Episode summary
Why Goldman Sees a $7.6 Trillion AI Buildout Before 2031
Summary
Dylan Grice argues that Goldman Sachs’ estimate of a $7.6 trillion AI buildout through 2031 marks a massive reindustrialization trend centered on compute, data centers, and power. He says the scale is larger than the U.S. federal budget and implies that capital spending at this level could create outsized winners in the same way Cisco and Qualcomm benefited from the 1990s internet infrastructure boom.
He frames the opportunity as an “AI stack” with five layers: energy at the base, then chips, data centers and infrastructure, models, and finally applications. His core view is that the deepest and most durable economic value will come from the layers closest to infrastructure and from the applications layer, while many companies are exposed to only one part of the stack.
Grice also says Elon Musk’s SpaceX should be viewed as more than a rocket company, pointing to the S-1 filing as evidence that it is becoming a holding company for AI. In his telling, SpaceX is using cash flow from its legacy launch business to fund AI ambitions, and Musk is already vertically integrated across energy, infrastructure, models, and applications, with chips the main gap he still needs to fill.
Key takeaways
- Dylan Grice says Goldman Sachs is projecting a $7.6 trillion AI buildout in compute, data centers, and power before 2031.
- He argues the AI spend wave is comparable to past infrastructure booms that produced huge winners like Cisco and Qualcomm.
- Grice says the AI stack has five layers, with energy at the base and applications at the top where much of the economic value will be captured.
- He says SpaceX is evolving beyond rockets into a broader AI holding company, using launch cash flows to fund the next phase.
- Grice says Elon Musk is missing chips in his vertically integrated AI strategy and has a $25 billion incentive to make moves this year.
Notable quotes
“They say that is going to be a 7.6 trillion-dollar build-out over the next 6 or 7 years before 2031”
“We see Cisco own the pipes in the '90s, stock jumps 67,000%. And you think of Qualcomm back then, up 10,000%”
“If you look at the AI stack, it's basically a five-layer cake. And at the very bottom, you have energy”
“SpaceX isn't really just the rocket company anymore. It seems that it's a holding company for AI”
“he's got a $25 billion incentive to move on chips to fill the five layers of the cake he's building”
How the episode opens
Morning. This is Dylan Jovaneer behind the markets. Happy Thursday.
Today is Thursday, June 18th. We have a webinar at 1:00 today. I'm very, very excited about it.
It's such an interesting story, really. Goldman Sachs is out with a report last week saying they add all up the spending, the targeted planned spending over the next 6 years, compute data centers and power. They say that is going to be a 7.6 trillion-dollar build-out over the next 6 or 7 years before 2031.
7.6 trillion dollars. That's bigger than, you know, most national budget. Our budget is 4 trillion.
So, you see this reindustrialization of America. You see it through Elon and his new SpaceX IPO. SpaceX isn't really just the rocket company anymore.
It seems that it's a holding company for AI. And he outlines it in his S1 very, very clearly, where he's joining that 7.6 trillion-dollar infrastructure build-out in a very, very, very big way. He's basically joined the race.
That's what this is about. We talked about it a little bit yesterday. If you look at the AI stack, it's basically a five-layer cake.
And at the very bottom, you have energy cuz again, every tech revolution in history is an energy revolution at root. You think of the railroads, you know, you think of coal for railroads, you think of gasoline for cars, you think of the internet, energy consumption went up greatly. They're all energy revolutions at root, okay?
So, that's the very foundation of the cake. You have chips, semiconductors, very, very important. Obviously, it's another layer of the cake.
You have data centers infrastructure, AI factories, all these things. You have models on top of that second to the top layer where you have, you know, you have a Grok, you have ChatGPT, you have a Claude.
From the episode’s captions: roughly the first two minutes, lightly cleaned. Keep watching on YouTube →
Topics
- ai buildout
- energy infrastructure
- data centers
- chip supply
- ai capex
Companies and tickers mentioned
- GS
- CSCO
- QCOM
- Goldman Sachs
- SpaceX
- Cisco
- Qualcomm
- OpenAI
- Anthropic
- xAI
Watch the episode
Watch “Why Goldman Sees a $7.6 Trillion AI Buildout Before 2031” on YouTube
Summary, takeaways and quote selection by Octagon. Quotes are short excerpts; the full conversation is in the episode.
More from Behind The Markets Podcast
-
Nvidia CEO Out: 4 Reasons I’m Not Panicking About Nvidia -
The Mag 7 Sell-Off: My 3 Secrets to Picking Stocks -
Morgan Stanley’s AI Winners List Is Full of Bank Stocks -
Anthropic IPO: Why I’m Not Rushing to Buy at $1 Trillion -
Chip Stocks Sell-Off: Why the AI Build-Out Just Went From 3 to 5 Years -
Midterm Election Results: The Stock Market’s Winners and Losers
Behind The Markets Podcast, this episode and its artwork belong to their owners. Octagon is not affiliated with the show; this page is an independent summary.
