Jensen Huang Maintained His Prediction That AI Infrastructure Spending Will Hit $3 Trillion to $4 Trillion by 2030 -- the Same Forecast He Made a Year Ago to the Day. Is Nvidia Still a Buy at These Levels?

Nvidia CEO Jensen Huang reiterated his forecast for global AI infrastructure spending to reach $3 trillion to $4 trillion by 2030, supporting the company's long-term growth case. Nvidia reported 106% year-over-year revenue growth in fiscal 2027 Q2 and has guided for 70% growth in fiscal 2028, while broadening its business from GPUs into Vera Rubin supercomputing systems, strategic investments and a $12.9 billion Hugging Face acquisition. The article argues that Nvidia's 27x P/E remains attractive despite substantial AI-growth expectations already being reflected in the share price.

Analysis

The investable issue is not whether AI capex expands, but whether NVDA can convert increasingly turnkey systems into durable incremental gross profit rather than financing demand through its ecosystem. Full-rack architecture raises switching costs and broadens content per deployment, but it also shifts the bottleneck to customer power, data-center energization, networking, and depreciation economics. That makes reported accelerator demand a less clean read-through to end-user AI monetization over the next 2-4 quarters.

The CoreWeave relationship warrants particular scrutiny: vendor investments and customer purchase commitments can pull forward orders while increasing credit, concentration, and residual-value risk. If neocloud utilization or financing costs deteriorate, the effect is asymmetric—NVDA can retain revenue initially but face slower repeat purchases, receivable pressure, and a valuation reset when the market questions the quality of demand. CRWV is therefore a higher-beta expression of sustained GPU scarcity, but also the more direct downside vehicle if AI capacity is oversupplied.

Consensus appears to treat platform integration solely as a moat. The counterpoint is that vertically integrated rack sales increase exposure to supply-chain execution and create openings for AMD, AVGO custom silicon, and networking vendors where hyperscalers seek bargaining leverage. Over 6-18 months, AI spending may remain robust while the value pool rotates from training hardware toward inference optimization, power infrastructure, and custom accelerators; NVDA can still grow, but its multiple will depend on preserving gross-margin resilience rather than simply delivering revenue growth.

Near term, this is not a clean incremental long catalyst from promotional commentary. A constructive stance requires independent confirmation that new-system backlog is backed by end-customer deployments and that customer financing has not become a material demand support. Thesis falsification is a sequential deceleration in data-center revenue coupled with gross-margin guide-down, elevated receivables/investment balances, or CRWV utilization and funding stress over the next two earnings cycles.

Key Decisions for Investors

• Maintain NVDA as a core long only against a 2-quarter verification framework; add on post-earnings weakness only if data-center revenue and gross-margin guidance remain intact and management quantifies cash-backed backlog. Exit or hedge if gross-margin guidance falls materially while receivables or strategic-investment balances rise, signaling lower-quality revenue.
• Express the financing-risk divergence as long NVDA / short CRWV over 1-3 months, sized modestly given NVDA's index support. The pair benefits if hyperscaler demand proves durable but neocloud economics weaken; cover the CRWV short if utilization, contracted backlog, and financing access improve materially without incremental vendor support.
• Watch for a relative-value rotation basket: long AVGO and ANET versus a reduced NVDA overweight over 6-18 months if hyperscalers disclose rising custom-silicon deployment or rack-scale networking content. Do not initiate solely on narrative; require evidence in hyperscaler capex commentary and supplier order trends.
• Set an event alert around NVDA's next earnings: a revenue beat without forward gross-margin support is not bullish confirmation. In that outcome, use liquid NVDA puts or a short-dated collar to protect gains rather than adding exposure, as multiple compression can dominate continued earnings growth.