Nvidia Just Guided for 70% Revenue Growth in Fiscal Year 2028. Here's What That Means for AI Stocks.

Nvidia reported $96B in quarterly sales (+106% YoY) and is guiding for $108B in Q3, with CFO Colette Kress projecting 70% revenue growth in fiscal 2028. The article forecasts revenue could exceed $400B in fiscal 2027 and approach ~$700B in fiscal 2028 (implying ~$448B in operating earnings if the 64% operating margin is sustained). It also notes Nvidia raised chip prices by 15% due to higher memory costs, which could pressure customer margins but should lift demand across the AI supply chain.

Analysis

The key mechanism is not just higher AI capex, but who captures the spread between scarce compute demand and the cost to deliver it. Near term, the cleanest beneficiaries are TSM and MU: TSM gets leverage to every incremental wafer/package shipped, while MU benefits from the parts of the stack where supply tightness still supports pricing discipline. ASML is a later-cycle winner; its upside is real, but order conversion will lag the market’s first reaction by quarters rather than weeks.

The weaker leg is the customer base, especially AMZN and MSFT, where AI capex is increasingly a quasi-fixed cost before monetization is fully visible. If they cannot push AI features into higher cloud consumption or pricing, margins and free cash flow can compress even if usage keeps rising. The second-order effect is a margin transfer from hyperscalers to the semiconductor supply chain, which can persist for 1-3 quarters before customers renegotiate architecture choices or slow deployment.

Contrarianly, the market may be over-indexed to NVDA’s headline growth and underweighting the fact that the most durable economics often sit one layer down the chain. A 70% growth path is bullish, but the risk is not demand collapsing; it is mix shift, custom ASIC substitution, and eventual customer pushback on cost per token. Falsifiers are clear: any guide-down in gross margin, slowing order cadence, or hyperscaler commentary that capex is decoupling from revenue growth over the next 1-2 earnings cycles.

Key Decisions for Investors

• Long TSM on pullbacks over the next 2-6 weeks; pair against AMZN or MSFT if you want to isolate the supplier-versus-customer margin transfer. Risk/reward: better fundamental durability than NVDA if AI spend stays elevated, with downside mainly from a broader semiconductor multiple reset.
• Add MU tactically for a 1-3 month trade while pricing power remains supported by AI memory demand; use a tight stop if memory lead indicators turn. Falsifier: any sign that HBM supply normalizes faster than expected or pricing weakens into the next earnings print.
• Avoid chasing NVDA outright at current levels; if expressing bullishness, prefer call spreads into the next catalyst rather than stock. The stock can still work, but the gap between growth and perfection is now the main valuation risk.
• Short AMZN/MSFT only as a relative-value hedge against a long semiconductor basket, not as standalone shorts. The thesis works only if AI monetization lags capex for another 1-2 quarters; cover if cloud commentary shows AI services driving incremental ARPU or margin expansion.
• Set an alert for any NVDA gross margin compression or reduced forward revenue growth cadence; that is the earliest sign the market should de-rate the entire AI complex.