Amazon Spent More on Data Centers Last Quarter Than AWS Earned in Revenue. Should Investors Worry?

Source The Motley Fool

Key Points

  • Amazon’s cash capital spending reached about $53 billion in the second quarter, topping AWS’ $42 billion in revenue.

  • Management says demand justifies the spending, but free cash flow will take a hit before the payoff arrives.

  • Motley Fool Issues Rare "Total Conviction" Buy Alert ›

Amazon (NASDAQ:AMZN) has proven itself to be one of the early winners of the artificial intelligence (AI) boom, and this is thanks to its cloud computing business. Amazon Web Services (AWS) is a hub for anything AI, offering the chips and systems needed to power AI workloads. And thanks to this, AWS' revenue has soared, reaching a $169 billion annual run rate in the second quarter.

But the report also shows us another trend that might not please investors. And that's the fact that the company spent more on data centers in the period than AWS earned in revenue. Does this spell trouble for the AI giant? Let's find out.

Missed AI’s "Act 1"? Act 2 Could Be 15x Bigger. Most investors think they missed the AI boat because they didn't buy Nvidia in 2005. But according to our analysts, we’re only at the end of "Act 1"—the R&D phase. "Act 2" is the global rollout. Continue »

Glowing AI cloud icon above a futuristic blue circuit board and digital data displays

Image source: Getty Images.

The Amazon story so far

First, a quick catch-up on Amazon's AI story so far. AWS, the company's biggest profit driver, is the world's No. 1 cloud services provider, and as such, attracts a considerable amount of AI business. Customers that already use AWS for non-AI services find it easy to broaden into AI services here rather than turn to another cloud company.

As a result, AWS' revenue has taken off, and the second quarter illustrates this. In that period, the unit generated about $42 billion in revenue, up 36% year over year, and $16 billion in operating income. At the same time, though, Amazon's cash capital expenditures totaled just over $53 billion. The company says that investment "primarily relates to AWS and generative AI."

That puts capex at about $10 billion above AWS quarterly revenue and more than three times its operating income. Though this figure refers to the whole company, most of the spending is flowing toward AI and cloud capacity. And the pace of investment isn't slowing. Amazon lifted its 2026 capex forecast by $20 billion to about $220 billion.

Why Amazon says it's spending now

Management's argument is that demand exists right now -- this isn't about possibilities down the road. AWS' backlog, or contracted future revenue, has reached $496 billion and is growing at a triple-digit rate. Even at the above capex level, Amazon still won't be able to meet 2026 demand, says chief Andy Jassy, and he expects the same in 2027. On top of this, he calls demand already in place for 2028 "striking."

Building to support such demand doesn't result in monetization overnight. The company invests in data centers about two years before servers can start bringing in revenue, but the good news is Amazon can use those facilities for more than 30 years. Servers and networking gear generate growth faster: The company buys them a few months before they're set to work, and they generally break even in less than three years.

Meanwhile, Amazon will face pressure on free cash flow until the new data centers open their doors and the servers inside them have been running for a few years, according to Jassy.

Does this signal trouble?

So, does spending that outpaces revenue signal trouble? Not necessarily. Amazon is highly profitable with a long track record of earnings growth and experience in investing in technology that will deliver over time: the creation of AWS is a clear example of this. And Amazon says it buys servers only when customer demand is there, which is different from spending based on hope.

Still, this is a bet on timing, and investors must be comfortable with that. Amazon has to build before it can earn, so investors should expect free cash flow to remain under pressure during this time. AWS' backlog gives the bet a solid foundation, and now the thing to watch is whether the unit's revenue keeps accelerating alongside capex. From what we've seen so far, though, Amazon is well positioned to deliver, making it a fantastic AI stock to buy and hold.

Should you buy stock in Amazon right now?

Before you buy stock in Amazon, consider this:

The Motley Fool Stock Advisor analyst team just identified what they believe are the 10 best stocks for investors to buy now… and Amazon wasn’t one of them. The 10 stocks that made the cut could produce monster returns in the coming years.

Consider when Netflix made this list on December 17, 2004... if you invested $1,000 at the time of our recommendation, you’d have $361,596!* Or when Nvidia made this list on April 15, 2005... if you invested $1,000 at the time of our recommendation, you’d have $1,303,935!*

Now, it’s worth noting Stock Advisor’s total average return is 910% — a market-crushing outperformance compared to 211% for the S&P 500. Don't miss the latest top 10 list, available with Stock Advisor, and join an investing community built by individual investors for individual investors.

See the 10 stocks »

*Stock Advisor returns as of October 24, 2026.

Adria Cimino has positions in Amazon. The Motley Fool has positions in and recommends Amazon. The Motley Fool has a disclosure policy.

Disclaimer: For information purposes only. Past performance is not indicative of future results.
placeholder
Dollar presses 102 as the 10-year yield sits at 5.14% — and October hike odds are only 17%The dollar index closed at 101.99 (+0.20%) on Monday after briefly topping 102.07, while the 10-year Treasury yield rose to 5.14% and the 2-year hit 4.774% — yet futures price only about a 17% chance of a hike at the October 27-28 FOMC, versus roughly 90% for December. The driver is a global bond selloff, not a Fed repricing: France's 10-year OAT jumped to 4.84% ~ 4.94% and the OAT-Bund spread widened to 140 ~ 146bp. Here are the levels, the cross-asset read-through and what to watch.
Author  Irene Q.
Oct 20, Tue
The dollar index closed at 101.99 (+0.20%) on Monday after briefly topping 102.07, while the 10-year Treasury yield rose to 5.14% and the 2-year hit 4.774% — yet futures price only about a 17% chance of a hike at the October 27-28 FOMC, versus roughly 90% for December. The driver is a global bond selloff, not a Fed repricing: France's 10-year OAT jumped to 4.84% ~ 4.94% and the OAT-Bund spread widened to 140 ~ 146bp. Here are the levels, the cross-asset read-through and what to watch.
placeholder
Meta's AI bill just hit $145 billion — and it issued a new share class to pay for itMeta reported record third-quarter revenue of $64.60 billion, up 26% from a year earlier, and raised full-year 2026 capital expenditure guidance to $135-145 billion from $130-145 billion. To fund that build-out without touching Mark Zuckerberg's voting control, the company announced a 2-for-1 stock split that hands every Class A holder a new non-voting Class C share. Here are the four lines that matter, the peer group that is being re-rated alongside it, and the bull and bear cases.
Author  Irene Q.
Oct 22, Thu
Meta reported record third-quarter revenue of $64.60 billion, up 26% from a year earlier, and raised full-year 2026 capital expenditure guidance to $135-145 billion from $130-145 billion. To fund that build-out without touching Mark Zuckerberg's voting control, the company announced a 2-for-1 stock split that hands every Class A holder a new non-voting Class C share. Here are the four lines that matter, the peer group that is being re-rated alongside it, and the bull and bear cases.
placeholder
The Fed will sit still on October 28 — but 100 of 100 economists now expect a December hikeA Reuters poll run October 16-20 found 94 of 100 economists expecting the Fed to hold at 3.75-4.00% on October 28 — and all 100 expecting a hike in December, against less than half a month earlier. The market is close to the same price, with CME FedWatch at roughly 17% for October and 82.8% for December. With the Fed in blackout since October 17, here are the five dates that decide the rest of the year, what each one is priced for, and the two directions each can take.
Author  Irene Q.
Oct 22, Thu
A Reuters poll run October 16-20 found 94 of 100 economists expecting the Fed to hold at 3.75-4.00% on October 28 — and all 100 expecting a hike in December, against less than half a month earlier. The market is close to the same price, with CME FedWatch at roughly 17% for October and 82.8% for December. With the Fed in blackout since October 17, here are the five dates that decide the rest of the year, what each one is priced for, and the two directions each can take.
placeholder
Gold Price Analysis Today | Forex Gold Analysis | XAU/USD Analysis - October 23, 2026Get the latest gold price analysis, forex gold insights, and XAU/USD analysis for October 23, 2026. Read on for the details.
Author  Mitrade
Yesterday 02: 56
Get the latest gold price analysis, forex gold insights, and XAU/USD analysis for October 23, 2026. Read on for the details.
placeholder
Nasdaq closes at 26,008.86 — 8.9 points above 26,000, and the credit market never blinkedThe Nasdaq Composite fell 2.60% to 26,008.86 on October 22, its worst session since June, as Meta sank 6.98% on 2026 capex guidance of $135–145 billion. But the equal-weight S&P 500 rose 0.73%, high-yield credit was flat and 10-year yields fell 20 basis points. This was an AI de-rating, not a risk-off event — here is the transmission chain, the 29,000 / 28,200 levels and what would prove it wrong.
Author  Irene Q.
Yesterday 06: 39
The Nasdaq Composite fell 2.60% to 26,008.86 on October 22, its worst session since June, as Meta sank 6.98% on 2026 capex guidance of $135–145 billion. But the equal-weight S&P 500 rose 0.73%, high-yield credit was flat and 10-year yields fell 20 basis points. This was an AI de-rating, not a risk-off event — here is the transmission chain, the 29,000 / 28,200 levels and what would prove it wrong.
goTop
quote