AI Spending: A New Engine for the Global Economy

**Seth Carpenter** (0:00)
Welcome to Thoughts on the Market. I'm Seth Carpenter, Morgan Stanley's Global Chief Economist and Head of Macro Research.

**Michael Gapen** (0:06)
And I'm Michael Gapen, Chief US Economist.

**Chetan Ahya** (0:08)
And I'm Chetan Ahya, Chief Asia Economist.

**Jens Eisenschmidt** (0:11)
And I'm Jens Eisenschmidt, Chief Europe Economist.

**Seth Carpenter** (0:14)
And today is going to be our third quarter economic roundtable, taking a wide angle view on the global economy and all the key forces shaping our outlook and the economy. It's Monday, July 20th at 10 a.m. in New York.

**Jens Eisenschmidt** (0:27)
And 4 p.m. in Frankfurt.

**Chetan Ahya** (0:28)
And 10 p.m. in Hong Kong.

**Seth Carpenter** (0:32)
Since our last roundtable in April, the global economy has continued to face all sorts of shocks, a mix of resilience and friction. Inflation pressures have not disappeared. Energy and geopolitical risks have come up. They've receded. They've come back. They've receded all over the place. But there is one underlying source of momentum that we have to talk about, and that is the AI-driven capex cycle.
Michael, let me turn to you because the US is a real focal point of all of this.
Tell me a little bit about where Morgan Stanley Research is thinking about hyperscaler capex, how big it is, and then for you, when you think about the US economy, just how big of a driver is it for what we're looking for in the US?

**Michael Gapen** (1:14)
We continue to revise higher our estimates for hyperscaler and AI-related capex in the US economy.
We were thinking a little over a trillion for 2027 Now, we're more like 1.2, 1.3 trillion, maybe as high as 1.4 trillion in 2028 So the level of hyperscaler spending continues to keep rising. The growth rate and its effect on the economy is likely to slow, but as you noted, it's still a major driver of momentum in the US.
You would look at that headline number and think, wow, that's 3.5% or so of GDP, must be a massive source of momentum for GDP growth. But roughly about 60% of that hyperscaler capex spending goes to items like computers and peripherals, equipment spending categories that have a very, very high import content. We still get a significant number that AI capex is probably contributing around 40 basis points to growth this year. It'll be a similar sized amount perhaps next year. So for an economy that's growing somewhere a little bit above 2% right now, maybe closer to 2.5% next year, that's a non-trivial amount. We just have to remember it's fueling growth around the world, just not here in the US.

**Seth Carpenter** (2:36)
Yeah, that's a really great point because I have seen some estimates where people say, well, if it wasn't for AI capex, the US economy wouldn't have grown at all, and that's clearly wrong as you point out. But US imports aren't necessarily exports from somewhere else. Chetan, if I can pull you into the story then, US firms are buying a lot of AI related equipment from Asia.
What does that mean in your part of the world? In particular, I'm thinking about Korea, Taiwan, and maybe some other economies in Asia. What's the critical story there?

**Chetan Ahya** (3:05)
For Asia, this has definitely been a big boom. If you look at Asia's exports, they have been booming, and particularly for the ones which are exporting semiconductors to the US., they are seeing semiconductor exports growing by 90%.
When we go back in time and compare Asia's semiconductor exports, it's very tightly linked to the US. IT capex. It's not surprising when Mike Gapen mentions about the imports going up, it's on the other side helping Asia's exports quite meaningfully. So, so far, we have seen this benefiting Korea, number one, Taiwan, and also Japan. All these three are big beneficiaries of US. AI capex. And of course, also not just US., but the other countries which are doing any little amount of capex on AI front, that's also helping these three economies in the region.

**Seth Carpenter** (3:59)
You've been doing a lot of work, Chetan, recently about how much the story can actually broaden out that the AI capex cycle has really contributed to Asian growth, but it doesn't tell the whole story that there's a broader industrial cycle. Can you give us a little bit of a flavor of that story?

**Chetan Ahya** (4:17)
That's right, Seth. So we are actually highlighting that there is a capex and industrial super cycle that is underway in Asia. And there are four components to this story, AI and semiconductors capex, which we just briefly discussed. Number two is energy. Number three is defense. And number four is industrial supply chain on-shoring related capex. I know that everybody still thinks that AI is the most important part of this story. But when I give you the numbers and the breakup of that. So for Asia, AI and semiconductor companies capex is about $380 billion in 2026