Amazon's AI Business Passed a $25 Billion Run Rate. Where Does That Put the Stock in 5 Years?

Amazon(NASDAQ:AMZN) put a number on its artificial intelligence (AI) business in its July 30 earnings report, and the number deserves more attention than it got. The AI business inside Amazon Web Services (AWS) has passed a $25 billion annual revenue run rate, management said, and it is growing at triple-digit percentages year over year.

(Amazon's custom chip business separately passed the same mark. The company hasn't said how much the two overlap.)

A business of that size, and growing that fast, inside a company priced on ordinary growth assumptions, can decide a five-year return on its own. What could it make the stock worth by 2031?

Image source: Getty Images.

Inside the acceleration

The AI line sits inside a segment that is itself speeding up. AWS grew net sales 37% year over year to $42.2 billion in the second quarter (growth that management called its fastest in 18 quarters), putting the unit at a $169 billion annualized revenue run rate.

To me, the margin move is even more impressive. AWS operating income reached $16.6 billion for the quarter, up about 64% year over year from $10.2 billion, expanding the segment's operating margin from about 33% to about 39%.

And the demand reaches years out. CEO Andy Jassy said on the July 30 earnings call that AWS's backlog of contracted work reached $496 billion in the quarter. Measured against the unit's current pace, the backlog alone is close to three times a full year of AWS revenue.

The $25 billion AI line is the fastest-moving of those AWS numbers. After all, triple-digit growth means it is on pace to double again within about a year.

Five years of compounding

If the AI business doubles just twice more by 2031, that works out to about 32% annualized growth (a hard deceleration from triple digits), and it would still take the line from $25 billion to about $100 billion of annual revenue.

The rest of AWS, about $144 billion of the current run rate, growing at a far more ordinary 12% to 15% a year, could reach somewhere between $250 billion and $290 billion. Put together, AWS could plausibly be a $350 billion to $390 billion business by 2031 -- about double the whole segment today.

Margins decide what that is worth. If AWS holds an operating margin anywhere near the second quarter's 39%, the segment alone could produce somewhere around $140 billion of annual operating income by 2031. For perspective, the entire company generated $27.5 billion of operating income last quarter, about $110 billion at an annual pace.

And none of that counts the retail and advertising businesses that still produce most of Amazon's revenue.

The price already assumes a lot

Amazon's market value is about $2.82 trillion as of this writing, with shares around $261, about 9% below their early August record. The stock costs about 28 times forward earnings. The price-to-earnings ratio looks far cheaper, but it leans on a one-time item: the $53.4 billion of non-operating income, mostly gains tied to the company's Anthropic investments, that landed in the second quarter.

For the stock to compound at 12% a year through 2031, Amazon's value would need to reach about $5 trillion. At 25 times earnings at that point (a slightly cheaper multiple than today's forward figure), the company would need about $200 billion of annual net income, roughly double the earnings base the current price rests on.

Doubling profits in five years works out to about 15% a year. That's the growth the current price quietly assumes.

Whether it's conservative or generous comes down mostly to AWS -- and within AWS, to the AI line. If the segment's margin holds while the AI business scales, the target gets easier each quarter. If heavy spending eats the margin instead, the 2031 numbers may prove out of reach.

Of course, the spending is enormous. Management raised this year's capital expenditure plan to about $220 billion, and Amazon's free cash flow over the trailing 12 months swung to a $7.6 billion outflow. And the depreciation bill from those data centers will weigh on reported earnings well past 2026.

But a $25 billion business growing at triple-digit rates is exactly what all of that spending was supposed to create. I think the AI line is what makes the five-year math plausible. At about 28 times the earnings expected of it over the next 12 months, the price assumes strong growth. Based on what AWS just reported, the assumption looks reasonable to me.

Daniel Sparks and his clients do not have positions in any of the stocks mentioned. The Motley Fool has positions in and recommends Amazon. The Motley Fool has a disclosure policy.