Better AI Infrastructure Stock: Nvidia vs. AMD

AI infrastructure spending is projected to reach ~$750B this year and exceed $1T next year, underpinning continued demand for data center chips. Nvidia outperformed in its latest quarter with total sales rising to $96.2B (vs. $92.1B consensus) and adjusted EPS of $2.22 (+120% YoY), alongside Q3 revenue guidance of ~$108B (+89% YoY) as data center revenue grew 117%. The article argues Nvidia offers better valuation than AMD, citing Nvidia’s ~29x P/E versus AMD’s 121x P/E, making it the preferred long-term AI infrastructure pick despite both companies benefiting from GPU/CPU/agentic AI demand.

Analysis

The market implication is less about "AI is growing" and more about which vendor captures the economics of that growth. NVDA remains the cleaner monetization vehicle because its product stack is closer to the budget owner and its share supports pricing power; AMD is a beneficiary too, but its upside is more dependent on continued share gains and on a second wave of CPU demand that is still more narrative than cash flow. That makes AMD more exposed to any slowdown in agentic AI deployment or any proof that hyperscalers are substituting custom silicon instead of broadening accelerator demand.

Second-order winners sit one layer down the stack: advanced packaging, HBM memory, and foundry capacity should keep taking share of every incremental dollar of AI capex, while networking and power infrastructure benefit from the physical bottlenecks of scaling clusters. The near-term risk is that the same capex intensity that supports the semis trade eventually becomes a drag on hyperscaler free cash flow and multiple expansion, especially if ROI metrics from AI workloads remain opaque. In that case, the first place the market will punish is the higher-multiple, more execution-sensitive names.

The contrarian read is that the consensus may be underestimating how much of the "AI spend" story is already crowded into the whole semiconductor basket, while still underpricing NVDA relative to AMD on both growth and quality. The bigger vulnerability is not a collapse in demand, but a deceleration in the rate of capex growth over the next 1-3 quarters. If management commentary starts to imply budget discipline or margin pressure, the relative-value trade should unwind quickly.

Key Decisions for Investors

• Long NVDA / short AMD as a 3-6 month relative-value pair. The spread trade benefits if AI capex stays strong but the market continues to reward scale, pricing power, and cleaner execution. Falsify if AMD data-center share gains accelerate meaningfully or NVDA guide/margins miss.
• If already long semis, rotate marginal exposure from AMD into NVDA on any post-earnings pullback. AMD's richer multiple leaves less room for disappointment; NVDA offers better downside protection if the AI buildout merely normalizes rather than reaccelerates.
• Watch-and-alert on hyperscaler capex commentary from MSFT, AMZN, GOOGL, and META over the next 1-2 earnings cycles. Any signs of ROI scrutiny or deferred orders would be the fastest reversal signal for the whole AI infra trade.
• Consider a barbell long in NVDA plus indirect picks-and-shovels exposure in TSM or MU rather than adding to AMD alone. This captures both GPU demand and the bottlenecks that accrue to the supply chain, with less valuation risk than buying the higher-P/E compounder outright.