The Zacks Analyst Blog Highlights NVIDIA and Advanced Micro Devices

NVIDIA’s fiscal Q2 2027 results showed $96.2B revenue (+106% y/y, +18% q/q) with Data Center revenue of $89B (+117% y/y, +18% sequential) and GAAP/non-GAAP margins at 75% (vs ~72.5% a year ago). The article highlights NVIDIA guiding fiscal Q3 2027 revenue of $108B ±$2B and says its Vera Rubin platform is ramping, positioning NVDA as the “better buy now” versus AMD, which guided Q3 2026 revenue to $13B ±$300M with a non-GAAP gross margin of 56%. Valuation is framed as NVDA trading at a forward P/E of 23.51 vs AMD’s 64.21, with Zacks ranks of NVDA #2 (Buy) and AMD #3 (Hold).

Analysis

This is more of a confirmation of the AI capex hierarchy than a fresh fundamental inflection. The cleanest second-order winners remain the picks-and-shovels with constrained supply and high content per rack: advanced packaging, HBM, and networking. If AMD keeps ramping, that is not immediately bearish for the space; it broadens the installed base and validates demand, but over 6-18 months it also creates a path to price competition that could cap the top-end multiple on the incumbent.

The immediate market risk is expectation saturation. NVDA now needs to keep converting demand into unusually high sequential growth and margin durability; any deceleration in Data Center growth or gross margin normalization would drive multiple compression even if absolute numbers stay excellent. For AMD, the issue is less revenue and more monetization quality: early ramps often look good on revenue but disappoint on mix, software support, and operating leverage, so the stock is vulnerable if gross margin or earnings leverage lags the optimistic ramp narrative.

Consensus is probably too focused on headline valuation ratios and not enough on cash conversion consistency. NVDA still looks like the higher-quality compounder, but the trade is better framed as relative value than outright direction: AMD can continue to work if the market believes it is gaining credible share, yet it needs proof beyond commentary. The thesis is falsified if NVDA’s next guidance implies materially slower sequential Data Center growth, or if AMD fails to translate its ramp into sustained margin expansion over the next 1-2 quarters.

Key Decisions for Investors

• Long NVDA / short AMD as a 1-3 month relative-value pair; thesis is that NVDA keeps earning quality while AMD must prove software and deployment breadth. Cover the short if AMD gross margin holds above the mid-50s and Data Center growth stays above 15% q/q for two straight prints.
• Buy NVDA call spreads into the next earnings cycle rather than outright stock if you want exposure without paying full multiple risk; the upside is continued estimate revisions, while the risk is post-print multiple compression on any guidance cadence slowdown.
• Use AMD weakness to add only if the company shows a second consecutive quarter of accelerating Data Center mix with margin expansion; otherwise keep AMD as a watchlist name, not a core long.
• Overweight the AI infrastructure beneficiaries with less execution risk than GPU shares — especially ANET and MU — as a way to express continued AI capex without paying the full idiosyncratic valuation premium on the chips themselves.