The Reason I Keep Buying Amazon That Wall Street Keeps Missing

I keep clicking buy on Amazon (NASDAQ:AMZN | AMZN Price Prediction), and I am not embarrassed to admit it. Every time the stock dips into the low $240s, I add. My cost basis keeps rising, my share count keeps rising, and my conviction gets louder with every filing.

Here is what pulls me back. Amazon is running the largest infrastructure buildout of my investing lifetime, and it already has the receipts to pay for it. The market keeps arguing about whether hyperscaler capex is discipline or waste. I look at the pre-committed contracts sitting behind the spend and see something Wall Street keeps glossing over.

The Backlog Nobody Wants to Talk About

AWS carried a $364 billion commercial backlog into Q1 2026, and that figure does not include the recent Anthropic deal announced for over $100 billion. Andy Jassy told analysts “there is reasonable breadth in that as well, it is not just one customer or two customers.” Against that pipeline, AWS grew 28% year over year to $37.59 billion, the fastest pace in 15 quarters, at a 37.7% operating margin.

Then the silicon. Amazon’s chips business is running at a $20 billion annual revenue run rate, growing triple digits year over year. Trainium alone carries over $225 billion in revenue commitments. Jassy said “at scale, we expect Trainium will save us tens of billions of dollars of CapEx each year and provide several hundred basis points of operating margin advantage”. That is the math I keep chewing on. Amazon owns the chip stack for the AI workloads it can move in-house.

Third, the boring engine keeps compounding. Advertising crossed $70 billion in TTM revenue while growing 24%, and unit growth in Stores hit 15%, the highest since the tail end of covid lockdowns. Consolidated operating margin came in at 13.1%, the highest ever.

Why Not Microsoft or Alphabet

The question I get asked most: why not Microsoft (NASDAQ:MSFT) or Alphabet (NASDAQ:GOOGL)? I own the neighborhood. Amazon is where I keep adding. Microsoft leans on OpenAI’s roadmap for its differentiated AI story, and Google Cloud remains the third player in enterprise share. Amazon just booked OpenAI itself for approximately 2 GW of Trainium capacity beginning in 2027 and Anthropic for up to 5 GW. The two leading AI labs are now paying to run on Amazon silicon. At a trailing P/E of 29 against 76.65% year-over-year net income growth, I am paying below-market multiples for the fastest hyperscaler in the group.

The Real Risk

The real risk is cash. Free cash flow TTM collapsed 95% to $1.2 billion as capex hit $44.2 billion in a single quarter, and long-term debt climbed to $119.1 billion from $65.6 billion year over year. The spend would scare me if it were speculative. Instead, Jassy said Amazon has “customer commitments for a substantial portion” of that 2026 spend, and the monetization window on new capacity runs 6 to 24 months against assets with 30+ year useful lives on data centers. The math works if you let it play out.

I keep buying because the backlog is signed, the silicon is shipping, the consolidated margin line just hit an all-time high, and the stock trades at $247.49 against an analyst consensus target of $312.91 with 62 buy ratings and zero sells. Until the receipts stop arriving, my finger stays on the buy button.

Contact [email protected] for any questions or corrections.