Best Artificial Intelligence (AI) Stock to Buy Now: Nvidia vs. Palantir

Palantir reported revenue growth of ~70% YoY and Nvidia ~73% YoY, but Palantir trades at ~114x forward earnings versus Nvidia at ~22.4x. The article argues Palantir's subscription software model is more sustainable, yet its valuation embeds several years of growth, making Nvidia the better buy given similar growth and a far lower multiple. Author disclosure: Keithen Drury holds Nvidia; The Motley Fool holds/recommends both and did not include Palantir in its Stock Advisor top 10.

Analysis

Nvidia’s dominant position in the AI compute stack creates a multi-year hardware replacement and build-out wave that cascades value to fabs, HBM vendors, and datacenter infrastructure — this isn’t a one-time capex spike but a drawn-out replacement and densification cycle that will keep gross margin leverage for incumbents if unit pricing holds. Palantir’s platform sits at the “last-mile” of decisioning where stickiness and high marginal revenues can compound, but that premium depends on demonstrable automation-led ROI across multiple large commercial verticals, not just defense; failure to scale commercial ARR or convert pilots into platform-wide deployments is the single largest value inflection risk. Second-order winners from continued AI expansion include TSMC/ASML (capacity and advanced nodes), memory/HBM suppliers (SK Hynix, Micron), and colocation/cooling vendors; losers include legacy systems integrators and narrow-stack analytics firms that get commoditized by integrated platform + LLM solutions. Timing matters: monitor datacenter build signals and hyperscaler purchase cadence over the next 3–12 months as the earliest reveal of sustained hardware demand, while Palantir’s re-rating hinges on sequential commercial net expansion and multi-year contract roll-offs over 12–36 months.

Key Decisions for Investors

• Pair trade (6–18 month horizon): Long NVDA (1.0–1.5% net portfolio notional) / Short PLTR (0.5–1.0% notional). Rationale: capture asymmetric upside from continued AI capex while shorting a stretched software multiple. Target: NVDA +40–80% / PLTR -30–50%; hard stop if NVDA declines >25% or PLTR rises >40% from entry.
• Options play (12–24 month horizon, directional): Buy NVDA Jan-2027 LEAP calls (near-the-money, delta ~0.55–0.65) sized to 1% portfolio risk. Upside scenario: AI build-out continuation -> >2x payoff; downside: premium lost (defined max loss). Use premiums as controlled call on multi-year secular demand.
• Hedge against Palantir downside (12 month horizon): Buy PLTR put spread (buy longer-dated put / sell lower strike put) sized to 0.25–0.5% portfolio. Rationale: protects against multiple compression/contract renewals while capping premium. Target payoff: 3–6x premium if PLTR compresses 30–50%; max loss = paid premium.
• Tactical infrastructure exposure (9–24 months): Overweight advanced-node supply chain (TSM, ASML, HBM suppliers) by 1–2% vs broad tech. Rationale: hardware build-out amplifies these suppliers’ margins with more predictable orders; take profits if quarterly fab guidance or HBM pricing weakens. Monitor TSMC capacity commentary and HBM spot pricing weekly as triggers to rebalance.