The Shift Back to Gas

The Shift Back to Gas

August 01, 2025

The federal government is making a concerted effort to promote construction of new power plants that burn fossil fuels. How much of a build out of new gas-fired power plants is it reasonable to expect in the next few years, given the current three- to five-year wait for new gas turbines and the risk for developers that a new administration four years from now will engineer another 180-degree shift in policy? 

A panel discussed this question at our 34th annual energy finance conference in San Diego in June. The following is an edited transcript.

The panelists are Declan Flanagan, CEO of Bluestar Energy Capital and former CEO of Ørsted Onshore and Lincoln Clean Energy, Joel Link, president of development at Tenaska, a leading developer of gas, renewables, carbon capture and hydrogen projects, Michael Pantelogianis, co-CEO of Apterra Infrastructure, the private capital lending arm of Apollo Global Management, and Jan Smutny-Jones, CEO of the Independent Energy Producers Association in California. The moderator is David Burton with Norton Rose Fulbright in New York.

Equipment Shortages

MR. BURTON: Joel Link, how long would it take to get a gas turbine if you place the order today?

MR. LINK: If you start today, I think you could possibly secure some slots for 2030 deliveries.

It is a pretty arduous task. There are only three vendors: Siemens, GE and Mitsubishi. We had to go through a vetting process for our equipment orders where they vetted all of our projects line by line. We spent days discussing what our prospects were. The script has flipped. There are only about 120 to 130 of the advanced class turbines available globally on an annual basis. There are not a lot of turbines to go around. A couple years ago, the Saudis ordered something like 30 in one order. There is a lot of global demand.

The advanced class turbines are 450 megawatts of output for a single combustion turbine. A one-on-one combined-cycle power plant with a combustion turbine, HRSG and steam turbine gives you about 750 megawatts of output. You get a lot of capacity for that configuration.

Technologies change. Supply is scarce. If you are willing to take  F-class turbines, which were the technology of the early 2000s, you could secure some of those for delivery in 2028 or 2029. Those are used mainly for simple-cycle peaking applications.

Steam turbines are becoming a long lead-time item as well. When we talk about buying combustion turbines, we really have to talk about buying the entire power island, which is the HRSG and steam turbine to integrate with the combustion turbine. That requires roughly a $25 million down payment per train.

MR. BURTON: For utility-scale solar and storage projects, the step-up transformer is the longest lead-time item. The last delivery schedule I saw for a step-up transformer was 18 months from placing the order. That is much more accelerated than for a gas turbine. Is that time period consistent with what others are seeing for main power transformers?

MR. LINK: I would like to know who those vendors are. We just purchased some 345-KV transformers, and it was more like 48 months. I think when you get into the lower voltages, you are possibly at 30 months as the best case.

Installed Costs

MR. BURTON: On average, what would you expect the all-in construction cost to be for a gas-fired power plant in dollars per kilowatt of installed capacity?

MR. FLANAGAN: Let's start with the math that Chris Seiple of Wood Mackenzie showed us yesterday, which I believe was $2,200 per kilowatt. There were two other numbers in Chris' slides: about $90 a MWh as a combined capacity and energy payment, assuming a little under $20 for energy. As Kevin Smith mentioned yesterday, you have to think of the capacity and energy separately for gas-fired generation.

Let's then do some simple math to take the $90 net off the energy cost at $20, so the $70. Take a utilization rate: I think a 60% capacity factor is what WoodMac would use. Run through the math and keep me honest. That is north of $30 a kilowatt month for capacity.

Storage contracts are being done in the low teens per kilowatt month for two- and four-hour storage. Whether the cost is $2,200 or $2,500 per kilowatt of capacity, there is no pulse in the pro forma at those numbers. It is off by two to three times.

I am a renewables guy, so I am here largely, as a German colleague of mine used to call it, as an “advocate of the devil,” as gas is needed in the transition.

The question is not the lead time, but how the installed cost got to this level and what policy needs to happen to get it back down below $1,000 per kilowatt of capacity. Of course, current policy is doing the opposite of what is needed, by trying to drive up the cost of batteries and leaving record capacity auction prices in PJM that appear to be politically untenable.

The conversation in this industry five-plus years ago was about the missing money problem. Now the costs have doubled or tripled, but we talk about lead times. The bigger issue is how to get these costs down. No one will pay the implied prices for electricity at these cost levels.

MR. BURTON: Joel Link, do you want to comment on the relative cost of gas versus solar?

MR. LINK: It is important to know the configuration when talking about gas-fired generation. If it is one combustion turbine, one HRSG and one steam turbine, the cost is $2,200 to $2,500 per kilowatt of installed capacity. That is in today's dollars. The cost by the time you build it four or five years from now will be higher by about 5% a year.

When you get into a two-on-one combined-cycle plant where you have two combustion turbines and an oversized steam turbine, you are talking about 1,500 megawatts of capacity. That is the ideal economy-of-scale configuration, but finding room on the grid to accommodate another 1,500 megawatts is difficult.

The vendors are pushing pretty much everyone to a one-on-one combined-cycle configuration because they do not want variability of design. They do not have the time to spend working on that.

You can do supplemental duct firing where you add natural gas to the exhaust of the combustion turbine to juice your steam production.   

The efficiency of the gas turbine depends on your elevation. The efficiency differs at sea level versus higher elevations.

Tenaska is a diversified company. We have a gas trading logistics business called Tenaska Marketing Ventures. We do not speculate on the price of gas, but we match up producers with end users of natural gas. We have positions in around 160 US pipelines and 50 billion cubic feet of gas storage capacity. We are the number one marketer. The reason I mention this is we have good visibility into the gas situation in the US. Gas is scarce, so not only are gas turbines and steam turbines scarce, but the pipelines and storage capacity to handle natural gas are scarce as well.

Obstacle Course

MR. BURTON: What other obstacles are you facing as you try to build more baseload generation using gas?    

MR. LINK: It is difficult to screen for available gas. You may find available gas, but there is usually some topographic or environmental issue or maybe you cannot obtain an air permit because the project is in a non-attainment zone for air quality. These are some of the challenges with finding suitable locations to build new gas-fired power plants.

My biggest concern is not whether more gas-fired power plants will be built. We need them to fill in gas between renewables generation.

Kevin Smith pointed out correctly yesterday that the starting price of $90 per MWh is based on a gas price today of $3 an MMBtu. When I set up a tolling agreement with a utility, I am pushing the gas-price risk on the utility. The utility is responsible for the price of the gas, because there is no way for a developer to take on that risk for a 20-year contract term.

My biggest concern is finding an EPC contractor who can build a new gas plant. Kiewit is the 800-pound gorilla. The others are Black & Veatch, Burns & McDonnell, Zachry, which is emerging out of bankruptcy, and MasTec. We have a labor shortage in the US. To build complex infrastructure, we need a bigger focus on the trade crafts and enabling more contractors.

The labor shortage is a serious constraint on the amount of new gas-fired power plants that can be built over the next decade, assuming the demand remains.

The potential for another 180-degree shift in policy four years from now creates a lot of uncertainty and a potential for whiplash like we are now seeing for renewables. How can anyone possibly plan for this?

MR. FLANAGAN: Joel's point is a key one on the labor shortage and all the non-turbine balance-of-plant stuff because we see this across technologies. We focus too much on the turbine or the solar panel.

Chris Seiple had data that I have been puzzled about for years why solar costs so much more to build in the US versus in Europe. We have 50% higher balance-of-system cost. Right away, every line item is higher. On the gas side, the data shows three times the cost on all of the non-turbine inputs.

There needs to be a long-term plan on driving down these costs and on building up the skill levels in the labor market. On the gas side, the projects require a much more complex multi-year construction period to build and have heavier reliance on trade skills that are in short supply.

You hear this from all of the contractors. With consolidation, the same contractors tend to be doing wind and gas. This administration seems uniquely ill-suited for that type of long-term, education-based strategy. That is really going to bite because the cost figures will only keep going up.

A $30-a-kilowatt month does not work for gas-fired power plants, and it is more like $50-a-kilowatt month.

We need an Apollo program to get the cost of electricity down, and we are doing exactly the opposite.

Electricity Prices

MR. BURTON: What does your math imply about the cost of electricity for offtakers who are buying firm power from gas-fired power plants?

MR. FLANAGAN: It is going to be a capacity payment and a fuel pass-through. I don't think it will work for under $100 a MWh. It is in that range. You compare that with the nuclear deal that was announced the other day with Constellation. The speculation is $80 a MWh.

I am reminded of the snappy line that Michael O'Sullivan, whom many of you know, used to use: “Soon we will be calling offshore wind the satellite phone of electricity.”   It is a very expensive niche product, and at these numbers, that is what new-build combined-cycle gas turbines are risking becoming.

It will happen for niche applications like data centers or in regulated uses.

The cost of capital and interest on construction debt are major line items as well. We saw $90 a MWh from Wood Mackenzie. I would not sign up to that because I think these projects need north of $100 a MWh.

MR. BURTON: Forty percent of data center demand is being met by gas-fired power plants according to the latest statistics, and the state of Georgia projects that its demand for electricity will be 17 times where it is today in a decade due to increased industrial use. Are the business plans of AI companies and energy-intensive factories anticipating the high cost of energy implied by this demand-supply imbalance? Can AI data centers really afford this energy?

MR. PANTELOGIANIS: The situation is more challenging that those numbers imply.

Three and four years ago, developers were talking about 50-MW to 100-MW data center facilities. Today the big guys are talking about 500-MW data centers. These facilities are coming up in 2026, 2027 and 2028. Tenaska can't get a gas turbine until 2030. How are we going to meet all of this demand in Georgia?

We have a ton of financing opportunities in the Atlanta region and in all the big markets. Forget Virginia, which is known to be the data center capital of the world. The planned capital expenditures across all markets are staggering. If I had to put my money on it, I would bet that we are going to have a huge problem.

How will that manifest itself?  Probably during peak periods when there will be more demand for electricity than we are able to provide.   If greenfield power plants take two and half to three years to get built and you cannot get a gas turbine until 2030, when are you going to be online?  In 2032 or 2033. I have data centers hoping to open in 2026 and 2027. How are we going to bridge this?

MR. LINK: That is exactly the current challenge. We have some merchant gas assets, so we dangle those in front of the data centers as a near-term opportunity to get power.

For example, we are in Loudon County, Virginia with a merchant gas plant. We won a resource reliability initiative deal in PJM where we were allowed to skip the interconnection queue for 1,500 MW in a market with a huge need. We also have a gas peaking plant just an hour southwest of Atlanta that has been contracted to Shell. We are planning an expansion there to four peakers. But one of the big issues in Georgia is lack of firm transportation for gas. We can get intermittent gas, but the pipelines cannot commit to firm deliveries.

Then you need fuel oil backup, because the peaking power plant cannot run all the time.

These are just some of the nuances that I think get overlooked in the data center debate.

MR. PANTELOGIANIS: We have data center developers that lock in leases to serve hyperscalers. The data centers get built and then the parties -- and lenders -- have to wait for the power in some instances for three years without causing any defaults in the underlying lease agreements with the hyperscalers.

MR. SMUTNY-JONES: You are in deep trouble. Welcome to California.

No one talks about this in California because we have not sited a new gas plant since 2017. When we did the last one, Jerry Brown was on his way to Bonn for the world climate-change conference. The gas plant was on the 95-yard line of being approved and the California Energy Commission said, “No thank you. We are not going to do any more gas.” 

The dirty little secret is about 44% of the power in California comes from the gas fleet. The other 56% of electricity comes from a diverse set of resources, including the Diablo Canyon and Palo Verde nuclear plants.

The real question in California is how we are going to sustain the existing gas fleet.  California spends a lot of time planning. We have an integrated resource plan. It has gas in the system until 2045. The capacity factors drop and some of the gas plants retire, but gas is expected to remain part of our resource mix for the next 20 years.

There are people who think the most important role for the gas fleet is fast ramping. Our system is built around a net peak between four in the afternoon and eight at night. It is not just data centers. We are electrifying everything. In the summer months through September, we have a lot of solar electricity that can fill the gap.

The issue is when we get to winter, we do not have as much sunlight as we do in the summer.

Having a portfolio of resources is really important. The gas fleet comes under constant attack for generating greenhouse gas emissions. We have a cap and trade system, so the gas generators pay for their CO2 emissions.  

If you look more broadly at the West outside California, I think we will see a lot of activity. Electricity demand is expected to increase by 20% in the WECC part of the grid in the next 10 years, which is twice the projected load growth just two years ago.

There will be a significant amount of development opportunities, assuming you can get equipment.

A five-year wait for a gas turbine, main power transformers and other key equipment is absurd. If we have a major earthquake in California and have to replace a bunch of substations, what happens then?

Nuclear

MR. PANTELOGIANIS: I was a rookie 31 years ago in this business and, when I was learning about the fleet, it was diversified in terms of baseload options. We had coal, we had gas, we had hydro, and we even had some old oil-fired power plants.

If you look at what we have done today, we have essentially gone all in on gas, and we are struggling to figure out how to find the additional megawatts of electricity required to satisfy rapid load-growth forecasts.

Is this a period for transformation an opportunity?  Is this the opportunity for another baseload fuel, like small modular nuclear reactors, to come in and transform the space?

MR. SMUTNY-JONES: I don't think you can assume nuclear can fill the gap.

MR. PANTELOGIANIS: I am not saying it has to be nuclear. I am open to asking what will fill it.

MR. SMUTNY-JONES: Find me a nuclear power plant that has been built on budget and on time, anywhere. The Vogtle plant ended up costing close to $37 billion. I am not anti-nuke. Diablo Canyon is not going anywhere. We cannot start retiring stuff like that, even though it is 40 years old.

MR. LINK: Can any of you tell me what the price would be for electricity from a small modular reactor that has an installed cost of $10,000 per kilowatt? Can you do that math quickly? 

MR. FLANAGAN: No math required.

MR. PANTELOGIANIS: Declan, when you give that answer, please note that the size of utility-scale nuclear facilities and the escalating costs require more banks, which means interest costs have to go higher to fill in the bank syndicate.

Brewing Political Storm

MR. FLANAGAN: This is a pretty bleak picture we are painting, and no one should take any joy from it. The electricity industry has a terrible habit of fighting technology versus technology. Storage fights with solar, and solar fights with wind, and gas fights with renewables.

The big issue is how deregulated markets have served the customers. We are perilously close to a situation where politicians across the spectrum turn and say, “This does not work for us, because you have messed up the baseload. It used to work 30 years ago.” 

Ten years ago, combined-cycle gas-fired power plants didn't pencil in ERCOT at $900 a kilowatt. They turned into solar projects. No one should take comfort in this. It is not an outcome that serves renewables well.

Not serving the growing electricity demand is going to be a political problem that leads to re-regulation. What does that look like?

We need bring down the cost of electricity across the board. Yet we are doing the complete opposite.

Across the board, we are driving up the cost of electricity. I saw this play out at a microcosm in Ireland where I lived for the last few years, where my utility bill in my house in Ireland is now $400 a MWh. Europeans are more used to these prices, but in the US, it is not going to fly.

Ireland let data centers grow disconnected from grid planning, and it caused big political and social affordability problems. You were getting backup generators on barges for winter peaks. That is expensive stuff. It feels like this is about to play out on a much more massive scale in the US.

MR. LINK: For the record, I do not think gas is screwed. Otherwise we would not be buying gas turbines. There is enough demand for all of us regardless of technology. We are still in the cycle of replacing the large number of coal-fired power plants.

Unless demand deteriorates, which we do not want of course, we are in for escalating power prices.

If you focus on the SPP, MISO and PJM parts of the US power grid, they are heading for reliability crises. This has to be figured out.

Maybe the simple fix is we cannot get as many data centers online as quickly as developers want until the equipment and manpower issues get resolved.

MR. BURTON: It is not just the construction trades where we are short manpower, how many bankers have underwritten a gas-fired power deal? Do the bankers even know how to do this anymore?  And if you graduated from school a few years ago, you have probably never seen a gas-fired deal.

MR. PANTELOGIANIS: We are very active in refinancings of both standalone CCGTs and portfolios of such projects. I have no concern over experience to provide project financing to CCGT projects. None.

What I have concerns about is costs being driven to levels that match costs in places like New York City. That requires a greatly expanded pool of lenders. There is not enough lending capacity among the traditional project finance banks.

We financed most of the activity in 2013 through 2020 through the bank market. But if we are going to be financing new CCGTs at $2,200 per kilowatt of capacity, we are going to need leveraged debt or CLO managers to participate, which means the rating agencies are going to have to provide a view for the CLO managers to be able to participate in greenfield financings. Otherwise, it means more equity, which adds to the cost. More parties means more complicated and slower moving financings.

MR. BURTON: Jan Smutny-Jones touched on the lack of appetite in California for more gas-fired power plants. The last gas-fired power plant in New England was the CPV Towantic Energy Center in Oxford, Connecticut that was commissioned in 2018. It took 20 years to build.

Are we going to see more gas-fired power plants built in New England and, if not, where in the East will they be built?

MR. LINK: New England is very difficult. Permitting is challenging. Even if you can get such a plant permitted, the potential supply of gas into New England is severely limited. We are focused on Virginia and Georgia.

On the west coast, we have a couple of biodiesel peaker developments in Washington state and that might be a way to thread the needle through biodiesel, but biodiesel is not as plentiful as natural gas and oil. The bottom line is the east and west coasts are pretty challenged. As you near the coasts, you get to the end of the line on gas pipelines, and gas supply is a real concern.

MR. SMUTNY-JONES: One of the problems that I see with the current federal policy is it is killing carbon capture and storage. I live in the Sacramento Municipal Utility District, and Calpine had a deal with SMUD to capture and sequester the emissions from a gas plant that was built in 1999. That just got killed. If you are going to "drill, baby, drill" and you want to be able to build more fossil fuel plants, there ought to be an effort to prove that carbon capture and sequestration work.

More effort should be put into mixing fuels.

I have members that are looking at mixing hydrogen with gas. In 2008, 18% of California's power came from coal. Now it is down almost to zero. It will hit zero at the end of this year when LADWP will finish converting its remaining coal plant into a combined-cycle gas plant that can run on a mix of hydrogen and gas. It is the gas fleet that helped take out the coal in the West.

Kat Gamache gave me a hard time yesterday because I held up a unicorn on a panel at this conference some years ago as a proxy for utility-scale battery projects. They did not exist. That was 2016. By 2020, we had 300 MW of battery capacity as a hedge against rolling blackouts. Now we are at about 14,000 MW just five years later. We are pretty good at problem solving. It is not all doom and gloom. We will figure it out, but let's also not underestimate the challenges, and it is not helpful if government policy is working at cross purposes.

Paradigm Shift

MR. BURTON: We have time for one audience question.

MR. STEUBE: Erik Steube, president of Ecoplexus. I was surprised not to hear more discussion about the fundamental shift in how we think about baseload power. Instead of the traditional categories of gas, coal, nuclear, etcetera, people are also thinking of it as renewables plus storage supplemented by gas. Many of us have seen analyses of energy parks that are 80% renewables with storage and with 20% gas backup coming in at something like $80 a MWh for electricity. I am curious to hear your views on that.

MR. FLANAGAN: I completely agree that is what is required. That should be the main takeaway from the economics and schedules.

We have a capacity challenge to overcome. Get more solar and wind at cheap megawatt hours and batteries to move the electricity to the right times of day when demand is greatest. I am sure Michael Skelly can move power much more cheaply than the cost to build new fossil fuel plants at $2,200 to $2,500 per kilowatt of capacity.

The key is transmission to move electricity and batteries to shift the time of use. That is a paradigm shift.

There is a role for gas. There will be some behind-the-meter, data center baseload. There will be regulated new build, but the main takeaway is you need this paradigm shift. It is not clear that anyone is pulling all those threads together into a coherent policy on a regional or national level.