Matthew Martin, senior U.S. economist at Oxford Economics, joins BNN Bloomberg to discuss the company’s report on the AI trade boom.
The global buildout of artificial intelligence infrastructure is reshaping trade flows and the freight networks that support them. Its physical footprint extends beyond chips to the equipment and materials needed to develop data centres.
BNN Bloomberg spoke with Matthew Martin, senior U.S. economist at Oxford Economics, about the firm’s report mapping AI-related trade and what cross-border shipments can reveal about the investment cycle.
Key TakeawaysU.S. AI-related imports are up 60 per cent year over year, compared with about eight per cent for other imports, Martin said.Strong demand is sustaining AI-related trade despite tariffs and higher freight costs, although prolonged trade fragmentation could slow productivity growth.Asia anchors the upstream supply chain, with Taiwan specializing in chip manufacturing and China supplying a broader range of equipment and infrastructure.The U.S. and Canada retain strengths in research, development and innovation, even as data centre construction depends heavily on imported goods.Disruptions to helium supplies from the Middle East highlight the AI supply chain’s dependence on specialized chemicals and materials.
Matthew Martin, senior U.S. economist at Oxford Economics Matthew Martin, senior U.S. economist at Oxford Economics
Read the full transcript below:
ROGER: The report by Oxford Economics is mapping the AI trade boom. It’s looking at the movement of goods needed to deliver the current global AI buildout. To discuss more, let’s bring in Matthew Martin, senior U.S. economist at Oxford Economics. And Matthew, thanks very much for joining us today.
MATTHEW: Thank you very much for having me. It’s a pleasure to be here.
ROGER: Can we safely say AI, I think, is pretty much everywhere now? What stood out for you in this report?
MATTHEW: Yeah, I mean, it’s hard to go a day without hearing about AI. It’s become pervasive in all parts of the economy, and certainly trade is not the exception to the rule. What we’ve looked at and mapped across countries is really trying to track, you know, not only semiconductors, which is really just the headline, the tip of the spear, but all the component parts that go into the AI supply chain.
And what we find is that, you know, in a world that is seemingly attempting to be less globalized, in one of the most important parts of the growing economy, AI, it is very, you know, obviously interconnected. You know, there’s a lot of trade at the start in Asia, whether that’s China, Taiwan, South Korea, and intermediate goods that flow across to the — to North America, whether that’s Mexico or the U.S., where the end goal is.
But there’s a lot of components that are not only going from one country to next, but flowing back and forth across borders before it ends up in those data centres that we’re seeing a large buildout. And there’s a lot of different modes of transportation and a lot of interdependencies that obviously make this a very, you know, fine balance and a balancing act between countries.
ROGER: And do we see that? I mean, we’ve seen so many other — with this trade wars that are going on and the tariffs that are going on. Are we seeing the AI net being stretched and strained, or are they? Is there a threat that that may break apart that global interaction?
MATTHEW: You know, I think that’s certainly a threat to things. You know, higher costs for goods certainly can be an impediment to trade and might, you know, divert production to other areas in other countries. What I would say, though, is that demand is so strong at the moment that, you know, any increased cost and tariffs is really not much of a, you know, a roadblock at the moment.
Another area to obviously watch is the rising fuel costs that we’re seeing because of the Iran war. You know, air freight is a large component of trade as well, and so higher jet fuel prices obviously contribute to higher freight costs and ultimately higher costs of goods. But again, demand is so strong that it seemingly isn’t an impediment to this trade.
And so, while I would say risks are out there, and certainly in the longer term, more fractured trade would, you know, certainly potentially slow down productivity, you know, increase cost of goods sold. It doesn’t look to be an impediment at the moment, and what we’re seeing at the — at this time is that it’s really full steam ahead.
ROGER: All right, and it’s centred in Asia. Is there one point in one place in Asia that is the actual absolute focal point of it?
MATTHEW: You know, I think the two countries that really stand out, of course, are China and Taiwan. Taiwan is really the manufacturing hub for chips. TSMC is obviously one of the leading companies in the world for that, and, you know, for them, that accounts for more than 80 per cent of their trade. So clearly, that is a very vital component of their economy.
But also, is a good leading indicator for, you know, what we can expect in investment cycles. If those exports continue to rise, that’s a good leading indicator for imports across the country, demand across other countries. Whereas if they start to fall, obviously that would be the reverse.
And the other end of the spectrum, you have China, which is, you know, you know, a big player in the chips, but they’re also a big player in kind of some of the bulkier items and the kind of the power infrastructure, computers, circuit boards that are, you know, obviously components that go into ensuring that those chips ultimately perform the function that they’re needed to in those data centres.
And so I think those are the two biggest countries, you know, at the start upstream of the value chain. But of course, South Korea, Singapore are also very important players, and so really Asia dominates the start of the value chain upstream. Whereas if we move further along the line downstream, it’s really U.S. that is the biggest, you know, demand for these components.
I think it’s really useful to look at, you know, trade data in the U.S. We look at AI-related imports that are up 60 per cent year over year. If we look at all other imports, they’re up around eight per cent. So, you know, there’s a large-scale difference between, you know, demand for AI products in the U.S. versus all other goods.
And again, it’s just a leading indicator to show that AI demand is, you know, still hasn’t been satiated. We’re still, you know, moving forward. Data centre growth is still very strong, and so we’ll continue to see those trade growth starting in China, Taiwan, and moving over to Mexico, Canada and the U.S.
ROGER: Is there a danger that North America — I mean, it’s downstream, but is it there a danger that we could be left behind, not being as involved in the upstream aspect of it?
MATTHEW: You know, I think it’s, you know, obviously a choice for companies, for countries, where they want to specialize. But I would say is that the U.S., Canada, while they not might not be the production hub, certainly they are the leaders in R&D, innovation and cutting-edge technology in these products, and so, you know, it’s really just a question of where to specialize.
You know, certainly there are limitations there. You know, what we talk about a lot in the U.S. is this booming investment growth for data centres and related equipment, and you see, you know, large capital expenditures from companies in hundreds of billions of dollars, if not trillions, at this point.
But in order to build those data centres, we’re very reliant on imports, and so a lot of that, you know, capital expenditures is really being, you know, sucked out of the country and goes towards Taiwan, China, where we’re, you know, obviously needing to import goods. So, you know, it’s a bit of a balancing act there.
But ultimately, you know, the U.S., Canada are really at the top of the downstream, the really the cutting technology innovation and R&D, and ultimately that will be a large, you know, driver of growth, profit, and a good place to be as well, not just at the production side.
ROGER: Okay, I mean, now chips, computers, cable and storage. I’m just looking at cabling and power are three of the big things of AI-related goods. Are there some things that people don’t think about that are maybe in that other category that might surprise?
MATTHEW: Yeah, I think one that a lot of people don’t think about is the, you know, chemicals and materials that ultimately are needed to, you know, you know, process this equipment. It’s a very specialized process. There’s a lot of, you know, materials that are needed even just to handle them, and I think that’s one area where, you know, maybe supply chains weren’t, you know, as readily available when Iran war kicked off.
The Middle East is a large producer of hydrogen, which is a — just needed to, you know, produce and handle these goods. Sorry, not hydrogen, helium. Apologies. And that is, you know, where a lot of the exports come from, and so with that being cut off, you know, that’s a, you know, critical area where, you know, supply chains could see stress, higher costs because there’s not readily good availables to, you know, produce.
And so it’s really not even just the equipment itself, but, you know, those inputs, chemicals, materials that are needed to handle them as well, and so, you know, there’s a fine balance that, you know, these supply chains have to strike. You know, certainly it’s a very globally interconnected one.
So where we see these geopolitical events occur, and we’re starting to see them occur more regularly, you know, certainly there is additional risk, you know, to be found.
ROGER: Okay, we have to wrap it up there, Matthew. But thank you very much for joining us.
MATTHEW: Thank you for having me.
ROGER: Matthew Martin, senior U.S. economist at Oxford Economics.
—
This BNN Bloomberg summary and transcript of the Oct. 14, 2026 interview with Matthew Martin are published with the assistance of AI. Original research, interview questions and added context was created by BNN Bloomberg journalists. An editor also reviewed this material before it was published to ensure its accuracy and adherence with BNN Bloomberg editorial policies and standards.