Better AI Stock: SpaceX vs. Nvidia

The article favors Nvidia over SpaceX as an AI investment, citing Nvidia’s Q2 revenue doubling and expected revenue of about $412 billion versus analysts’ roughly $45 billion estimate for SpaceX this year. SpaceX CEO Elon Musk said the company will build exclusively on Nvidia, while Nvidia’s Vera Rubin architecture is said to cut inference costs 10-fold and training costs fourfold versus Blackwell. The author views SpaceX as higher-growth but less certain and more expensive, noting valuations of about $5.5 trillion for Nvidia and $2 trillion for SpaceX.

Analysis

The investment distinction is less “safer versus faster growth” than transparent hardware economics versus a bundled, capital-intensive platform. SpaceX’s stated Nvidia exclusivity may improve demand visibility at the margin, but without purchase volumes, delivery timing, or pricing it does not establish material incremental earnings for Nvidia. For SpaceX, the same commitment creates supplier concentration and ties xAI’s unit economics to the cost and availability of external compute. A sharp fall in inference cost can expand usage, but it can also intensify price competition and lower revenue per token; usage growth alone is not proof of attractive margins.

The valuation comparison is also incomplete: consolidated SpaceX sales include businesses other than xAI, while the AI opportunity is only one component of its equity story. Treating those sales as an AI denominator risks overstating the apparent premium. Conversely, Nvidia’s lower implied sales multiple is not a margin of safety if customer capex slows or custom accelerators displace workloads. AMD and Broadcom, alongside in-house chips at large cloud providers, remain relevant substitution checks.

Near term, the likely reaction is sentiment-driven; over 1–3 months, verify Nvidia order conversion and SpaceX/xAI segment economics rather than relying on architecture claims or reported growth rates. Over 6–18 months, lower compute costs are bullish only if workload expansion outruns price compression and capital intensity. The thesis weakens on Nvidia guidance/order deterioration, evidence of meaningful GPU substitution, or xAI growth failing to translate into improving monetization and margins.

Key Decisions for Investors

• Prefer NVDA to SPCX for direct AI exposure, but stage entry on weakness rather than chase the bullish narrative. Reassess if Nvidia order conversion or forward guidance deteriorates.
• Do not short SPCX solely on the cited sales comparison: its consolidated revenue is not an xAI-only valuation base. Keep it on watch pending disclosed xAI revenue quality, compute costs, and profitability.
• Treat SpaceX’s Nvidia commitment as a conditional demand signal, not an earnings estimate. Monitor order size, delivery schedule, and whether Nvidia’s next disclosures show incremental demand beyond existing customer plans.
• Track AMD and Broadcom plus cloud-provider custom silicon as substitution indicators; sustained share gains or weaker Nvidia pricing would challenge the hardware-leadership premium.
• For the 6–18 month view, test whether inference-cost reductions produce faster token/workload growth than price declines. If monetization lags usage, reduce exposure to the application-growth thesis.