Amazon's Free Cash Flow Is Nearly Zero. Here's Why That's Actually Bullish.

Amazon's EV/FCF ratio is approximately 301x right now.

That's one of the most alarming valuation metrics in the entire S&P 500. At 301x free cash flow, Amazon looks wildly overvalued — more expensive than Tesla on some traditional metrics.

But Amazon generated $139.5 billion in operating cash flow in 2025, up 20% year-over-year. That's one of the highest operating cash flow figures of any company on Earth.

How can a company with $139.5B in operating cash flow have near-zero free cash flow?

The answer is a single number: $131.8 billion in capital expenditure in 2025, with guidance to spend ~$200 billion in 2026.

That gap — between $139.5B coming in and $131.8B going straight back out — is the whole Amazon investment thesis. Understanding it is the difference between thinking Amazon is broken and recognizing it as one of the most deliberate long-term capital allocation stories in market history.

Not financial advice. All figures from Amazon's FY2025 earnings release and company filings as of March 2026.

The Number That Explains Everything

Here's how Amazon's cash flows have evolved:

| Year | Operating Cash Flow | Capex | Reported FCF | |------|--------------------|---------|----| | FY2022 | ~$47B | ~$58B | -$11B | | FY2023 | ~$85B | ~$52B | ~$33B | | FY2024 | ~$116B | ~$78B | ~$38B | | FY2025 | $139.5B | ~$131.8B | ~$7–11B | | FY2026E | ~$155–165B (est.) | ~$200B | ~$0 to negative |

The operating cash flow trajectory is exceptional — from $47B in 2022 to $139.5B in 2025, a 3x increase in three years. That's a business generating cash at a remarkable rate.

The capex trajectory is the other side of the story: from $52B in 2023 to $131.8B in 2025, now guided to ~$200B in 2026. Amazon is deploying nearly every dollar of operating cash flow back into infrastructure.

The 301x EV/FCF isn't a sign of a broken business. It's a sign of an intentional capital allocation decision by a company that believes its return on invested capital justifies near-total reinvestment.

What Amazon Actually Is: Three Businesses, One Ticker

Most of the confusion about Amazon's valuation comes from treating it as one company with blended economics. It isn't. Here's what the FY2025 segment data shows:

| Segment | Revenue | Operating Income | Operating Margin | |---------|---------|-----------------|-----------------| | North America (retail + ads) | $426.3B | $29.6B | 6.9% | | International (retail) | $161.9B | $4.7B | 2.9% | | AWS | $128.7B | $45.6B | 35.4% | | Total Amazon | $716.9B | $80.0B | 11.2% |

AWS is 18% of Amazon's revenue and 57% of its operating income.

That's the fact that restructures every other number in this analysis. You aren't buying a thin-margin e-commerce company at 30x earnings. You're buying a 35%-margin cloud computing business that happens to also run one of the world's largest retailers — and using the retail revenues to obscure what the cloud business actually earns.

The Hidden Engine: Amazon Advertising

There's a third business inside Amazon that almost nobody discusses properly.

Amazon Advertising Services — FY2025:

Revenue: $68.6 billion (+22% YoY from $56.2B in FY2024)
Growth rate: 22% for the third consecutive year — accelerating
Estimated gross margin: 60–70%+ (software-level economics)
How it's reported: buried inside the North America and International retail segments

At $68.6B, Amazon's advertising business is larger than YouTube's total revenue ($60B+ in FY2025). It's growing at the same rate as Meta's advertising business. And it has a data advantage that neither Google nor Meta can fully replicate: Amazon knows not just what users search for, but what they actually buy.

Purchase-intent advertising — someone searching "best noise-cancelling headphones" who then buys Sony WH-1000XM6 through Amazon — is the most valuable ad unit in digital marketing. Advertisers pay enormous CPMs to reach buyers at that moment. Amazon has 200+ million Prime members generating purchase data at scale.

If Amazon Advertising were a standalone public company, it would likely be worth more than $1 trillion.

The Normalization Math: What Amazon Is Really Worth

Here's the exercise that changes the valuation picture entirely.

What if capex stayed at 2023 levels (~$52B)?

Normalized FCF = Operating Cash Flow − Normalized Capex
               = $139.5B − $52B
               = ~$87.5B

Normalized EV/FCF = ~$2.2T / $87.5B
                  = ~25x

At normalized capex, Amazon generates approximately $87–90 billion in annual free cash flow — putting it among the top 3–4 FCF generators on Earth alongside Apple ($108B) and Microsoft (~$70B).

The normalized EV/FCF of ~25x is the honest multiple. You're paying roughly 25x steady-state free cash flow for a business that's temporarily reinvesting nearly everything into infrastructure it believes will generate far higher future returns.

That's not 301x. That's a reasonable price for one of the most profitable underlying businesses in the world.

The bear's response: The $52B normalization assumes capex eventually does plateau and those data centers do generate the expected returns. If AI demand proves lower than projected — if more efficient models require less compute, or if Google Cloud and Azure take meaningful AWS share — the capex creates stranded assets, not compounding returns. Management's stated verdict window: 2027–2028.

The $200B Bet: What's Amazon Actually Building?

Amazon's 2026 capex guidance of ~$200B breaks down approximately as follows:

AWS infrastructure (~65–70%): Data centers, servers, networking — demand-driven by signed AWS contracts. Management explicitly described this as capacity-constrained, not speculative. AWS is currently turning away some customers due to supply limits.
North America logistics (~25%): Same-day delivery expansion, fulfillment robotics, automation that structurally reduces per-unit cost
International + other (~5–10%): Regional expansion, Project Kuiper satellite internet

The key phrase from Q4 2025 earnings: management described the capex as having "high returns on invested capital as demand accelerates." This isn't a vision statement — it's a specific ROI claim about AWS contracts that are signed and waiting for capacity.

The historical analogy: From 2010 to 2015, Amazon's capex regularly exceeded FCF as it built the fulfillment network that now gives it 1–2 day delivery to most of the US. Investors who couldn't see through the FCF distortion sold. The fulfillment network is now an unassailable moat. Amazon believes it's building the equivalent moat in AI cloud infrastructure.

Sum-of-Parts: What Each Piece Is Worth

| Segment | Operating Income | Multiple | Segment Value | |---------|-----------------|---------|--------------| | AWS ($128.7B rev, 35% margin, 20% growth) | $45.6B | 30x OI | $1.37T | | Advertising ($68.6B rev, ~65% margin, 22% growth) | ~$44B | 25x OI | $1.10T | | Retail — North America ($426B rev, 6.9% margin) | $29.6B | 15x OI | $444B | | Retail — International ($162B rev, 2.9% margin) | $4.7B | 15x OI | $70B | | Subscriptions ($49.6B rev) | ~$15B | 15x OI | $225B | | Net debt/other | — | — | ~-$80B | | Implied total equity value | | | ~$3.13T |

At a $2.2T market cap, Amazon is trading at approximately a 30% discount to sum-of-parts.

The discount exists because blended reporting obscures the segment economics. Investors applying a single multiple to $716.9B of mixed-quality revenue will always underprice Amazon relative to what its parts are actually worth.

DCF: The Three Scenarios

Bear case — AWS decelerates, capex doesn't pay off ($175–195):

AWS slows to 15% growth; retail margins stay flat
Capex ROI disappoints; FCF stays near zero through 2028
22x forward P/E on FY2026E EPS of ~$8.00 = $176

Base case — AWS sustains 18–20%, operating leverage improves ($255–295):

AWS reaches $160B+ by FY2027 with improving margins
Advertising reaches $83–85B at 22% growth
North America retail margin expands to 8–9% on automation
30x forward P/E on FY2026E EPS of $9.50 = $285
Analyst consensus: $280–287 (44 analysts, Strong Buy)

Bull case — AWS re-accelerates, advertising compounds ($330–380):

AWS grows 25%+ as AI workload mix drives margin expansion
Sum-of-parts re-rating toward $3T+
35x on FY2026E EPS $10.00 = $350

At ~$209, the stock trades ~$70–80 below the base case. That gap is the capex uncertainty discount. The question is whether you believe the $200B investment will generate the returns management is projecting.

The Risks Worth Taking Seriously

1. Capex ROI window: 2027–2028 is the verdict Management has a stated ROI window. If AWS growth is still 15–18% in 2028 with $200B+ in annual capex deployed, the investment case weakens materially. Watch Q1–Q3 2026 AWS growth closely — any deceleration below 20% while capex stays at $150B+ is the early warning signal.

2. AWS competitive pressure Azure grew 38–40% YoY in the most recent quarter. Google Cloud grew 48%. Both are growing faster than AWS's 20%. AWS still leads in absolute cloud revenue and has the deepest enterprise relationships — but its share of incremental cloud growth is declining. The long-term structural question is whether AI consolidates toward the most capable models (likely benefiting Google/Microsoft DeepMind/OpenAI ecosystem) or toward the most accessible infrastructure (AWS's traditional strength).

3. Advertising regulation The FTC has scrutinized Amazon's practice of prioritizing Amazon-brand products over advertisers' competing products in search results. A settlement requiring structural changes to ad ranking would impair the 22% advertising growth trajectory.

Run Your Own Amazon Valuation

The scenarios above use our assumptions on AWS growth and operating leverage. Your view on capex ROI timing, advertising durability, and retail margin expansion will differ.

Run a free AMZN valuation on EvidInvest — stress-test the capex thesis with your own inputs →

Frequently Asked Questions

Why is Amazon's free cash flow near zero if it's so profitable?

Amazon generated $139.5B in operating cash flow in FY2025 — but also spent $131.8B in capital expenditure. The two nearly cancel out, producing near-zero reported FCF. This is entirely intentional: Amazon is reinvesting nearly every dollar of operating cash into AWS infrastructure and logistics automation it believes will generate high returns. The normalized FCF (at 2023 capex levels of ~$52B) would be approximately $87.5B.

What percentage of Amazon's profit comes from AWS?

AWS generated $45.6B in operating income in FY2025 — approximately 57% of Amazon's total $80B operating income, on only 18% of revenue. AWS operates at a 35.4% operating margin; the rest of Amazon blends to about 5–7%.

Is Amazon's advertising business really bigger than YouTube?

Yes. Amazon Advertising generated $68.6B in revenue in FY2025, growing 22% YoY. YouTube's total revenue (ads plus subscriptions) exceeded $60B in the same period. Amazon Advertising is now one of the top 3 digital advertising platforms globally, behind only Google Search and Meta.

What is Amazon's intrinsic value in 2026?

Base-case DCF and analyst consensus converge at approximately $255–295. At ~$209, the stock trades at a ~25–35% discount to base case — primarily reflecting the capex uncertainty discount. Bear case (capex ROI disappoints): $175–195. Bull case (AWS re-accelerates): $330–380.

Is now a good time to buy Amazon stock?

At $209 vs. a 44-analyst consensus target of $280–287, the data suggests meaningful upside if you accept the 2027–2028 capex ROI verdict window. The investment requires patience and a view that the $200B in annual infrastructure spend will generate the returns management projects. Investors who couldn't see through the FCF distortion in 2012–2015 (when Amazon built its fulfillment network at the expense of reported FCF) missed one of the greatest investment returns of the decade.

Bottom Line

Amazon's near-zero FCF is not a warning sign — it's the fingerprint of a company spending aggressively on infrastructure it believes will generate high returns. The underlying operating cash flow ($139.5B, +20% YoY) tells the real story: this is a business generating cash at an extraordinary rate and choosing to reinvest it rather than return it.

The normalized picture — $87.5B in FCF at 25x = reasonable — looks nothing like the scary 301x headline.

The investment thesis in one sentence: Buy the capex overhang at a discount, own the AWS + Advertising compound machine. The verdict comes in 2027–2028. At $209 with 36% upside to analyst consensus, the market is pricing in a fair amount of doubt that the thesis is right.

Try EvidInvest free — run AMZN and see the sum-of-parts breakdown →