name risk than accelerators, especially networking/power/advanced packaging exposures, as a hedge against valuation compression in GPU leaders over the next 6-12 months.
• For event-driven protection, hedge NVDA long exposure with short-dated puts around next guidance cycle if consensus has become crowded; the main downside catalyst is not competition, but a pause in hyperscaler capex sentiment.