Jensen Huang's Nvidia Guided for 70% Revenue Growth in Fiscal 2028, Far Above the 44% Wall Street Expected, as Amazon Agreed to Buy 2 Million Nvidia GPUs. Is the Growth Forecast Believable?

Nvidia guided to 70% year-over-year revenue growth for fiscal 2028 (CFO Colette Kress, reflecting supply constraints); absent supply bottlenecks, management suggested revenue could more than double. The ramp of its Vera Rubin (CPU+GPU) platforms is starting to ship, and Amazon’s planned order of 2 million Nvidia GPUs across 2027–2028 reinforces the two-year revenue outlook. With prior growth accelerating (e.g., 106% YoY in fiscal 2027 Q2) and guidance implying ~12% sequential growth at the midpoint in fiscal 2027 Q3, the article frames Nvidia’s ramp as support for meeting (or exceeding) the 2028 target.

Analysis

The real market signal is not that demand exists; it is that the constraint has shifted to industrial capacity. That usually means the highest-multiple asset in the chain can keep compounding, but the incremental economics migrate upstream to whoever controls scarce packaging, memory, and cluster build-out capacity. In that setup, NVDA can keep winning on revenue, but the more asymmetric rerating may be in suppliers with underappreciated supply optionality rather than in a name the market already treats as a monopoly.

AMZN’s GPU commitment matters less as a direct earnings catalyst and more as a proof point that hyperscalers are still in an arms race. Near term, that is bullish for AWS stickiness and AI monetization credibility; over 1-3 quarters it also means heavier capex, higher depreciation, and more scrutiny on free-cash-flow conversion. If monetization does not accelerate by the time those assets hit the P&L, the market could start penalizing AI spend instead of rewarding it.

The contrarian miss is that “more demand” is not the same as “more profit” for NVDA shareholders at this stage. If supply remains tight, the upside over 6-18 months may accrue disproportionately to capacity holders and memory vendors, while NVDA’s stock becomes more sensitive to any hint of digestion, lead-time normalization, or hyperscaler capex discipline. The key falsifier is a step-down in sequential growth or commentary that supply bottlenecks have eased faster than expected; that would compress the 2028 narrative quickly.

Key Decisions for Investors

• Prefer a relative-value long SSNLF / short NVDA basket over the next 6-12 months if liquidity is acceptable; thesis is that bottleneck rents flow to memory capacity before they fully re-rate into NVDA. Risk: if NVDA sustains 100%+ growth without margin erosion, the pair underperforms.
• Do not chase NVDA after the news; wait for a 3-5% pullback or a failed post-earnings extension to add. Risk/reward is better on a retrace because the market has already started discounting the 2028 setup.
• Maintain a constructive view on AMZN, but size it as a capex-cycle beneficiary rather than a clean near-term FCF story. Consider financing via covered calls into strength over the next 1-2 quarters; the trade works if AWS AI monetization catches up before depreciation pressure becomes visible.
• Set an alert on hyperscaler capex and memory pricing into the next earnings cycle: if HBM/DRAM prices roll over or guidance implies slower GPU deployment, cut the NVDA-through-2028 thesis quickly. If lead times stay extended, add to suppliers rather than the headline AI leader.