On a recent episode of The Investor’s Podcast (838), Daniel Mahncke and Shawn O’Malley argued that Amazon offers more asymmetric upside than its hyperscaler peers: “Amazon has literally built one of the largest chip businesses in the world in the last couple of years, and barely anyone has even noticed.”
Amazon’s silicon business now exceeds a $25 billion annualized run rate, is growing at a triple-digit percentage, and has expanded from about a $10 billion run rate in under a year.
The $25B Chip Business Hiding Inside AWS
Amazon (NASDAQ:AMZN | AMZN Price Prediction) has scaled its custom silicon operation faster than most investors realize. On the Q2 FY2026 earnings call, CEO Andy Jassy told analysts AWS grew 36.7% year over year, the fastest growth in 18 quarters. On the show, the hosts suggested the current $25 billion run rate for the chips business “could even be double that or closer to $50 billion” if Amazon began selling chips externally.
Amazon’s Trainium2 is fully subscribed with 1.4 million chips landed, powering the majority of inference on Bedrock. Project Rainier is the world’s largest operational AI compute cluster with more than 500,000 Trainium2 chips training Anthropic’s Claude, and OpenAI committed to roughly 2 GW of Trainium capacity beginning 2027. On the CPU side, Graviton is used by 98% of the top 1,000 EC2 customers.
AWS’s $496B Backlog Supports the Silicon Expansion
AWS revenue reached $42.2 billion in Q2 FY2026, with operating income of $16.6 billion and a 39% operating margin. Even more exciting, the company’s $496 billion backlog tells the forward story, growing at triple-digit rates year over year.
Amazon’s capital spending is climbing to match. Q2 CapEx was $53.1 billion, with roughly $200 billion planned for FY2026. Amazon’s CEO Andy Jassy has framed AWS as capable of becoming “a trillion-dollar annual revenue business for us in time.”
Google’s AI Rally Has Left Amazon Far Behind
Alphabet (NASDAQ:GOOGL) offers a natural comparison, with its own custom TPU stack and hyperscale cloud. Google Cloud accelerated to 82% growth in Q2 FY2026, reaching $24.77 billion, with Q2 CapEx of $44.9 billion. The show hosts flagged that both companies are guiding to roughly $200 billion in CapEx and that free cash flow has turned negative as a result.
What differs is the market’s reception. At the time of recording, Alphabet was up roughly 75% while Amazon was up just 0.5% over 12 months. The hosts laid out the case for Amazon stock today, saying: “Since Amazon has benefited less from the AI hype cycle, there’s probably less for the stock to lose in terms of giving up gains,” and “It does feel like [Amazon] has a lot more room to catch up.”
Amazon’s Valuation Creates an Asymmetric AI Setup
After adjusting for a $17 billion markup due to Amazon’s Anthropic stake, Amazon trades closer to 17-18x operating cash flow, which the hosts called “pretty attractive for a company of that quality.” The forward P/E sits at 28.3, with analyst consensus firmly bullish at 59 buy or strong buy ratings versus 3 hold and no sell ratings, and a target price of $327 vs a current share price of $260.
Amazon’s custom-silicon business, Trainium and Graviton, has become a core part of AWS. This business supports Anthropic, has attracted a major OpenAI commitment, and generates more than $25 billion in annualized revenue. The risk is that Amazon must spend roughly $200 billion this year to satisfy an AI demand curve that remains difficult to forecast.
Yet with Amazon trailing dramatically behind Alphabet, investors may be getting one of the world’s fastest-growing chip franchises without paying the same AI premium attached to its peers.
Contact [email protected] for any questions or corrections.
Top Gaining Stocks
Top Losing Stocks