Amazon's $132 Billion Capex — What Ugly Financials Are Hiding

Amazon's financials look broken on the surface. Net income of $77 billion last year, free cash flow of $7.7 billion. That's not a typo. The company's free cash flow is one-tenth of its net income because it spent $132 billion on capital expenditures in a single year.

Most investors see that and walk away. That reaction might be the opportunity.

The Capex Trajectory Tells the Real Story

Looking at Amazon's capital expenditure over the past decade reveals a company that invests for the future with almost reckless aggression:

$7.8B → $8B → $12B → $13.5B → $12B → $17B → $40B → $61B → $64B → $52B → $83B → $132B

That's a 16x increase over ten years. Outside of Berkshire Hathaway, it's hard to find a company this committed to reinvesting in its future.

Admittedly, a decade ago, this pattern looked like financial recklessness. No cash flow, margins that seemed permanently thin. The conventional wisdom was that Amazon couldn't convert revenue into profits. That turned out to be completely wrong. They were building AWS, the logistics network, and the advertising infrastructure that now drive massive earnings.

Why Traditional Metrics Fail Here

Amazon's P/E of 29 looks reasonable. But the P/FCF ratio is essentially meaningless because the company reinvests almost everything it earns.

The five-year averages make the gap clear:

• 5-year average net income: $39.6 billion
• 5-year average free cash flow: $8.23 billion
• The difference is almost entirely capex

What's actually improving underneath all that spending:

• Operating margin: from 6% a decade ago to 10.5% last year
• Gross margin: expanded to ~50%
• ROIC: trending up — 7.5% five-year average, 9% last year

The core business is becoming dramatically more profitable. Amazon just keeps pouring profits back into the next phase of growth.

The AI Capex Question

The market has been pressuring Amazon over AI capital expenditure concerns in 2026. The worry is legitimate — every big tech company is spending aggressively on AI, and much of it is "keep up or fall behind" spending rather than proven-return investing.

But Amazon's position is fundamentally different. AWS is the infrastructure AI runs on. Other companies building AI services need AWS to do it. On top of that, retail margins are expanding and the advertising business is quietly becoming one of the most profitable segments.

Analyst estimates project:

• EPS more than doubling over the next four years
• Double-digit revenue growth sustained through 2029

These are enormous expectations. But Amazon has a track record of making bets that look expensive, then delivering returns that dwarf the investment.

Running the Numbers

Ten-year analysis assumptions:

• Revenue growth: 4%, 8%, 12%
• Operating margin: 8%, 12%, 16% (above historical highs)
• FCF margin: same range (expecting eventual convergence)
• Terminal P/E: 20x, 23x, 26x (premium business)
• Required return: 9%

Result: low $103, mid $234, high $469. Current price: $209.

At the mid-case, that's roughly a 10.5% expected return. The question every investor needs to answer: is 10.5% enough compensation for the uncertainty around whether $132 billion in annual capex will actually pay off?

My assumptions may be conservative. Pushing operating margins higher shifts the fair value up meaningfully. But conservative assumptions are where I prefer to be wrong.

The Core Question

Every analysis of Amazon comes down to one thing: will the capex convert to free cash flow?

History says yes. AWS proved it. The logistics network proved it. Prime proved it. But AI is spending at a different scale, and not every dollar invested will generate returns.

Amazon is better positioned than most to win this game. Whether the current price adequately reflects both the opportunity and the risk is the judgment call each investor has to make with their own numbers — not someone else's recommendation.

FAQ

Q: Why is Amazon's free cash flow so much lower than net income? A: Capital expenditures. Amazon spent $132 billion on capex last year — building data centers, logistics infrastructure, and AI capacity. Free cash flow = operating cash flow minus capex. When capex is this massive, FCF shrinks dramatically even while the business is highly profitable on an earnings basis.

Q: Should Amazon be paying a dividend instead of spending this much? A: No. Amazon's historical pattern shows that reinvested capital generates far higher returns than what shareholders could earn from dividends. AWS alone generates more value than decades of potential dividend payments. As long as the company can deploy capital at above-market returns, dividends would actually destroy shareholder value.

Q: At what price would Amazon be a clear buy? A: Based on my mid-case assumptions (8% revenue growth, 12% margin, 23x terminal P/E), the fair value sits around $234. Anything below $200 would offer a meaningful margin of safety. But this heavily depends on your view of whether AI capex will convert to cash flow at historical rates.