Nvidia Isn't The Expensive AI Stock Anymore—AMD, Palantir And ARM Are

Nvidia’s valuation has compressed to ~30x trailing earnings (~22x forward), well below several AI peers (e.g., AMD ~77x, Palantir ~79x, Arm >156x), suggesting investors are paying smaller premiums for the current AI leader. The article notes AI investing is becoming more selective as investors prioritize converting AI demand into durable earnings, with capital rotating toward memory and AI infrastructure names. Overall, the piece is more positioning/valuation-focused than a fundamental earnings catalyst, implying limited near-term price impact.

Analysis

The key signal is not that NVDA is cheap, but that the market is using it as the “cash flow anchor” inside AI while pricing adjacent names as long-duration options. That usually happens late in a thematic cycle: winners with visible earnings become the lower-multiple names, while the highest multiples migrate to companies with more narrative and less near-term proof. In the next 1-3 months, that creates a valuation trap for ARM and PLTR if earnings revisions do not stay ahead of price.

Second-order effects favor the AI supply chain over the most richly valued software/semis. If investors keep demanding proof of monetization, capital should continue rotating into picks-and-shovels where revenue is tied to capex budgets rather than sentiment: memory, foundry, and semiconductor equipment remain better positioned than names whose upside depends on multiple expansion. NVDA can also benefit relative to peers because it owns the strongest balance-sheet/cash-generation profile, which matters if real yields stay sticky.

The contrarian view is that the consensus may be underestimating how durable NVDA’s earnings power is versus the smaller AI “story” names. If Blackwell demand or inference spend re-accelerates, NVDA can grow into its lower multiple while the expensive peers de-rate on even modest execution misses. What would falsify this is a broad AI spend slowdown or a quarter where NVDA guide-downs while ARM/PLTR keep surprising higher on revenue and operating leverage.

Key Decisions for Investors

• Relative-value: long NVDA / short PLTR into the next earnings cycle; use a 1-3 month horizon. Best if the market continues rewarding earnings visibility over narrative. Falsifier: PLTR revenue growth re-accelerates while NVDA data-center bookings soften.
• Tactical short: buy put spreads on ARM over 6-10 weeks if it remains priced like a perpetual compounder. This is a multiple-compression trade, not a fundamental collapse call. Falsifier: any upward revision to AI silicon content per device or a sharp revenue beat.
• Stay overweight NVDA versus the broader AI basket for 6-12 months. It is the cleaner quality-duration trade inside AI if investors keep demanding cash conversion. Risk/reward improves on 3-5% pullbacks, not after momentum extensions.
• Do not chase the richest-multiple AI names here; wait for earnings revisions to confirm. If forward estimates fail to rise over the next two reporting cycles, expect a fast de-rating across ARM/PLTR even if the AI theme stays intact.