How Will Amazon and Alphabet Handle Custom Silicon Efficiency and Grid Access Now?
Amazon keeps its custom chips in-house while Google sells its chips to outside customers, and both strategies now collide with the same limit: power. One approach protects margins, and the other leans on borrowed money. Which one holds up?
This post may contain links from our sponsors and affiliates, and Flywheel Publishing may receive compensation for actions taken through them.
Amazon (NASDAQ:AMZN | AMZN Price Prediction) and Alphabet (NASDAQ:GOOGL) reported third-quarter results a day apart in late October, and both quarters ran on homegrown chips. AWS used Trainium and Graviton to post its fastest growth in nearly five years. Google Cloud more than doubled, and TPU systems became a product Google sells directly. Both companies now need much more power to keep growing.
AWS Keeps Its Chips In-House. Google Sells Its Chips to Customers.
AWS revenue reached $45.98 billion, up 39%, with a 38.7% operating margin. Management linked part of that margin to custom silicon savings. Management said customers using Amazon’s chips saved more than $2 billion in Q3 alone. Trainium revenue commitments rose more than 20x. Most of Bedrock already runs on Trainium, so Amazon keeps those savings as AWS profit.
Google Cloud revenue hit $30.71 billion, up 103%, and operating income rose to $11.12 billion from $3.59 billion. That jump coincided with strong demand for TPU systems. CEO Sundar Pichai cited “strong demand for our TPU systems across a wide range of customers.”
| Silicon Lens | Amazon | Alphabet |
|---|---|---|
| Flagship chips | Trainium, Graviton | TPU systems |
| How chips earn | Mostly inside AWS and Bedrock | Sold as Cloud product revenue |
| External chip sales | Late 2027 | Underway now |
| Q3 capex | $77.26B | $51.36B |
Amazon Controls Its Sites While Google Borrows to Build
Amazon said large customers reserve hundreds of megawatts or up to a gigawatt of capacity in advance, though it did not give a utility timeline or megawatt allocation. CEO Andy Jassy said neocloud capacity runs about 3,500 basis points worse in operating margin than long-leased data centers. Owning or long-leasing sites can give Amazon more control over its power connections. Amazon also listed energy prices as a risk and committed $1 billion to U.S. data center communities.
Google’s pressure shows up in its cash flow. Free cash flow fell to negative $9.06 billion, and Alphabet raised $28.7 billion in notes on top of equity sold through its at-the-market (ATM) program. Google has also used demand-response arrangements that let utilities reduce data center load during peak periods. That flexibility helps, but debt and share dilution are paying for the power-hungry buildout. All of that capacity has to be powered and cooled by someone, and we pulled seven of those suppliers into a free report here.
Trainium’s Debut and Google’s Cash Burn Come Next
The focus will be on whether Amazon’s roughly $220 billion in 2026 capex keeps reducing payback periods as management promised. You should keep an eye on whether TPU sales can cover Google’s rising interest costs. Alphabet’s GAAP EPS of $2.12 missed the $2.97 estimate. An equity securities loss accounted for $0.55 of that shortfall.
Why I Lean Toward Amazon’s Chip Strategy Right Now
Amazon rose 7.88% over the past week, and Alphabet fell 18.75% over the past month, though those windows are not comparable. Amazon’s model holds a structural edge because its Trainium savings stay inside AWS margins, and its owned and leased sites give it more control over power. Google is growing faster, but it is funding that growth with debt and dilution. That edge would narrow if Google’s free cash flow turns positive while Cloud keeps doubling.
Contact [email protected] for any questions or corrections.
Alex Sirois is a financial writer with experience spanning both retail and institutional investing. He has written for InvestorPlace and held roles at BNY Mellon and Bernstein, giving him a perspective that bridges Main Street portfolios and Wall Street analysis.
Alex holds an MBA from George Washington University and has built his career across multiple industries, including e-commerce, education, and translation — a breadth of experience that informs how he breaks down complex financial topics for everyday investors. His writing is conversational, actionable, and grounded in long-term, buy-and-hold investing principles.
At 247 Wall St., Alex focuses on delivering analysis that is both accessible and useful, with a clear emphasis on helping readers make more informed decisions with their money.
