Amazon Has Become an E-commerce and Cloud Computing Behemoth. Now It's Set Its Sights on Another Billion-Dollar Market. And It Could Spell Trouble for Nvidia.

Amazon’s AWS AI-chip business has reached an annual revenue run rate of more than $20 billion, and CEO Andy Jassy said it could be about $50 billion if sold to third parties. Trainium2 is nearly sold out and Trainium3 is almost fully subscribed, underscoring strong AI demand and potential upside for Amazon. The piece is positive for Amazon’s growth story, while noting a modest competitive risk to Nvidia if AWS expands its chip offering.

Analysis

The market is underestimating how quickly AWS can turn internal AI demand into an external distribution advantage. The key second-order effect is not just chip substitution away from NVIDIA, but the creation of a lower-cost, vertically integrated procurement layer that can compress pricing across the entire AI infra stack. That should be structurally bullish for AMZN margins over the next 6-18 months if utilization stays tight, because chip scarcity supports AWS pricing power while improving customer stickiness.

For NVDA, the real risk is not immediate share loss; it is margin pressure from having to defend the performance frontier while AWS monetizes the value segment below it. If Amazon successfully monetizes Trainium as an AWS-only product first and then expands to racks for third parties, the competitive threat broadens from "alternative chip" to "cloud-native supply chain." That can force NVDA to spend more aggressively on cadence, software, and customer lock-in, which is manageable but can cap multiple expansion if growth inflects slower than expected.

The contrarian read is that this is still a positive for NVDA in the medium term because more AI deployment expands total compute spend faster than it cannibalizes premium demand. The bigger overlooked beneficiary may be AWS itself: every dollar spent on in-house silicon improves gross margin on compute relative to buying third-party accelerators, and that delta compounds with scale. The main reversal catalyst would be a rapid normalization in AI capacity supply over the next 2-3 quarters, which would weaken Amazon’s scarcity premium and reduce urgency around alternate chips.

Key Decisions for Investors

• Stay long AMZN for 6-12 months; the risk/reward improves if AWS chip scarcity persists, because in-house silicon can expand cloud margins while supporting share gains in lower-cost AI workloads.
• Hold NVDA but trim into strength on any multiple expansion above peers; the right framing is not a thesis break, but a higher execution bar over the next 4-6 quarters as AWS pressures innovation cadence.
• Pair trade: long AMZN / short a basket of AI infra beneficiaries with weaker vertical integration over the next 3-6 months; the trade works if cloud-native silicon becomes a durable procurement lever.
• Buy NVDA downside protection 3-6 months out rather than shorting outright; the near-term headline risk is competitive, but the stock still benefits if AI capex remains elevated and demand broadens.
• Set an alert for any customer-facing third-party launch of AWS chips; if announced, expect a re-rating event for both AMZN and NVDA within days, with AMZN likely outperforming on operating leverage.