Is Amazon's Capex Bet Worth Stomaching the Free Cash Flow Drought?

April 27, 2026·3 min read

Direct answer: The roughly $70B gap between Amazon's free cash flow and net income isn't accounting trickery. It's the largest AI infrastructure bet in corporate history. The real question is whether today's price embeds enough margin of safety to wait for that bet to pay off.

The first time you look at Amazon's numbers, you stop. Market cap of $2.7T, enterprise value of $2.98T, debt around $300B. And free cash flow of just $7.7B. Net income, meanwhile, is $77.7B — roughly 10x the free cash flow. That gap is the most extreme I see in any large business I track.

If you don't understand where the gap comes from, you can't analyze Amazon. The answer is one line in the cash flow statement: capital expenditures.

The Capex Curve, in Plain Numbers

Capex over the last several years:

2019: ~$16B
2025: $130B+
2026 plan: ~$200B

That's a 7–8x increase. This isn't an accounting drift — it's a deliberate choice. CEO Andy Jassy is funneling cash directly into AWS infrastructure and AI training data centers. The optical similarity to Enron is superficial: Enron faked the profit. Amazon is investing real capital in real future revenue streams. AWS itself looked like a money pit for years before becoming one of the most profitable businesses ever built.

So What Is It Worth?

Two assumptions matter when modeling Amazon.

Will margins and free cash flow converge over time? If you don't believe they will, don't waste time analyzing Amazon.
Is the return on invested capital high enough? If yes, prolonged reinvestment compounds value rather than destroying it.

My 10-year DCF setup:

Revenue growth: 4 / 8 / 12%
Operating and free cash flow margin: 8 / 12 / 16%
Terminal P/E in year 10: 20 / 23 / 26
Required return: 9%

Scenario	Revenue Growth	Margin	Terminal P/E	Fair Value
Bear	4%	8%	20	$103
Base	8%	12%	23	$230
Bull	12%	16%	26	$470

The base case at $230 is just below today's $247. Worth noting: the 12% margin assumption is a level Amazon has never actually hit. That's why I keep my own watchlist at $200, not at the model's mid output. If Amazon comes to me, I re-evaluate. If it doesn't, I can sell cash-secured puts at lower strikes and get paid to wait.

Would You Pick Amazon Over a Car for a Year?

Warren Buffett's framing of iPhone intrinsic value asked: "If you had to give up your car or your phone for a year, which goes?" Try the same thought experiment with Amazon. If you met someone today who had never used Amazon, your reaction would be, "Where have you been living?" Advertising — built on top of that habitual usage — is now a multi-billion-dollar business growing fast.

A business that integrated into daily life doesn't trade at a 15–16x P/E. The premium is justified. The question is just how much premium you're willing to pay, which is exactly what defines your buy price.

More in this Category

Four Mega-Caps in 48 Hours: The Hinge Week for 2026

Over 20% of the S&P 500 reports next week, and Microsoft, Meta, Amazon, and Apple are clustered into a single 48-hour window. Here's why this is the hinge week for 2026 — and what to listen for in each print.

Beats Aren't Enough — The Week Guidance Beat EPS

Texas Instruments rallied 10% this week. IBM and ServiceNow dropped double digits. All three met or beat consensus. Here's why the outcomes split — and the framework I'm taking into next week's mega-cap prints.

SMH Just Printed 17 Up Days in a Row — Chase or Wait for the Pullback?

SMH has now closed up 17 sessions in a row — possibly its longest streak ever. Intel +22%, AMD +12.6%, Nvidia +2.75% in a single day. My read: don't fight it, but don't chase it here. Wait for the pullback.

GE Aerospace Beat Expectations and Still Dropped 5.56% — Here's What Wall Street Saw

Orders up 87%, revenue up 29%, a record $210B backlog — yet shares fell 5.56% on print day. Held guidance, a cut departures outlook, and 230bp margin compression explain the gap.

Why GE Aerospace Is the Toll Booth of Global Aviation: A 30-Year Razor-and-Blade Anatomy

GE powers about 75% of narrow-body and 55% of wide-body flight cycles globally and controls roughly 40% of the engine MRO market. The engine is the razor; three decades of service work is the blade.

GE Aerospace Bull vs Bear: A Framework for Deciding Whether 35x Earnings Is Worth It

Bull case: 95% of next-quarter spare parts revenue locked in, two-thirds of the CFM56 fleet still ahead of its second shop visit. Bear case: 70% spare parts delinquency, double-digit Middle East departure declines, and a 39x forward multiple.

Big Four AI CapEx in One Year: $300B–$500B and a Build-Out Bigger Than Most Countries' GDP

Amazon, Meta, Microsoft, and Alphabet are deploying $300B–$500B on AI infrastructure in a single year. Amazon alone has committed ~$200B, Microsoft $80B in one fiscal year — bigger than most countries' GDP.

Is AI CapEx the Next AWS or the Next Fiber Optic Bubble — Both Analogies Matter

The AI CapEx debate is really a fight between two historical analogies — the 1990s fiber optic glut (bears) versus the early-2010s AWS build-out (bulls). Both are partly right; what decides outcomes is the price you pay.

Microsoft at $417 and $3.12T Cap — A Conservative Valuation of the AI Bet

Microsoft trades at $417 with a $3.12T market cap and $77.5B FCF. My conservative valuation (7/9/11% growth, 34/37/40% margin, 20/23/26 terminal multiples) gives a $484 midpoint — about 11% annualized at today's price.