Episode summary
Can the AI Spending Boom Pay Off?
Thoughts on the Market (Morgan Stanley)
· · 5 min · Guests: Brian Nowak
Summary
Brian Nowak argues that the huge AI infrastructure buildout can still create attractive economics, even with major cloud providers set to spend more than $1.4 trillion on AI next year alone. He says compute capacity could quadruple from 2025 to 2028, reaching roughly 120 gawatt, but the bigger question for investors is not the size of the spend; it is the return on invested capital that these assets can generate.
He lays out three business models and says all of them can produce meaningful returns. Renting compute power, the infrastructure layer of the AI economy, can in his base case deliver about 30% ROIC, with outcomes ranging from the low 20s% to nearly 40%. A vertically integrated AI lab that owns both the model and the infrastructure can do better, with roughly 75% incremental operating margin and 40% plus ROIC.
The least attractive case is when AI developers rent infrastructure instead of owning it, but even there he sees roughly 30% incremental operating margin and 25% post tax return potential. He says the key variables are token pricing and infrastructure efficiency, which makes chip and software improvements essential to the long-term economics of the AI ecosystem.
Key takeaways
- Brian Nowak says major cloud providers are set to spend more than $1.4 trillion on the AI buildout next year alone.
- He sees compute capacity potentially quadrupling from 2025 to 2028, reaching roughly 120 gawatt.
- His base case for renting AI compute power is roughly 30% ROIC, with a range from the low 20s% to nearly 40%.
- An AI lab that owns its model and infrastructure could generate roughly 75% incremental operating margin and 40% plus ROIC, according to Brian Nowak.
- Even AI developers that rent infrastructure instead of owning it could still earn roughly 30% incremental operating margin and 25% post tax return potential.
- Brian Nowak says token prices and token throughput per GPU are the two biggest drivers of whether the economics work.
Notable quotes
“we estimate that the major cloud providers are going to spend more than $1.4 trillion on this AI buildout next year alone”
“We see paths to roughly 25 to 50% return on invested capital or ROIC across three emerging AI business models”
“In our base case, a large next generation data center can generate a return on invested capital of roughly 30% simply renting AI compute power”
“our base case generates a roughly 75% incremental operating margin and a return on invested capital of 40% plus”
“our base case still produces roughly a 30% incremental operating margin and 25% post tax return potential”
How the episode opens
Welcome to Thoughts on the Market. I'm Brian Noak, Morgan Stanley's US Internet analyst. Today, can the enormous investment behind Gen AI actually pay off?
It's Wednesday, September 9th at 9:00 a.m. in New York. AI has moved quickly into everyday life.
It helps people write software, research purchases, automate work, find information, among a myriad of other use cases. But we need an infrastructure buildout of extraordinary scale to support all of this activity and more activity to come. In all, we estimate that the major cloud providers are going to spend more than $1.4 trillion on this AI buildout next year alone.
But compute capacity is potentially going to quadruple from 2025 to 2028, reaching roughly 120 gawatt. But all of this spending has raised a lot of questions for investors. One of the most common questions is what kind of return on invested capital can these companies earn from all of these trillions of dollars of data center infrastructure investment.
Well, our bottomup work points to encouraging answers to this question. We see paths to roughly 25 to 50% return on invested capital or ROIC across three emerging AI business models. Now, ROIC is a useful way of measuring whether investments pay off.
Think of it as how much after tax operating profit can be generated relative to the capital required in the first place. The first business model we've analyzed is renting compute power. This is the infrastructure layer of the AI economy.
Cloud providers build data centers filled with advanced graphics processing units or GPUs and rent that compute capacity to customers. In our base case, a large next generation data center can generate a return on invested capital of roughly 30% simply renting AI compute power. And even if rental prices move around, our scenarios still produce returns ranging from low 20s% to nearly 40%.
From the episode’s captions: roughly the first two minutes, lightly cleaned. Keep watching on YouTube →
Topics
- ai spending boom
- gen-ai business models
- big tech capex
- investor returns
Watch the episode
Watch “Can the AI Spending Boom Pay Off?” on YouTube
Summary, takeaways and quote selection by Octagon. Quotes are short excerpts; the full conversation is in the episode.
More from Thoughts on the Market (Morgan Stanley)
Thoughts on the Market (Morgan Stanley), this episode and its artwork belong to Morgan Stanley. Octagon is not affiliated with the show; this page is an independent summary.
