The ‘Capacity Tax’: How AI Data Centers Alter Freight Markets

Speaker 1 [0:00] We have Blake Azell. He’s the Vice President of Customer Success and Support at IntelliTrans. Blake, welcome to FreightWaves Today. Your comment or your theory on this, you have argued that this is not the same recovery freight market that everyone’s talking about. What is your, what is your sort of view of the market right now?

Speaker 2 [0:20] Yeah, good afternoon. Essentially, you know, I think it’s everyone has assumed for, for many years that a healthy market, you know, is a tight market with high prices. Essentially leads to a healthy market. And, you know, I don’t think that’s the case today as we look at, you know, I, as I talk with many shippers and what they’re experiencing, there’s a big gap between what’s going on and some of the key market indices.

Speaker 1 [0:46] Now, for those that aren’t familiar with IntelliTrans, you’re part of Roper Technologies. DAT is a sister company, competes, we compete with DAT. We love to hate them here at Freightways Today, but we’ll let you off the hook there. But you guys are at TMS really for the bulk and break bulk market, which is different than a lot of our core audience. You know, when we look at the bulk rate data, which we have inside of Sonar, it hasn’t seen the same volume increases and high rates that you see in some of the over-the-road market. Why is the bulk market not as, not as vibrant, we’ll say, as the overall truckload market?

Speaker 2 [1:29] Sure. Well, I think if we look at 2026 at the freight market, it’s just genuinely tight, right? So we know we’ve got a number of different industries that are either at historic extremes, at all-time, all-time records, or even rates, you know, we’re still seeing rates from, from 2021. The issue becomes the— is the volumes, and we’re looking at volume, softened volumes from about 2023. And so what we’re trying to understand is why is it that, you know, what’s really driving— why do we have, say, on the bulk and break bulk side, a flatbed index of 73 to 1, you know, load-to-truck ratio? Why is that happening and what’s really driving it? When we look at the volumes, we look at, you know, a softened market of, you know, call it 4%, you know, volume erosion.

Speaker 1 [2:19] What is the— in your customer base, what are customers thinking about right now? I mean, if the market volume isn’t as robust as perhaps just the overall capacity situation, what, what are they doing to navigate this?

Speaker 2 [2:33] Yeah, yeah, good question. So, um, now specific, we’re in the bulk and breakbulk space, so think, uh, chemicals, plastics, forest products, building materials, metals, mining, etc., etc. Um, you know, it really is what is going on, what’s going on in the market, you know, what are their structural changes that’s happening, but where is the capacity going, right? Where’s the capacity going? And why are those rates going up? You know, I think a few things to look at. We’ve got, we’ve got a serious kind of a paradox, right? So like I said, we’re essentially paying premium, premium pricing for softened volumes. And that’s, that’s being driven by a few things. So going back to, you know, roughly around 2020, we started seeing erosion in driver availability. And I think this is, this is a key piece to it. But then there’s also what I would say there, you’ve got, you’ve got driver erosion outside of retirement. You’ve got cash flow economics right now, especially when we look at fuel surcharges and the impacts of that on independent carriers. And we see a lot of that within the flatbed markets, the specialized asset market. And then at the same time, there’s also kind of a silent competitor that has come into play now that appears to be absorbing a lot of that capacity at different margins than what these commodity shippers have been able to withstand.

Speaker 1 [4:00] What is that? You say silent competitor. What do you mean by that?

Speaker 2 [4:03] Yeah, so, you know, AI, obviously AI, it’s the big boom everyone wants to talk about, right? So cloud computing, but behind cloud computing is it’s the data centers, right? It’s large data centers. And it’s all the freight that moves into those data centers. Those aren’t moving typically in in a van, right? I mean, it’s going to be on your flatbed freight and specialized. So that data center world really is, you know, I think I’ve seen numbers as much as, you know, roughly worth, you know, at this point around $83 billion, like projected to grow into about $150 billion over the next 5 years. All of that demand from that data center growth, that’s where I believe we’re starting to see the capacity moving from our traditional freight markets, moving in that direction.

Speaker 3 [4:52] Yeah, I mean, I think there’s no question about that. We’ve talked a lot about the freight needs and the amount of freight that the AI construction and data center boom has created. Craig could quote the stats better than I can.

Speaker 1 [5:05] 7% flatbed. I mean, we’ve got this.

Speaker 3 [5:08] We’re spending every 5 weeks the same amount as we spent every 10 weeks?

Speaker 1 [5:11] No, it’s every 2 weeks.

Speaker 2 [5:12] Okay.

Speaker 1 [5:12] $700 billion. So the interstate highway system was $20 billion a year over 35 years. The hyperscalers are spending that every 2 weeks to build out data centers. Most of that, or over half of that, is going by flatbed. And it, look, it’s not obvious to people. People are like, well, GPUs. No, we’re talking about concrete, steel, PVC pipe, generators, transmission, copper, fiber optics, steel. All of that has to move to build out AI data centers. 30,000 truckloads, Julie, to build a 500-megawatt data center. And now there’s an announcement not too long ago of a 10-gigawatt, which is 20 times the size, in Utah data center that Kevin O’Leary is trying to put up. The numbers are astounding.

Speaker 3 [5:59] So I want to make sure I understand your thesis here about what makes a healthy freight market in the situation we’re in. So I think there’s no question we’ve all felt like this this cycle has been driven by lack of capacity, not by increased demand. However, there has been new demand specifically for flatbed coming out of these AI data centers. And then when we track the bulk and the other commodities, many of which you mentioned in the rail numbers each week, we have seen gradual increases and surge in volumes in all of those. So I want to understand where you feel like the pressure is and what you think is the most imbalanced part of our freight market currently.

Speaker 2 [6:41] Well, we’ve got— so if we look kind of on, like, on the rail side, we, we’ve seen marginal volume increases year over year. Uh, chemicals, forest product, non-metallic minerals, motor vehicles, all basically flat to, you know, a deterioration. We’ve really seen— so if we look at kind of the, the key pieces of business that, uh, we would anticipate moving within the data centers, that’s where we’ve seen the growth. We’ve seen, um, it’s essentially we’re seeing this digital capacity. Um, the digital world is creating a— call it a parking lot— so much within our physical assets for the supply chain that’s moving, moving flatbed. Um, again, we’ve seen roughly, I think it’s, it’s, uh, 4% volume decline, right? So 16% tender rejections. We know the markets are soft within, within the key, uh, the key, you know, typical commodity segments. It’s soft, volumes are increasing, and, and it’s our, our shippers, uh, our, our commodity shippers are now almost getting into a bidding war with the tech giants, right? And so that’s really where the concern is. Um, you know, Craig, like you said, they’re moving just about everything in, from, from HVAC to wire to concrete to structural steel. Everything is moving into these markets. And, and, you know, for logistics managers, it’s what can they do, you know, how can they best position themselves moving forward? Because for the next foreseeable future, unless there are structural changes within the data center environments and regulation, then we would infer that those markets will continue to stay very tight for the near future.

Speaker 1 [8:29] Blake, it’s not going away. We did a long-form, we call it a sitrep, which is a deep dive into AI data centers. Only 8%, this is when we did it a couple months ago, 8% of the data center announcements that have been announced and contracted are under any level of construction. And so if you look at this chart that we got pulled up in terms of the amount of truckloads, by the way, this is truckloads. And the good rule of thumb is 50% to 70% of that number, depends on how you classify it, is in the flatbed open dimensional markets. Some of it is in bulk, because concrete moves in bulk, gravel moves in bulk. We’re talking about this being a multi-year, even with some of the data centers delayed, or states and cities and municipalities doing moratoriums on data centers, it’s not going away. And so this is going to be something that’s going to be around. I mean, look at the fact we’re at 3 million truckloads this year, and we’re ramping to just every single year is an increase. And it would be foolish to think that AI data center construction is somehow going to go away. These are By the way, these are public— a lot of people like to compare the dot-com fallacy and the dot-com days as like this is a bubble. Maybe it is. But these companies that are investing CapEx are the most profitable cash-flowing, operational cash-flowing companies on the planet in the history of mankind. They’re not going anywhere. And this is not only the race to build AI, you also have competition with China, you have competition with Europe, you have competition with the Middle East. Saudi Arabia is going to invest hundreds of billions of dollars in— they had this massive line across the desert city. They’ve turned it into a massive data center, because they’re like, that’s a lot more interesting to us. This is not going to go away. The Department of Defense, a big part of their budget is towards data centers. I would argue that this is a permanent structural change in our economy.

Speaker 2 [10:29] No, absolutely. The irony is, uh, it’s AI and cloud computing that is what is creating it, and it’s meant, you know, for all intent, you know, to make us more efficient, to get us information faster, um, you know, to save cost, etc., etc. But at the, at the heart of this, it’s also driving the problem that is making everything else more expensive when we look at the markets that these other commodities actually support— our food, our housing, it’s all getting more expensive and AI is somewhat at the heart of that.

Speaker 1 [11:01] Yeah, construction workers, if you’re building homes, one of the things that isn’t being talked about is construction workers are going to these AI data centers. And the AI data center customers, the hyperscalers, are price insensitive. They do not care about what normally you have a very rational market with traditional shippers in our world are very sensitive to small changes in price. Because capital markets, because the stock market has rewarded them so kindly with massive surge in stock price, they’re price insensitive. They don’t care that the price went up by 10% to build out the AI data center. And I think that’s one of the big impacts. The amount of energy and power that these AI data centers are consuming is power consumption went up very linear over the last couple of decades. This is an exponential growth in terms of power demand. And yes, you can argue that the AI data centers are a big cause of that. The biggest cause of it is that we didn’t build power centers, utilities, or pipelines across the United States or transmission lines. We were stunting in terms of we just underinvested. And I think that’s— we’re having to play catch-up to that, Blake.

Speaker 2 [12:18] That’s right. That’s right. And I think we’re starting to see now, you know, a resurgence in coal growth as well, you know, moving across and powering these facilities.

Speaker 1 [12:26] So what are shippers doing about this? I mean, this is something they’re going to have to live with. How do they handle it?

Speaker 3 [12:31] Yeah.

Speaker 2 [12:31] So there’s really, I mean, 3 things that I always think about. The first is to be, you know, so this is what I would say. This is a capacity tax, right? So it’s not— it’s supply constrained. It’s not a tax that’s because, hey, the market is good and rates are up because people are buying and companies, companies are selling more, people are buying. That’s not the case. Rates are up because, you know, just overall supply is tight. There’s fewer drivers. We’ve had driver, uh, again, through driver enforcement, we’ve lost drivers, about, about a quarter of a million drivers since 2020. We’ve lost drivers due to, um, due to retirement. Uh, but, but we’ve definitely— this is a, a capacity tax that they’re they’re paying, uh, in order to move the freight. So understanding that tax, um, the first thing, don’t rely on those headline indices. Stop using— completely discard the national averages. It’s no longer applicable. Uh, use, you know, look at your specific lanes, look at your, uh, your regional averages, understanding your market and your freight type. But definitely discard the national average. Have conversations with the carriers. Unfortunately, rates and contracts negotiated a year ago, they may be signed on paper Um, you need to make sure that those rates still support the capacity commitments, right? So we know, uh, carriers, especially the small independent carriers, they’re bleeding, uh, bleeding OpEx, uh, fuel surcharges, uh, that’s, that’s, that’s hurting, um, that’s hurting in this environment. Um, but they are, you know, dispatchers and, and, uh, the carriers, they’re definitely being selective with freight. So make sure that your, your current rates and rates coming up in the future those support those capacity commitments, and then build out a tiered backup plan, carrier backup plan. The third opportunity—

Speaker 1 [14:19] go ahead.

Speaker 2 [14:21] I’d say the third opportunity is really to evaluate modal arbitrage. We still have rail— rail transportation still offers a compelling opportunity when you look at transloads, but understand that by lane where you have opportunities. But rail rates are up roughly 2%, capacity sitting around 70% capacity utilization, sitting around 70%. So rail still is a very cost, uh, advantageous opportunity. It’s just a matter of going through, going through the exercise and understanding where you can move, uh, through, through transload systems. And then fourth, it’s kind of the, the final piece of it, Craig, is really it gets back to AI. Everybody’s talking about AI and how they’re going to use it into their business. So AI is kind of the lead in this, but it’s also the laggard because it’s causing the problem. But as, as, as the shippers are looking at the various AI opportunities and technology that’s moving towards them and being offered to them, really understanding where their data is and in what state the data is and understanding that data strategy. Because, you know, in this market, data is going to be everything. And it truly is. I think it’s the race to the underlying data to help with this problem. And it’s not the race to the AI algorithm.

Speaker 1 [15:39] Yeah, well, Blake, so thanks for coming in. Great points on rail, and particularly you see it in the intermodal market, but bulk is also really attractive. And we’re going to get into that tomorrow. We have the Future of Rail Symposium.