Amazon just closed out 2025 by spending $131.8 billion on capital expenditures. In a single year, the company spent more on data centers and infrastructure than the entire annual GDP of Hungary! Free cash flow collapsed as a result. From $32.9 billion in 2024 to just $7.7 billion in 2025. If you saw that headline and thought Amazon was on a downtrend, you’re not alone. It is exactly the story making the rounds on “expert” sites right now.

Here is what those takes are missing. Amazon is building, not breaking. The company is in the middle of the largest AI infrastructure buildout in history, and the cash going out the door today is what produces the cash that comes back tomorrow. The question is whether the business behind the spending justifies it.

There is a second story building around Amazon right now, and it is just as important. In April 2026, Amazon announced it is acquiring Globalstar for $11.6 billion. Globalstar is a satellite operator that owns something Amazon does not: a global wireless spectrum license. “Spectrum” sounds technical, but the idea is simple. Spectrum is the slice of the radio airwaves that wireless signals travel on. Owning a spectrum license is like owning a private highway in the sky. No one else can drive on your lane without paying you.

This deal does two things at once. First, it instantly gives Amazon Leo (Amazon's low-earth-orbit satellite network) the legal right to beam signals directly to phones around the world. Second, it hands Amazon the existing customer relationship that Apple's Emergency SOS feature runs on top of. Apple was already paying Globalstar for satellite connectivity. Now those payments come to Amazon.

The market Amazon is buying its way into is called direct-to-device satellite, or D2D. Today, if you are out of cell range, your phone is a brick. D2D changes that. With D2D, your phone connects straight to a satellite, without needing a cell tower. Emergency calls, basic texting, eventually full data — all from anywhere on Earth.

This is the next big wireless market, and there is one major player Amazon is now squaring off against: SpaceX. Elon Musk's Starlink already has thousands of satellites in low-earth orbit and is rolling out its own D2D service through partnerships with T-Mobile and others. Amazon has been behind Starlink in the satellite race for years. The Globalstar deal closes the gap in two ways: Amazon gets ~24 working satellites and the rights to launch 50+ more, and it gets the spectrum licenses needed to operate globally. SpaceX and Amazon are now the two serious players in a market that did not exist five years ago.

For shareholders, this is another bet on top of the AI capex bet. Amazon is spending heavily today to own the highest-margin parts of tomorrow (cloud, advertising, and now satellite connectivity).

That is what we are going to figure out together. By the end, you will know what Amazon actually does, how much real cash it generates when you strip out the AI cycle, whether the balance sheet can handle the spending, what management is doing with every dollar, and whether the current price gives you a reasonable entry. Spoiler: this stock is going on The Investing Department watchlist.

Amazon is not just “an online retailer.” That description was true in 2005. It has not been true for a decade. Today, Amazon is a holding company that runs seven distinct businesses. Each one successful, some are dominant.

Online Stores ($269B, 38% of revenue) is the part most people think of. You order something, Amazon ships it. Profitability is low because retail is hard.

Third-Party Seller Services ($172B, 24%) is what happens when a small business sells through Amazon’s platform. Amazon takes a cut, handles logistics, and the seller carries the inventory risk. High-margin revenue. Amazon owns the customer and the data, and lets someone else hold the inventory.

Amazon Web Services - AWS ($129B, 18%) is the cloud computing division. AWS rents out servers, storage, and AI compute to other companies. It is the most profitable part of Amazon by a wide margin and the engine behind the AI capex story.

Advertising Services ($69B, 10%) is the fastest-growing segment at 22% growth. Brands pay Amazon to show up at the top of search results, the same way they pay Google. This business has very high profitability.

Subscription Services ($50B, 7%) is mostly Prime, plus video and music.

Physical Stores ($23B, 3%) is Whole Foods and Amazon’s brick-and-mortar footprint.

Other Services ($6B, 1%) is everything else.

Add all that up and you get $717 billion in 2025 revenue, up 12.4% from 2024. That growth rate matters. When a company this large is still compounding at 12% a year, there buying windows are much wider than most companies.

On market position, Amazon is dominant but not monopolistic. In U.S. e-commerce, roughly 38% of market share, which is #1 by a wide margin. In cloud services, AWS holds about 30% of global infrastructure, second behind Microsoft Azure. In digital advertising, the third-largest player behind Google and Meta. In none of these markets is Amazon the upstart. They are the leader or the strong contender everywhere they compete.

Total revenue in 2025 came in at $716.9 billion, up 12.4% year over year. That is on top of 11.0% growth in 2024 and 11.8% in 2023. Three straight years of double-digit growth at half-trillion-dollar scale.

The mix is shifting in a direction that should make you happy. The high-margin businesses (AWS, Advertising, Subscriptions) grew faster than the low-margin retail businesses. AWS grew 19.7%. Advertising grew 22.1%. Online Stores grew only 9.0%. Every year that mix tilts further toward the higher-margin segments, the whole company’s profit profile improves.

You can see it in the margin trend.

Gross margin crossed 50% in 2025 for the first time. Operating margin more than doubled in two years. Net margin went from negative to nearly 11%. This is the mix shift showing up in the financials.

Now let’s talk about the elephant in the room. Free cash flow.

Look at 2025 alone and you might be alarmed. Look at the full picture and you see something else. Operating cash flow grew 20% in 2025 to a record $139.5 billion. The business is getting healthier, not weaker. CapEx jumped 59% to $131.8 billion — management doubling down on AI infrastructure for AWS.

Here is the normalization reasoning. FY2025 free cash flow of $7.7B is artificially depressed by a $48.8 billion single-year jump in capital spending tied to AI infrastructure. This is not indicative of the business’s ongoing earning capacity. FY2024 FCF of $32.9B, which clusters tightly with FY2023 FCF of $32.2B, better represents normalized free cash flow. I am using $32.9B as my valuation base.

This is the same playbook Amazon ran in 2021-2022 with the COVID-era warehouse buildout. They spent heavily for two years, FCF went negative, and then operating cash flow exploded on the other side. The market did not believe it then either.

This is where the AI capex story either holds up or falls apart. Let’s look through three lenses.

Debt to Free Cash Flow. Using normalized FCF of $32.9B and long-term debt of $65.6B (per the 10-K), the ratio is 2.0x. This means Amazon can pay their debt off in two years based on the cash flow they generate. That is Great by our framework (≤3x). Even if you load in the long-term lease obligations of $87.3B and treat them as debt, you get a total debt load of $153.0B, which is 4.65x normalized FCF — still Solid (≤5x). However you slice it, the leverage is manageable.

Debt to Equity. The D/E ratio is 0.37, down from 0.46 in 2024 and 0.67 in 2023. The trend is clearly improving, even while management is taking on more debt to fund AI buildout, because equity is growing faster than debt.

Excess Cash. Amazon ended 2025 with $123 billion in cash and short-term investments against $66 billion in long-term debt. That is $57 billion of net cash on the balance sheet. This is a balance sheet with real flexibility. Management can keep funding the AI cycle without breaking anything.

Interest Coverage. Operating income of $80.0B against interest expense of $2.3B gives an interest coverage ratio of 35.2x. That is not a typo. Amazon could lose 95% of its operating income and still cover interest. There is no debt risk here.

Dilution is a risk most investors ignore. They look at total profit, not profit per share. Big mistake.

Over five years, the diluted share count grew from 10,200M to 10,827M. That is total dilution of 6.2%, or about 1.2% per year on average.

The driver is stock-based compensation. SBC was $19.5B in 2025, $22.0B in 2024, and $24.0B in 2023. Notice the trend. SBC is decelerating, not accelerating. Management is being more disciplined with stock grants, even as the company grows.

A 1.2% annual dilution rate for a business growing free cash flow at double digits is acceptable. You are losing 1.2% of your ownership stake each year, but the per-share value of what you own is still climbing because earnings are climbing faster. It is something to monitor, not something to fear.

Every business has cash at their disposal. It cost the business to have this cash in their position. This is called the cost of capital. This is the hurdle rate of the business, what they must earn to break-even on the money invested in the business. Beat that rate, and the company is creating value for shareholders. Miss it, and the company is destroying value, even if profits look fine on the surface.

Return on Invested Capital (ROIC) is what the company actually earns. Weighted Average Cost of Capital (WACC) is what it has to earn. The spread between the two is the test.

Five-year average ROIC: 8.4%. Five-year average WACC: 7.1%. Average spread: +1.28 percentage points.

Amazon is creating value. The spread is not enormous on a five-year average basis, but the trend is what matters. 2022 was a disaster year (the company essentially broke even on operations), and that one year drags down the entire five-year average. Pull 2022 out, and the average spread is closer to +3%.

The trajectory is clearly improving. ROIC has been above WACC for three straight years and is accelerating. Management is putting capital to work at returns above what investors require. That is the definition of value creation.

Operating cash flow tells you what a business generates. Where management spends that cash tells you what management believes about the business.

Here is the FY2025 breakdown.

Read that table again. Amazon spent 94.5% of all operating cash flow on capital expenditures in 2025. They borrowed an additional $10 billion on top of that to fund more spending.

This is a single message from management: we believe the best return on every dollar of cash is to put it back in the business. Specifically, into AWS data centers and AI infrastructure.

Is that the right call? Look at the segment growth rates. AWS grew 19.7% in 2025. Advertising grew 22.1%. Both are higher-margin than retail. If management can convert today’s CapEx into more AWS and Ad revenue tomorrow, the answer is yes. If the AI cycle disappoints, the answer is no. This is the core risk of the stock. It is also the core opportunity.

Now we put it all together.

DCF Inputs:

DCF Output:

Sum of present values: $2,749 billion. Divided by 10.757 billion shares:
Intrinsic value of approximately $255 per share.

Current price: $273.55. That is +7.1% above intrinsic value.

Fair Value Range (±10%): $230 to $281.

The current price sits inside the upper half of my fair value range. Not at a discount, but not stretched either. This is what fair value looks like for a quality business.

Implied growth check. Working backwards from the current price of $273.55, the market is pricing in a free cash flow CAGR of approximately 18.9% per year for the next five years. My assumption is 16.85%. The market is pricing in about 2 percentage points more growth than I am. That is a modest premium, not a stretch. If AWS keeps growing 19% a year and Advertising keeps growing 22%, the market’s number is achievable. If the AI cycle disappoints, the market’s number is too rich.

Amazon is a premium quality business trading inside its fair value range. Premium businesses deserve premium prices, and at $273.55, that is exactly what you are paying — a fair price for a high-quality compounder. The balance sheet is strong, the margin profile is improving every year, the mix is shifting toward higher-margin segments, ROIC is rising and is comfortably above WACC, dilution is decelerating, and management is putting every available dollar of cash back into the highest-return parts of the business.

This is not a deep value play. It will not double in a year. But it does not have to. What it offers is a quality compounder at a fair entry price, with a clear path to higher free cash flow once the AI capex cycle moderates.

Adding $AMZN to The Investing Department watchlist.

Amazon analysis performed on 10/17/2024

Disclaimer: This analysis is published by The Investing Department for educational purposes only. It is not financial advice. Always do your own research and consult a qualified financial advisor before making any investment decisions.