John Igo: There’s a reason a lot of things in supply chain happen just in time and just in the right quantity that matters, is because it’s efficient. If there’s a big-time lag between transmission and generation and load, that puts upward pressure on rates everywhere.
The dynamics of the utility space right now is just off the charts, and flexibility has never been more important. I think co-ops are in an incredible position as emperors of their system between transmission distribution and capacity to say, how can we co-optimize the least cost system, so we keep up with this new load growth, that we continue to manage this high renewables penetration and do it in least cost and eventually decarbonized way.
Teri Viswanath: That’s John Igo, the vice president of business development at Mainspring Energy, and he’s talking about re-envisioning the grid. The emergence of mega-loads, coupled with greater renewable supply, necessitates the need for flexible generation. But just where will we find that flexibility?
Hello, I’m Teri Viswanath, your co-host of Power Plays and the energy economist here at CoBank. As always, I’m joined by my colleagues, Tamra Reynolds, a managing director here at the bank. Hello Tamra.
Tamra Reynolds: Hey, Teri. Let’s be clear, in this rising cost environment, no one wants to pay for excess capacity that’s not going to be immediately utilized. We’ve spent time recently considering what a “no regrets” decision might look like from a system planning perspective as co-ops stay the course on their dual mission of affordability and reliability. And our speakers today, John Igo from Mainspring, and a policy and economic expert from Wärtsilä, David Millar, suggest that today’s energy markets require a deeper conversation on smaller, modular, flexible and decentralized power plants. So, that’s the conversation that we are going to engage in.
Viswanath: Our technology conversation begins with David Millar from Wärtsilä. Wärtsilä is a well-established, widely respected manufacturer of power generation technologies, specifically reciprocating internal combustion engines that have a proven track record for timely grid-balancing. These engines are capable of starting and stopping quickly, adjusting power rapidly, and operating at various loads.
Reynolds: David begins our discussion by taking stock of the current market conditions that have prompted our collective industry discussion on power supply and the possible need for new solutions.
David Millar: There’s no doubt that we have resource adequacy challenges in this country, and you can see that in the NERC report. There’s a few reasons why. One is we have a lot of renewables that are coming on the system. In fact, 95% of the projects in the interconnection queues are renewables and storage, but that actually has the impact in energy markets of depressing prices and adding more volatility. That’s a pretty challenging environment for legacy baseload generation.
On top of that, we are finally seeing load growth for the first time really in a couple decades. That is a fundamentally new world. This of course, is driven by data centers and electrification of different types of industry.
Then the third piece is that we’ve seen more and more extreme weather events, like Winter Storm Uri, Winter Storm Elliott. We’re seeing through those events that a lot of our thermal fleet actually has some reliability issues and have lots of forced outages, probably more than we thought and more than we had accredited those capacities to deliver.
We see, basically, there’s more fragility in the system. It highlights an urgent need to reinvest in our power supply. That’s what you see, I think, in PJM. Especially in PJM is where Northern Virginia is, and that’s really the epicenter of where the data centers are all trying to get built.
Viswanath: It’s certainly showing up in price. When you think about the capacity auction that just finished, the auction produced a price of $269 per megawatt day. This compares to what was $28 a year ago for the 2024/2025 auction. We are seeing prices respond to a market that is now apparently short. Are these price signals sufficient to see a power supply response in your view?
Millar: Well, we’ll see. Today, I think the power prices in terms of just the absolute prices in wholesale markets, between renewables and very, very cheap gas, the consumer has benefited from pretty cheap wholesale power prices for a long time. There really hasn’t been much incentive from the capacity side or the energy side to build. We really haven’t built much new generation in a long time. All of a sudden, as an industry, we got to get in gear very quickly and start to build.
I think there’s some initial signs that we’re going to have a response. I think if you are trying to buy a gas turbine today, it’s going to take you several years to even get the parts that you need to build. So we are really running into a supply chain issue there as well. Interconnection queues across the country are still jammed up, and a lot of projects in those queues will evaporate. It’s been a really difficult challenge to try to respond quickly. Frankly, the load is being built without any sign of abating.
Reynolds: David makes an excellent point: that a functioning market is one where if you show a high enough price for a long enough period, suppliers will respond. But there might be harder-to-move barriers that could prevent getting power supply to market. That backlog grew by 30% last year, with nearly 2,600 gigawatts of generation and storage capacity now actively seeking grid interconnection. This is more than double currently installed generation capacity. The last big build-out in transmission — for high voltage, long distance wires — occurred more than a decade ago. And this is a problem if the goal is to transition the grid efficiently and reliably. Here’s what David had to say.
Millar: Yes, there’s no doubt that we need to build a lot more transmission in this country, and a typical new transmission project is going to take more than a decade. If just looking at past history, it’s an extremely complicated process that involves a lot of local land, permitting decisions, and then generally a lot of back and forth about which states benefit and which states have to pay.
It’s super complex. There’s a new FERC regulation Order 1920, which is supposed to provide guidance and try to facilitate greater regional planning and delivery of transmission, especially we need very high voltage superhighways of lines, that cross multiple states in order to deliver things like remote wind or large solar sources from where the resource is to where the population centers are. It’s just a fundamentally different power paradigm than when you just have many fewer, but very large generation sources.
There’s no doubt that we could be doubling the size of the number of miles of transmission or to enable the kind of renewables we’re talking about in a decarbonized future. There’s definitely a challenge, and it’s certainly one of the things that’s going to slow the deployment of renewables. In the interim, we still need to maintain resource efficacy and reliability when we have a lot of power generation stations that should be retired. We call them “zombie plants” if they have to keep going, keep going and keep going just to maintain reliability. In general, we’re trying to figure out how to let those zombie power plants finally retire.
Viswanath: “Zombie plants.” This is a funny expression that David uses to talk about higher cost resources that could be economically retired (similar to the natural gas switching that happened in the early 2010s with coal plants) but simply can’t be taken out of commission because the “highway” can’t get built.
Reynolds: A confluence of factors currently are creating a gridlock in this country. Utilities and regulators are overwhelmed working to modernize and decarbonize the grid, while managing queues of generators and new loads seeking interconnection. This may lead to suboptimal outcomes if grid decision makers only see limited near-term options such as delaying new large load interconnections and/or delaying retirement of existing fossil fuel generators.
Viswanath: And large consumers such as data centers are taking it upon themselves to find their own solutions. My colleague Jeff Johnston and I recently published analysis on some of these novel solutions. One of which is that tech companies are increasingly looking to directly connect data centers to nuclear plants. Three Mile Island, known for being the site of the worst U.S. nuclear energy accident, is now at the forefront of efforts to expand nuclear capacity to meet rising electricity demand. But what are other options? David expands the discussion to include modular, flexible and distributed power supply resources.
Millar: There’s a lot of challenges with trying to take off some of our legacy thermal fleet, and especially in the face of large growth. A lot of utilities, we look at their IRP plans and they have this default move towards big gas combustion turbines, combined cycles, or they’re just going to run their existing coal, their existing steam generators, that were built in the ‘60s, ‘70s, ‘80s. They’re just going to keep riding those for as long as they can, right? We need to actually have a better, more flexible thermal fleet support, renewables, and storage as well.
Today, we have commercially solar and wind, and not everywhere in the country has great solar and wind resources. Batteries are a great technology for supporting renewables, but commercially and economically, we’re really only looking at about four hours of duration of battery storage. We still have to cover the winter, we still have to cover nighttime. We’re fundamentally going to need thermal generation for a long time. The question is what is the kind of thermal generation assets we want to invest in to support more and more renewables? The answer to that question is you want to have highly, highly flexible generation.
With reciprocating engines, they’re like a car engine. They can turn on and turn off, on a dime. It can go from zero to full in less than five minutes, and you can turn them on and off really without any penalties. It’s a perfect complement to a more volatile power system that’s driven by variable renewable resources.
They’re also small and modular. They’re typically in 10 or 20 megawatt engines and you build them serially. You don’t have to take the whole plant down for maintenance. You can just take one engine at a time. Also, there’s just an effect of when you build more, distribute it more modularly, you tend to get a lot of learnings from deploying one thing many times at scale. That means that when you build your projects, they very rarely go over budget and you can actually keep costs down.
That’s one reason why large nuclear plants, while they benefit from economies of scale, they have inherently a huge amount of risk because they’re so complicated and they take so long to construct. That’s what happened with the Vogtle Plant. It was supposed to cost $7 billion, then it was $14 billion, then it was $30 billion because of various factors. One of which is it takes so long to build those things and you don’t get any power until it’s totally complete.
Reynolds: In addition to cost, there is some concern about fuel dependency. Let’s be clear, a significant amount of natural gas-fired generation was built in the early 2000s and may not be able to operate in the next decade because of emissions concerns.
Viswanath: Both of the technologies that we are exploring in today’s discussion have fuel flexibility here, as David explains.
Millar: I will say that from Wärtsilä’s perspective, we’re agnostic and we’re neutral. We basically create an engine technology that will take any fuel. We recently announced that we have a 100% hydrogen-capable engine that’s going to be available in 2026. We can already burn ammonia and methanol, which are leading candidates to decarbonize the shipping industry because those are easier to keep as liquid fuels. Those would be green hydrogen-derived synthetic fuels. We can already burn those today.
Reynolds: Reciprocating engine technology is proven and has a very long history, finding traction in fact dating back to the late 1800s. But today’s energy needs are great. We wanted to explore other resources that can fill the gap as we search for flexible, reliable generation.
Viswanath: For this conversation we turned to John Igo at Mainspring. Mainspring was founded more than a decade earlier by three Stanford engineers seeking a new approach to generating clean, resilient, affordable electricity. There was a terrific Bloomberg article entitled, “How to Sell a Power Generator No One Has Heard Of” and it talks about the genesis of this company. Most of the technologies powering the clean energy transition—the solar cells, wind turbines, batteries and fuel cells—have been commercially available in one form or another for decades, and now they’re really taking off. Not so with the linear generator, which Mainspring started deploying in 2020. But John does a terrific job with getting us up to speed.
Igo: Mainspring’s commercialized a new category of power generation products called the linear generator. The linear generator at the highest level is just a very simple way of converting a clean reaction of air and fuel directly into high-efficiency, low-emissions power.
We leverage this kind of window between the software world and the hardware world that is power electronics to control a reaction of air and fuel on a very fast time scale and directly convert that reaction into power.
By leveraging this flexibility afforded us by power electronics in this way, we get these really interesting and important dimensions of flexibility. That’s the flexibility in terms of ramp rate and duty cycle. We can ramp up and down just as fast as a battery can,
We can also run a peaking duty cycle like a traditional peaker or we can also run prime or base load.
The second dimension of flexibility is the one that I get the most excited to go to work for every day is our fuel flexibility. We can switch fuels on the fly, leveraging these power electronics for both fuel redundancy, being able to switch over to a backup fuel in the case the primary fuel gets interrupted or is unavailable.
Another dimension of flexibility is scale. We can scale to effectively any project size. These are containerized and modular linear generators.
Then you can also, with this modularity, customize the level of availability and redundancy the system needs.
Then the last two are around siting and location. We have incredibly low emissions, categorically different than traditional thermal generation. Because of these very low emissions, you can permit our generators closer to load.
Lastly, because they’re containerized, getting to the siting and location perspective, we can move these things. If you picked a place to put these, maybe you put them next to a data center, and that data center goes away someday, you can move these containers across your system, and make sure that they’re in the right place as you move forward to ensure they’re not stranded.
If it’s okay, Teri, I got to use a term you used on one of your previous podcasts. These dimensions of flexibility that we think affords co-ops and their members the luxury of that no-regrets, right-size capacity investment.
Viswanath: I want to unpack a couple of the comments you made, so for folks that are trying to get their head around this technology. Compared to a peaking gas unit, how does it compare?
Igo: Take a traditional gas peaker. Let’s just pretend it’s a 150-megawatt frame machine as a starting point. Then slice it into 600 pieces of modularity that can be used to ensure that the power plant is always on, even if a small portion of that power plant is offline. That’s one thing that we can do that a peaker cannot do. If we need to bring one of our modular units down for maintenance, the rest of the plant stays up. You’re never losing 150 megawatts at one time. That just will not fundamentally happen.
Then take the efficiency of that plant and increase it to 45% efficiency. That’s on an LHV basis, so at 8,400 BTU per kilowatt hour, higher heating value heat rate, very high-efficiency product. Then know that no matter what load point you set at that 150-megawatt project, you’re getting that level of efficiency because you’re just turning off modular nodes at a time to run that plant.
Now, make it as fast-ramping as a battery. There’s no start time, there’s no thermal inertia to spin up. These units can turn on instantaneously like a battery. They can participate in all the fast-ramp ancillary services markets across the country. We have the flexibility to do that. Now, make it truly hydrogen-ready. Now, make it truly ammonia-ready. Now, say, whatever fuel you want to put in there, go ahead and put the fuel in there and run the project.
Then maybe one more thing. Hey, we need 20 megawatts of this project. Go bring up a load over a new data center over here. Let’s just take part of this plant and move it to where it’s needed on our system.
Viswanath: Let’s talk about the current marketplace and where you see application for this technology.
Igo: I’m just going to start from smaller use cases and go to the larger ones. We started behind the meter at grocery stores and other C&I—commercial and industrial—loads. This helped them not only reduce the retail energy costs but also manage their retail energy cost risk because they have more certainty. And they had the ability to ride their grid outage.
Then we have a great use case for Mainspring. I wish there was infinitely abundant biogas on planet Earth for us to run off because biogas is a great use case for us, but we run prime and base load off of biogas coming from landfills, wastewater treatment plants and dairies
Then we have microgrids. We’ve got microgrids at logistics hubs. We’re affirming rooftop solar at large logistics operations. We are powering fully islanded EV microgrids for the likes of Prologis. They’ve talked a lot about a project we’re both very proud of, which was their ability to bring on and electrify their drayage fleet years before it would have been possible with the electric utility. It’s not the electric utility’s fault necessarily. It’s in Southern California where it’s really hard to build out that infrastructure.
Then the larger scale version of that islanded microgrid is bridging power for data centers. We’re in the middle of with data center customers partnering on what the design looks like for a three or a five, nine installation in a place where they can’t get power for, sometimes, 10 years.
Then on the utility muni co-op side, we got a non-wires alternative going on with AEP in Oklahoma, more traditional grid scale opportunities that look like large battery energy storage projects that transmission voltage. A large one that we have currently being developed as a 50-megawatt project we’re building on behalf of P-ESCO in Colorado, where we were chosen as part of the least-cost portfolio for their last large round of procurement.
Something else we’re really proud of is that we beat out a traditional peaker at the same location because of our ability to provide that least cost portfolio given the constraints in Colorado.
Then I think one of the use cases most applicable to this audience is scalable, strategically sited capacity, particularly at distribution voltage, and this being for co-ops and munis as well, but especially for co-ops. Why distribution voltage? There’s a couple of reasons. One is capacity can be brought online faster because 12 and a half and 34.5 kV transformers aren’t as long lead as 69 kV and higher. If you need capacity before three or four years from now, you’re not waiting that long. The long lead equipment’s about a year. You can build a project and bring that capacity online.
We have a lot of customers that are looking at their battery installations as part of their least-cost portfolios, and saying, “We need those batteries, but they aren’t capacity. They cannot run through the multi-day heat wave.”
Just want to make sure it’s clear, even though we can flex and ramp fast like a battery, we can provide the long duration capacity to run through that scarcity event and you can keep the lights on during the next crazy thing that happens.
Reynolds: As John highlights, there’s a reason a lot of things in supply chain happen just in time and just in the right quantity that matters. That’s because it’s efficient. If there’s a big-time lag between transmission and generation and load, that puts upward pressure on rates everywhere. That’s what’s happening in PJM. That’s what’s going to happen in other markets because what we need is for capacity to keep up with the load.
Viswanath: Flexibility has never been more important and these new options might prove to be a very important piece of the puzzle.
Reynolds: I hope all of you have enjoyed today’s technology conversation, and will join us as we wrap up the year next month, speaking to the energy journalists covering this year’s major stories and the ideas they share about what the next year might hold.
Viswanath: Thank you for listening and goodbye for now.
