WEBVTT 1 00:00:09.000 --> 00:01:32.000 Welcome back to Powering Progress, a podcast series exploring the infrastructure, energy, and market dynamics shaping the next generation of data centers and AI development. I'm your host, Aubrey Bishai, Chief Innovation Officer at Vinson & Elkins. As AI and data center development accelerate, the competition for power, infrastructure, and emerging technologies is intensifying, just as regulators are taking a closer look at consolidation across these markets. In this episode, I'm joined by Evan Miller, a partner in our antitrust practice in Washington, DC, who advises clients on merger investigations and antitrust counseling across energy, technology, and infrastructure markets. We discuss why antitrust analysis in power markets can differ from traditional transactions, and what companies should expect as sectors adjacent to infrastructure — from cooling technologies to energy storage solutions — continue to evolve and potentially consolidate. Welcome, Evan, and thank you for joining. 2 00:01:32.000 --> 00:01:33.000 Thanks for having me. 3 00:01:33.000 --> 00:01:45.000 So what makes merger reviews involving generating assets and power infrastructure different from more traditional M&A transactions? 4 00:01:45.000 --> 00:03:38.000 Yeah, it's a great question. So the Department of Justice, which has jurisdiction over power deals, uses a specific merger analysis for power markets, and it's different from what drives most M&A reviews. So in a typical deal, you have two manufacturers merging or two software companies combining, the DOJ might focus on market share. The basic question is, does this combination give the merged company enough share of the market that it can raise prices without losing customers? And that's a relatively straightforward analysis. The DOJ recognizes that the power markets work differently. And when they're reviewing a merger involving generating assets, the concern isn't just how big is this company now — it's about a withholding tactic. So most listeners of this podcast are probably already aware, but electricity markets in the United States use a clearing price system, right? So meaning every generator dispatched in a given period gets paid the price set by the most expensive unit, the last unit called to run to serve power to the market. If a company controls enough generation, it has an incentive to withhold some of its cheaper capacity from the market, which forces the system to call on more expensive units up the supply curve to meet that demand. That pushes the clearing price up, and the company that withheld capacity gets paid that higher price on everything it does sell. So the DOJ isn't just asking, what's the combined company's market share? They're asking, does the company now control enough capacity to manipulate the power supply curve? And that's a more technical, more market-structure-specific analysis. And so I think the practical implication for the deal makers that we work with a lot is that two relatively modest generation portfolios can trigger serious scrutiny from the DOJ if they are concentrated in the same load zone or control pivotal capacity during peak demand periods, even if the combined market share is quite low. And I think that surprises some people to hear that. 5 00:03:38.000 --> 00:04:01.000 We've been talking about on this podcast how it really feels like a race. We're seeing companies race to secure power capacity, cooling technologies, and other infrastructure advantages. At what point does strategic growth begin to trigger heightened antitrust concerns? 6 00:04:01.000 --> 00:05:26.000 It's really kind of the topic of the moment. And the answer, of course, like most things, is that it really depends, right? So for hyperscalers, similar operators that are competing in this race — building new power capacity, signing long-term power purchase agreements, investing in efficiency — that all is generally fine, and actually quite pro-competitive, and the DOJ recognizes it as such. The concern really arises when companies start controlling limited resources in ways that foreclose competitors from the market. So if one company, for example, signs a non-exclusive agreement for renewable power off a new wind farm or something like that, that's a normal commercial agreement. But if that company signs agreements with every available provider in a constrained grid region — not because they need all that power right now, but as a strategic move to prevent rivals from accessing it — that starts looking like what we in the antitrust bar call foreclosure. At that point, you're not competing on the merits. You're competing by denying others the inputs they need to compete at all. Those are obviously examples at opposite ends of the spectrum, right? The practical takeaway for companies in this space is to be mindful of their cumulative control over power capacity or other essential inputs within a specific region. The fact that developers — thinking ahead for arguments in defense of this — the fact that developers have flexibility in most cases to build in other regions that may not be as power constrained will often be an available argument to address this concern. But this is an area where I can see the agency keeping an eye on. 7 00:05:26.000 --> 00:05:43.000 As adjacent industries consolidate, I'm thinking about cooling, energy optimization, AI enablement technologies. What types of competitive concerns are regulators likely to focus on most closely? 8 00:05:43.000 --> 00:07:47.000 I think this is kind of related to what we just talked about as well. We talked about the potential risks of locking up power supply or other essential inputs, but there is a second story that could happen in parallel, and that's consolidation across the adjacent technologies that you just mentioned. And these are the technologies that make AI infrastructure actually run, right? So cooling systems, power management software, all the things you just mentioned. And this is where some vertical integration concerns might creep in from the DOJ. And what I'm talking about here is when a large player in AI infrastructure acquires a company that makes essential cooling technology, the antitrust question might shift a bit. It's no longer about whether you're cornering the supply of an unaffiliated input like power. It's about whether you can now use ownership of a critical technology to disadvantage rivals who depend on it. Can you degrade their access? Can you slow down their service? Can you price them out? Those are the theories of harm that the DOJ would be looking at. But the concern that I could see driving the most investigation risk in the short term is the so-called killer acquisitions. And the classic example of this is when an industry leader buys an emerging competitor not to develop the technology, but to shelve it, to make sure it never matures into a real competitive threat. And we saw this theory gain traction in review of pharmaceutical and big tech deals, and the logic maps well onto the AI infrastructure space. So you can think about an industry incumbent acquiring an early-stage company that has developed a differentiated cooling approach or a novel energy optimization platform. If that technology, left independent, could have lowered the cost curve for the whole industry, and in doing so made it easier for smaller players to compete, the acquisition by a competitor could face significant antitrust scrutiny. And the other dimension that I'd flag here is also data, right? A lot of these energy optimization and AI companies have accumulated extraordinary data sets about grid performance, load patterns, efficiencies within the infrastructure, and regulators are increasingly treating data advantages as a competition concern. So that's another area that people look out for. 9 00:07:47.000 --> 00:08:03.000 Evan, you work extensively on emerging technology issues. Do you expect AI infrastructure and energy availability to become central themes in antitrust enforcement over the next several years? 10 00:08:03.000 --> 00:09:57.000 Without question. And I think at this point, we're past speculating on this, because we have a handful of data points. Going back to the Biden administration, Jonathan Cantor, who led the DOJ antitrust division then, has identified choke points in AI infrastructure as a central antitrust enforcement priority of the future. Then moving more into the present, Gail Slater, who led the antitrust division under President Trump until just recently, also emphasized the importance of effective antitrust enforcement in the power sector, specifically to support domestic AI growth. So for me, this means that interest in how AI infrastructure markets develop, especially around electricity, is not a feature of one administration's priority, but likely a fixture of the industry for the next five, ten years or more. And this isn't just talk from a couple of former regulators. There's actual enforcement action to back this up, right? So going back fairly recently, in December of 2025, the DOJ challenged Constellation Energy's acquisition of Calpine. And this was the first time in nearly 15 years that the DOJ sought structural relief in the form of divestitures in a merger involving electric generation assets. And the theory of harm was the same withholding concern that we discussed earlier in the podcast. The DOJ's complaint focused on how the merger would increase Constellation's ability and incentive to raise wholesale electricity prices through strategic withholding. And critically, and again, to circle back on the point that we made at the beginning, this theory was advanced even though the combined company would hold well below the kind of 30% market share threshold referenced in the merger guidelines as presumptively anti-competitive. So I think, again, the takeaway here is that for anyone building in the space, administrations from both parties with very different antitrust philosophies have landed in the same place on power and AI infrastructure. And while it's likely true that the antitrust risks will grow with the market, and that will take some time to develop, it's certainly not an area of focus that is likely to go away anytime soon. 11 00:09:57.000 --> 00:10:22.000 Thank you, Evan, and thanks everyone for listening to Powering Progress from Vinson & Elkins. For more insights on AI infrastructure and the forces shaping the data center economy, visit velaw.com.