Teri Viswanath: America’s running into a weird power problem. We’re not short on natural gas; we’re short on the turbines that turn it into electricity. The big manufacturers are essentially sold out until the early 2030s. Even after utilities place their equipment orders, it can take years to build the power plant — meaning new power might not show up until a decade from now.
Tamra Reynolds: Today, my co-host, Teri Viswanath, and I ask, what happens to reliability and electricity bills when the fastest, most familiar way to add power can’t arrive on time?
Hello, I’m Tamra Reynolds, a managing director at CoBank, and we are joined by a respected expert on this topic, Tom Freeman. Tom is a former GE Vernova gas turbine executive turned gas turbine coach who offers us a rare inside view of what’s driving today’s turbine bottlenecks and what it means for the grid.
Viswanath: Why should our audience really take an interest in understanding what the upstream manufacturing sector looks like for gas turbines right now? Why is that critical?
Tom Freeman: We want things to be affordable, and we want them to be reliable. Gas turbines have been a big part of that. That’s what really fascinates me about this whole situation we’re in now. It’s 180 degrees from what we were saying, let’s say 2018 or so. The first swing at that was Saudi Vision 2030, where the king made a move to upgrade his country in a lot of areas, but in the energy sector, what it ultimately did was take production line volume, a lot of it, from the big OEMs.
The second thing that happened is this AI boom has hit. It’s really not just the AI boom. It’s really reshoring, bringing manufacturing back home. We’re heading into another bubble. There’s been three bubbles in the industry; 1965 to ‘75 was the first bubble. It followed the initiation of industrial gas turbines in 1949 as a success at Belle Isle, Oklahoma. Then, in 1980, natural gas deregulation in the United States.
After that, ‘95 through 2005, another 10-year bubble. We were making a crazy number of units in Greenville, where I was living at the time. Now we’re entering into another one. Is it going to be 10 years long? It looks like it might be more like 15. Here’s what’s challenging about that. If you can divide the gas turbine industry up, you’ve got new units that are being built. You’ve also got existing units that are now 20, 30 years old that are coming up for renewals of equipment. You’ve got two there.
You’ve got a third one. You’ve got aerospace. Those are gas turbines. You have three portions peaking at the same time. Raw material rise, we have a problem. Casting and forging houses are overrun. Then craft labor, we don’t have the people. All of the big OEMs are busy trying to figure out a way to pump more through their factories. It’s super challenging.
Reynolds: How should people be thinking about how this stacks up and what that actually looks like from start to finish? How can they be penciling that out to see what real time would be to get something from an initial decision, all the way to actually running and operating?
Freeman: GE and Siemens and MHI, they’re the big three. Those guys are really selling 2031, 2032 now. Now, one of the EPCs construct these sites; those guys are saying, “Hey, it used to take us 12, 18 months to do this. It’s now going to take us three to four years.” It’s not going to make a kilowatt until, let’s say, 2034.
Viswanath: You mentioned the big three. Two-thirds of global turbines are going to be manufactured by either GE Vernova, Siemens, or Mitsubishi. If we just look at the largest of that is your former employer, which is GE. We take a look at the numbers from last year. Their backlog at year-end 2025 was 80 gigawatts. Their manufacturing capacity expectation is 24 gigawatts in the next two years. We’re talking about multiples of the existing manufacturing capacity. How long does it take for us to consider expanding the manufacturing capacity for those existing plants?
Freeman: GE and Siemens have both announced massive, massive expansions to their manufacturing, huge amounts of money. I would say that at least 50% to 60% more units will get through those factories, but that’s going to take some time. It’s not going to just turn on a dime here. The units coming out of the factory will probably go from some number that maybe is 80-ish to some number that’s like 130-ish type of number, but it’ll be like that kind of jump. But we won’t see that until 2028 or 2029.
Reynolds: Hey, just to put a few numbers behind this, GE is investing nearly $600 million to expand U.S. manufacturing, and they’re aiming for 20 gigawatts of annualized heavy-duty gas turbine production this year. Siemens will invest about $1 billion to ramp up manufacturing with plans to expand their global production by around one-fifth. Mitsubishi Power is committed to doubling their production, but I’m not really clear on the details here. They’ve already sold out into 2028. To be clear, even if all three delivered to plan, total output might only rise by 25% before 2030, and that’s nowhere near enough to meet the demand we have.
Freeman: The other thing that’s complicated about this is, in the last bubble, GE was making pretty much 7Fs, and Siemens was making their version of that called the 501F. Those two, they were making pretty much one thing. It’s the diversity of need in this bubble is different. They’re having to make a certain size of unit for this manufacturing, this new manufacturing, the reshoring. That’s a smaller unit. They’re having to make units for the LNG industry, which is booming. That’s a mid-range unit. Then there’s the bigger things.
Viswanath: Right, because this is made up of a diverse set of products. We have those numbers, but it really depends on the application.
Freeman: That’s right. I think everybody’s running the same strategy or same playbook that they used in the last bubble, for the most part, but it’s with a different situation. It’s a huge challenge for them. Then we’re only talking about the gas turbine. The gas turbine doesn’t do any good without a generator. Turning that spinning shaft into electricity, and it doesn’t do any good further if there’s no transformer.
Viswanath: That’s a really important point. The turbine shortage is not just an isolated equipment problem. It exposes the fragility of the entire industrial stack: the transformers, EPC labor, switch gear. All of these components depend on long, brittle supply chains. Turbines just happen to be where the pressure shows up first.
Freeman: There’s a lot of hurdles here to get through this. A lot of creativity. Not just people trying to upgrade existing units. The other thing that’s happening is refurbished units have become big. If you can go find refurbished units elsewhere in the world, whatever, and bring them to the U.S., flip them from 50 hertz to 60 hertz, that’s a big deal because there’s no way to get units from the OEMs.
Reynolds: Tom, I think one of the things that struck me about the conversation that I had a chance to sit in back in September with you at the member advisory group meeting within NRECA was fascinating to hear some of the evolutions and how people are thinking about some of the technology that exists. Maybe we can talk a little bit more about aeroderivatives. Any thoughts on some of those things that you’re seeing that are a little bit more unique and creative that are coming from utilities with some of the work that you’re doing?
Freeman: I’m glad you mentioned the aeroderivative piece. I feel like that’s going to be a key backbone here. There’s a couple of companies that are trying to take CFM56 core engines, which would be, when you get on a 737, that’s usually what they are. There’s some companies that came in and said, “Hey, what if we took the CFM56 because they’re retiring some of these units and we could give them a second life on the ground?” It’s not a short put at all. That’s a way to go out and find, within the bigger ecosystem, a different path to power that was not open.
I think that you’ll see data centers are going to gravitate toward an aeroderivative starting point, especially mobile aeroderivative units, and then transition them over to medium-range units. I think we’re going to see some of the smaller units play in here as we progress here in the infrastructure. There’s a company called ProEnergy that was taking units off the wing, and they’re basically making the GE LM6000. The other thing that’s happened is Siemens, who had walked away from aeroderivatives, has now jumped back in. That should tell you something right there. Then Mitsubishi already has the FT8s.
Viswanath: We’re seeing The Wall Street Journal article, how jet engines are powering data centers, which really speaks to this moment that we’re living through, that we really need an all-of-the-above solution. Is this a bridge, or is it a bridge based on the reality of where we’re at, the inability to turn on a dime, and the fact that we have a resource crisis? Am I going to have to, down the road, then get back to a different solution?
Freeman: I think one of the things people are thinking is, if I can have temporary power that’s mobile, that I can roll away and start another data center, I can do that while I wait for these units that are taking so long to come out. I think one of the things that happens with data centers is we hear reports that says, “Oh, we’re going to build this 5-gigawatt data center,” and we assume that we have to have 5 gigawatts on day 1, or we assume that you’ve got to have 5 gigawatts by year two. That’s not how it’s going to go.
That’s why I like the aeroderivative piece as a pathway. The reciprocating internal combustion engines, called RICE engines, have some appeal. It’s smaller power, but it can carry some of the load, and they’re fast, of course. They’re pretty quick compared to the other stuff. Here’s the thing, though, and this is the question mark. It’s one thing to be hanging on the wing, and then when that engine needs maintenance, you just drop it off the wing and you throw in another unit.
That’s not done quite the same way in the energy industry. Those CFM56 units have to be industrialized so they can last like we want them to last. We want them to turn on and go for four years. We don’t want to shut them off. We cannot have a bunch of forced outages. I’d say the same thing on RICE. RICE engines have to run differently. This is not Formula One. We’re not running one race and then changing out the engine. That’s going to go on for four years. We’re going to have to learn fast to try to update some of the things we’re doing.
Viswanath: The problem is that this shift is creating a shadow power grid that brings severe climate, health and logistical risks, right? One of the most visible examples of this is unfolding right now in Memphis, Tennessee, where more than 30 natural gas turbines are intended for daily use at the new Colossus data center. It is the industrial version of what homeowners might do when we get through a hurricane. Only in this case, some technology companies are planning to rely on off-grid gas for many years, not just through that emergency.
This also brings to light there’s a lot of problems because this gear is a lot less efficient than the equipment that we would otherwise bring to power big plants, meaning that we’re going to have to burn a lot more natural gas to generate the same amount of electricity, which is not great.
Reynolds: The problem used to be much easier to solve, right? You had simple cycle turbines. You had combined cycle turbines. You had reciprocating engines. From a cost standpoint, there used to be a pretty clear delineation between those different units and how efficient they were, and how you thought about least cost power dispatching. Today, that feels and looks different when we think about the need and how we’re going to have to solve the challenge of having enough power to meet demand. Maybe the turbine that they end up with and the ones that are in a backlog are twice as expensive or three times as expensive as they were in, say, 2006, 2007, 2008.
Freeman: The prices right now from 2020 to 2026, they’re over 3X. The numbers are just mind-boggling, what people are having to pay for these right now. The OEMs are requiring non-refundable reservation fees. Not only has the price gone up, but the percent that they’re asking for has gone up. It’s an unbelievable amount of money. I feel like we’re getting ourselves into a place where, financially, people are just going to have to wait this out because these prices are demand-driven.
Yes, there are some things that are costing more money now. True, true, true, but not 3X. When people are looking at units that they can get on the gray market or refurbished or whatever, what they’re doing right now is they’re sitting back waiting for the project to align before they buy the equipment. The danger in that is the prices keep going up and the number of units that we available are going to diminish and diminish. I figure there’s about a four-year run on refurbished units. I think it has to flip. Otherwise, it’s going to cost a fortune.
Viswanath: We know there’s a concentration in supply, but there’s also a concentration in where the turbines are being demanded. China and the U.S. represent about half of that order book. Then, even within the U.S., you’ve got a third of this occurring in Texas alone. Is it a possibility we could get ahead of our skis? We have a surplus of turbines on the market.
Freeman: Everybody’s waiting for that to come. I said those last bubbles were 10 years long. I think I’d be right in still saying that everybody believes that this is probably a bigger bubble than the last ones. This is probably a 15-year bubble.
Viswanath: Why do you think that this looks a little different?
Freeman: Part of it is because we can’t build the diversity of units fast enough. We’re trying to catch up on this stuff. It’s very, very difficult. I think it’s just going to stretch out just on production levels.
Viswanath: That last bubble, which was our last dash to gas that occurred in the early 2000s, we saw that rising natural gas prices really meant that those investments, we saw combined cycles that were operating at 30% capacity utilization, Tom, and it really damped down enthusiasm for development. I’m wondering, the consumer has a limited appetite for what they’re willing to take on with regard to rising electricity costs.
Freeman: I think the threat that we should talk about is making sure that if a data center comes in and produces this demand, that the residential consumer does not pay for all of these new plants. I want to make sure that the investments, that the residential part of this equation, doesn’t have to bear the weight of all this growth.
Reynolds: Tom leaves us with the warning: Make sure residential consumers aren’t left carrying the cost of the new buildout. Rising costs of strategic equipment could lead to stranded investments, passing the buck onto residential consumers if safeguards are not in place. In the early 1980s, the Washington public power supply system pursued a large nuclear buildout partly to meet forecasted industrial demand, but several plants were canceled mid-construction after cost rose and load projections weakened. The debt and other stranded costs didn’t vanish. Customers ultimately bore much of the burden through long-lived financial obligations and rates.
Viswanath: Tamra, that’s a really important point. We have been here before. I was also struck by the fact that Tom labels this accelerated upstream response by turbine manufacturers as a bubble. I recall the last dash to gas, which ultimately led to an excess of equipment that was left in warehouses for years. He predicts that because this buying spree is consumer-led rather than developer-led, utilities probably shouldn’t expect equipment discounts anytime soon.
That said, lining up all of the elements of a new natural gas combined cycle project is challenging, so timing other major purchases and readying the site will probably add to delays. There’s roughly 150 gigawatts of natural gas generation that is currently on the planning table, but these delays mean very little of this will come online in the next few years.
Reynolds: This was a terrific conversation, and I hope our audience comes away with three simple takeaways: Turbine supply is now a real constraint on adding power. Waiting can mean higher costs and longer timelines. As large industrial loads grow, we have to get the cost allocation right so residential members aren’t left holding the bag. Thanks for listening to Power Plays. We’ll see you next month.
