The Coming Energy Shakeout: Data Centers, LNG, ESG, and What Breaks in 2026

---

Support CleanTechnica’s work through a Substack subscription or on Stripe.

---

As it’s the start of a new year, Redefining Energy’s annual predictions show has dropped. Every year Laurent Segalen, Gerard Reid, and I assess each other’s predictions about global energy and decarbonization events and milestones from the previous year, scoring them and proclaiming a winner. Every year, we make more predictions. Read or listen, as you prefer.

Laurent Segalen [LS]: Happy New Year.

Michael Barnard [MB]: Happy New Year to the both of you. It is going to be an awesome year. There’s going to be so much interesting stuff happening in 2026 that’s positive.

Gerard Reid [GR]: I agree totally with you, Michael. I agree totally. It’s going to be a crazy year.

[LS]: But first, gentlemen, before talking about the new year, we need to revisit the predictions we made a year ago. Let me remind our listeners of our six predictions. Michael said that oil production would be down in the US in 2025. Gerard said oil prices would hit $40 per barrel in 2025.

[GR]: Can I correct you there? I said $50.

[LS]: No, no, you said $40.

[GR]: I didn’t. You couldn’t understand my accent, Laurent. I mean, sorry, I. 50 is what I said.

[MB]: Sadly, the record is against you on this one, Gerard.

[LS]: Fifty, starting with a four. Third one: geopolitics, stress, supply chains, and an energy bonanza will bring a more innovative and better world. That’s me. That’s a Gerard-style prediction.

Number four, Michael: a bloodbath for hydrogen in the transportation sector.

Number five: record installations of solar at 700 gigawatts, EVs at 20 million, and batteries at 200 gigawatt-hours, from Gerard.

And finally, from me, number six: the end of all financial products labeled ESG, climate, and carbon.
Let’s start with number one, Michael.

[MB]: I failed. I’m off by about eight months. Oil prices and oil demand were buoyed because China has massive reserves and kept buying oil far past the point of consumption. Even though actual diesel and gasoline demand are down in China, their petrochemical industry has gone up a bit. So their actual oil consumption is relatively flat.
But their reserves reflect a view that the geopolitics merit having significant запас—significant reserves. That kept demand up, which kept prices up.

That led to the second point, which is that shale wells in the United States did get finalized. But drilling rigs in the United States are way down. They’re not completing nearly as many wells. China’s reserve buildup has to be complete at some point.

I’m not making this one of my predictions, but I’m off by months. It didn’t happen this year.

[GR]: And by the way, building on that, my point about $40 oil relates to the same thing. At the end of the day, demand was stronger than we expected. But I would also agree with Michael and say that in the coming year we will see prices move down toward $40.

[LS]: The price of oil has gone from, say, $80 to $60. That’s a lot. But let’s rate it. Michael, I give you a 2 out of 10. Gerard…

[GR]: Oh, I give Michael 5 out of 10. It was just the timing was slightly out.

[MB]: That’s very generous of you, Gerard. Thank you.

[GR]: It’s 2026. It’s going to be a great year.

[MB]: I would have given myself a three out of ten. So a two and a five. I’m in the range. I’m happy with that.

[GR]: But I do want to say one thing: Mike gave a great explanation. The geopolitics of energy are what really made 2025 complex.

[LS]: Gerard, your prediction was very simple, and you got it wrong. So I give you a zero because it was very binary. You get a zero out of ten. Mike, what do you give Gerard?

[MB]: I’m just going to say that my prediction was binary as well, and you were more generous to me. I’m going to give Gerard a three on that. As he said, it got down to 50. It was lower than any of the oil boffins thought it was going to be. It was a bad year for oil.

[GR]: Agreed.

[LS]: My prediction was vague, so it was a Gerard style. The winning strategy was that geopolitics, stress, supply chains, and an energy bonanza would bring a more innovative and better world. A better world, I don’t know.

But what I do know is that the stress on the grids has driven progress. We’ve seen advances in transformers, advances in digitization, and incredible development of batteries. We even saw fuel cells evolve. If you look at a company like Bloom Energy, they are no longer doing fuel cells for hydrogen, but for gas. We’ve seen advances in grid studies with AI companies like Tapestry. We have advanced geothermal and a lot of work done on flexibility.

So for me, geopolitics and rising energy demand have pushed innovation forward.

[MB]: Gentlemen, my take is slightly different. I’m not disagreeing with you, except on Bloom Energy, which has always been a natural gas fuel cell company with very dubious environmental claims. It tried to ride the hydrogen hype and didn’t do much there. Now it’s riding the data center hype, but it’s still just a solid oxide fuel cell running on methane.

That said, the geopolitics have been very interesting. Pakistan is trying to find a new home for 24 LNG ships’ worth of long-term LNG purchases from Qatar. That is seismic. Those long-term contracts were supposed to buoy LNG, yet Pakistan is saying no because of the 17 gigawatts of rooftop solar it has installed.

Similarly, in the first six months of this year, India and China dropped LNG imports by double digits. That coincided with India cutting gas-fired generation by 34% year over year.

The world is changing tremendously for the better because of cheap solar and batteries. I see evidence of this globally. The LNG industry is not ready for it. It is going to be a bloodbath in the LNG industry, which isn’t my prediction, but it is the reality.

[GR]: Let me talk about the wording of your prediction. You said geopolitics, stress, supply chains, and an energy bonanza will bring a more innovative and better world.

Let’s start with stressed supply chains. I don’t see stressed supply chains. I see oversupply of almost everything in the energy space, whether it’s oil, gas, coal, or electricity. We’re swimming in electricity in Europe. There’s too much solar, too much battery, too much wind.

I take the point about a more innovative and better world, but I don’t know about that. What’s more innovative than last year? I don’t really see any huge innovations over the last 12 months. There’s no doubt the world is a better place, and the Pakistan example is a good one.

Because it was so vague, I’m going to give you a five out of ten.

[LS]: What about you, Michael?

[MB]: I think Gerard is being generous, given how vague the claim was. I pushed back quite hard on that point. Predictions are only useful if they are testable and quantifiable. If they are not, they are closer to narrative than analysis. I try to be consistent about that standard, including when it is uncomfortable. On that basis, I cannot justify a higher score. I am giving it a four.

Which gives you the highest score in the first round of predictions out of the three of us, because I failed completely on mine and Gerard failed completely on his. Generously, you both gave me more than zero, which I thought was very nice of you.

[LS]: The fourth one, Michael, you nailed it: a bloodbath for the hydrogen transportation sector. Do you have a few examples?

[MB]: Oh my God. It has just been schadenfreude city. Here in Vancouver, my schadenfreude cup has been overflowing. I’ve been talking about hydrogen and transportation, publishing and analyzing it, for over a decade. This year, all the dreams came tumbling down.

Let’s start with one of the big ones: heavy trucking. The claim was always that batteries weren’t going to cut it for heavy trucks in the freight industry. Except in China, which actually ran the experiment with hydrogen versus battery electric. In the first nine months of this year, they sold 90,000 battery-electric heavy trucks. Meanwhile, the number of hydrogen trucks collapsed, going from the low thousands to even lower thousands, down by something like 40%. That’s data point one.

Globally, we’re seeing battery-electric trucks being delivered, not hydrogen trucks.

Similarly, fuel cell cars have essentially disappeared. Alstom is now out of the hydrogen train business. They said they’re closing down that division. If customers still want the hydrogen trains they already ordered, Alstom will deliver them, but they’re clearly signaling that customers shouldn’t even keep them and should take battery-electric instead.

Looking across the data more broadly, I track a list of 171 firms and organizations that were active in hydrogen transportation. This year, 36% have formally dropped hydrogen or pivoted away from it. And that 36% significantly understates the reality, because my standard is high. They have to explicitly say they’re dropping hydrogen, not just quietly slink away. I think the real number is well over 50%.

Airbus dropped its hydrogen aviation program. Stellantis dropped hydrogen trucks. Plug Power and FuelCell Energy are struggling with reverse stock splits just to stay listed. They’re cutting maintenance at their hydrogen plants, which is dangerous.

It is really ugly in hydrogen transportation right now. And, as I said, I’m enjoying every minute of it.

[LS]: Michael, well done. So I give you eight out of 10.

[GR]: I give you a nine because you’ve been calling it for quite a while. Well done. Great job.

[MB]: I’m wondering how much more of a bloodbath do you want? What would cause it not to be a 10?

[LS]: Based on what I’ve shared, it’s very simple. I have quantitative criteria, and I think your answer was a bit vague.
Job number five. Now, Gerard, on that one, he gave numbers, and the numbers were pretty good.

[GR]: I was put under pressure, so I said 700 gigawatts of solar. Looking back, we don’t really have the full numbers yet, but I think they were close to that, maybe around 680. EVs at 20 million.

We’re all maybe a little bit above that, and definitely batteries too. I said around 200 gigawatt-hours, and I think it’s probably even higher than that.

So I think I did pretty well on my numbers.

[LS]: But it was a very good trick on your side, because first you said 200 gigawatts and then you said hours. In fact, because all the batteries are four hours. Well done. Well done on this one.

[GR]: And record numbers, right? They really were. It’s amazing what’s gone on.

[LS]: All right, so on this one, you deserve a 10 out of 10.

[MB]: I’m giving you a 10 out of 10 as well. We’ll reward you for predicting the bleedingly obvious. You picked numbers, you were precise, and you had three different numbers. And yes, 2025 has been an amazing year.

[LS]: The last prediction from last year was mine: the end of all financial products labeled ESG, climate, and carbon. If I look at the number of funds that were named climate or ESG, about 25% have been rebranded from sustainable to transition. In the US, there have been 12 consecutive quarters of outflows.

It comes down to decarbonization versus electrification. Everything that falls under decarbonization writ large, and I’m talking about carbon capture and hydrogen, is just burning money, often only to be bailed out by taxpayers, and probably less and less going forward.

By contrast, electrification funds are doing great because, as I’ve always said, switching fuels makes more sense than fixing fuels. Gentlemen.

[GR]: Great speech, Laurent, but I’m about facts. I just go into Google and type “ESG investing,” and what comes up is PIMCO. So I go into PIMCO and look, and there’s the PIMCO Climate Bond Fund, Global Investment Grade Credit, ESG fund, and so on.

So there’s no end. End means it’s finished. It’s still here. That’s the reality. There’s no way you’re getting good marks from me.

[MB]: You got the direction of travel right, but like oil for Gerard and me, you missed the timing. You were far too aggressive on timing. Yes, they’re down, but 25% is less than 36%.

From a quantification perspective, you certainly aren’t getting a perfect score. I’d give it a 5 out of 10. You were right that a bunch of them dropped and a lot of changes occurred in that space, but it wasn’t a bloodbath in ESG and climate funds in the same way.

[GR]: Yeah, I’ll give you four.

[LS]: So I recount as usual. I’m last. And we have a winner. Again, it’s Michael.

[MB]: I was expecting not to be the winner this year because I missed one really badly.

[LS]: You got the hydrogen one pretty good.

And Gerard, the first one was such a disaster. That was solved, but it wasn’t enough to catch up with Michael. So as usual, the winner is Michael. Well done, Michael.

[MB]: I’d just like to say that my uncle Nostradamus is rolling over in his grave. I was far too precise and testable in my predictions.

[LS]: There were two additional predictions we made along the way. The first was that we thought BP would not survive to 2025. They did, so good for them.

[GR]: Yeah, they did.

[LS]: The other one was a prediction Gerard made 15 months ago on the price of silver. He was very bullish on silver when it was at $30, and now it’s at $60. This one is for Gerard. Well done.

[GR]: We got one right.

[LS]: That’s good. Gentlemen, it’s time to move to our predictions for 2026. Who wants to go first?

[MB]: I’ll go first this time. My first prediction is that in China we’re going to see battery storage auctions at $40 per kilowatt-hour for grid-scale battery energy storage systems, with 20-year contracts maintained and deployed. We were at $65 last December, which was astounding. I think we’re going to hit $40 this year.

[GR]: Can I ask you a question, Michael?

[MB]: Sure.

[GR]: What technology Is it lithium ion or is it sodium ion or what’s your view?

[MB]: LFP. The Naxtra sodium ion batteries that CATL is bringing in are promising, but they’re also economically heavily challenged by LFP being dirt cheap and having most of the advantages.

[LS]: What’s incredible is the new form factors that are arriving. Everybody used to work with 314 amp-hour cells, and now it’s 587. We’re even seeing cells above 1,000 amp-hours. The bigger the cells, the lower the cost of the system.

[GR]: It has huge ramifications, doesn’t it? If you just think about solar and batteries together, the prices are crazy.

[MB]: And it’s not just the cells, Laurent. My analysis, which was wrong and trite and based on first principles, was that it would be the same as what we see with hydrogen electrolysis plants, where the balance of plant doesn’t get cheaper because it’s commoditized. That was true for battery energy storage systems for a while.

But in China, with LFP, they assemble the cells, put them in a rack, slap a bus on top with aluminum conductors that are dirt cheap and easy to assemble, shrink-wrap them, put them in a box, and drop the box in a field. It’s an insanely light balance of plant. They’re still innovating across the entire system. There’s still juice to be squeezed out of that lemon.

[LS]: Good job. I’ll go second, and of course I’m going to talk about data centers. Data centers and the new Panic of 1873. We’re going to see a gigantic purge. People talk about a bubble, especially at the end of last year, and there has been some extremely unhealthy behavior.

You have startups borrowing money to build data centers for other startups. Both are losing tremendous amounts of cash, yet somehow they’re still able to raise debt capital to fund these buildouts, without customers or any visibility into whether those investments will pay off.

I predict that, like last decade with Adam Neumann at WeWork, this decade’s equivalent is Sam Altman at OpenAI, but at a much larger order of magnitude.

Google and Microsoft are going to be okay. But on the periphery, there are a number of companies, and I even wonder whether we should include Oracle in that group. Blue Owl, CoreWeave. Everything is built on credit.

If you look at CDS, and I say this as someone who was at Lehman 20 years ago when everyone knew what a credit default swap was, people seem to have forgotten since then. Over the past three months, CDS have started coming up in conversation again. The CDS spreads of those peripheral companies look like Egypt, Ethiopia, or Sri Lanka.

It’s all well and fine, Gerard, to point to PIMCO. But at PIMCO, they’ve put out $28 billion of credit.

[GR]: So could you give us one line, please, so we can actually quantify it and see whether you’re right or wrong?

[LS]: Okay, 50% of the data centers announced will never be built.

[MB]: That’s remarkably low. But I’ll take 50%. That’s an easy target to hit. I agree. I have specced data centers in my professional career. I’ve even put loads in data centers. The first time in the 80s I put loads in the clouds and cloud based data centers. And I’ve been doing AI professionally on and off for 15 years including in one of my current firms. And I’ve been a performance engineer on major software programs. So I know three rules for software. First, never optimize early. Just assume you’ll be able to throw hardware at it because that’ll be the cheapest fix. Second, if you have a performance problem, throw hardware at it because that’s the cheapest fix. Third, if you can’t fix it with cheap hardware, then get intelligent people to optimize it. And that’s what Deepseek did this year.

Deepseek had intelligent people, they optimized the software, they got amazing results with back level chips in a fraction of the training and processing time and they’re selling that model globally. A lot of the innovation in stuff is scripting that isn’t using the large language models. So I know the demand bubble is gone. Now I’d like to talk about just a little bit about the corollaries, the implications of that. When I analyzed the AI spike, what I found was that vast majority of the data center build out is in the United States. All of the stock gains of the US stock market have been in the seven big tech companies. We’re building massive data centers. Everything else is flat or declining. The United States is in structural recession except for the AI data center build. So that’s kind of economic indicator one.

Economic indicator two is that the top 10% of income earners are spending 50% of the consumer spending. They’re spending that because their 401ks are good because of this data center thing. They’ve got index funds that are over invested in these data centers. There is a bloodshed coming. This is not my prediction, but there’s bloodshed coming in the American economy next year as these economic indicators collide and the top 10% stop spending.

[GR]: I think you’re both wrong. Oh, I think it’ll be 2027, not 2026. You got your timing wrong.

[LS]: Gerard, you’re going to have to explain this to me. Texas has received 226 gigawatts of requests for new data centers, and the peak load is 85.

[GR]: No, on what you said about data centers, I agree. They’re not going to be built. It’s the same as all the battery development projects in Germany that aren’t going to be built.

The point is, you’re going to see record capex expenditures in 2026.

[LS]: No, no, no. You are a romantic. You are looking at equity, but…

[GR]: You’re thinking about 50%. When everybody rushes for grid connections, they apply everywhere, including projects that are never going to be built. So for me, your prediction isn’t a proper prediction.

[LS]: You’ll say that. What I’m saying is that you forced me, you cornered me into giving you a number. But my point is simple: the debt market will not finance all the data centers being announced. Period. There’s just too much debt. Now the providers of debt are starting to freak out. The risk premium is going up.

So that’s it. We agree to disagree. We’ll count our chickens in 12 months.

Gerard, your first prediction.

[GR]: I’m not going to be vague. What you’re going to see is lower global wind and solar installations in 2026 than in 2025. In fact, I’d even go so far as to say we’ve seen peak global wind and solar installations.

This is not to say the energy transition is failing. What’s actually happening is that we’ve hit massive system constraints. Grids are full, permitting is slow, financing costs are higher, curtailment is rising, and capture prices are falling. That’s what’s going on.

In some ways, wind and solar have become victims of their own success. But as I said, I see this as part of the revolution we’re going through.

[LS]: Solar peaks? Solar. We did an episode last year, and wind was set to grow into 2025. So yes, absolutely.

[MB]: I’d like to reflect on this, because it actually leads into my prediction, which I won’t share yet. There’s one key country I always look at, because what China does moves the needle. China changed its solar price points for developers about three quarters of the way through the year, and everyone rushed to get solar built at those price points.
We see the same behavior everywhere in the world. Whether it’s feed-in tariffs, the investment tax credit or production tax credit in the United States, fiscal incentives and price points from governments make a difference.

This year, about 300 gigawatts of solar went in. Next year, the best estimate is likely closer to 200 gigawatts in China. So I think Gerard may be right.

But in my prediction, I’m going to hold out a very interesting potential. For now, let’s finish with Gerard’s. Laurent has said almost nothing. He hasn’t abused you yet. What’s going on?

[GR]: No, that’s all right. That’s good. He’s been nice to me. Go on, Michael, go straight into yours.

[MB]: China drops 100 gigawatts of capacity compared to last year. Let’s assume that’s true. There’s still more than 300 gigawatts of solar panel production capacity. And China isn’t an insular country. It trades.

So that 100 gigawatts of production isn’t going to be used domestically. It’s going to be used internationally. The same dynamic we saw in Pakistan, which was suddenly flooded with cheap solar panels and, to a lesser extent, batteries, is going to play out elsewhere. Pakistan installed 17 gigawatts of mostly behind-the-meter solar in a single year, blowing past everyone’s expectations and, as we discussed earlier, disrupting the global LNG market. That dynamic is going to occur globally.

We’ve already seen the breakout of solar in Africa. There’s a flywheel effect there. There are a million Chinese expats living across Africa. The Belt and Road Initiative has built ports, railways, and 12,000 kilometers of highways. Africa is now moving toward a free trade zone between countries.

China has a massive glut of solar panels and will have significant battery capacity as well. What this tells me is that we’re already seeing an uptick. Last year, about 2.5 gigawatts of solar were deployed in Africa. My prediction for 2026 is 20 gigawatts. I’d like to say more.

[GR]: Wow, wow, wow.

[LS]: So prediction, Michael, 20 gigawatts of solar.

[MB]: In Africa in 2026. I think I’m being conservative, frankly. We’re all going to be surprised.

[GR]: I’m actually totally with you on that, Michael. I actually agree with you totally.

[LS]: My second prediction is that the fight over the Greenhouse Gas Protocol is going to turn ugly. Let me remind you of the previous episodes. The Greenhouse Gas Protocol is a non-binding, voluntary system for accounting for CO₂ emissions. It was established 25 years ago and is revised on a regular basis. It is now in the process of negotiating its third revision.

The most important issue is whether we count electrons or certificates. Historically, emissions were counted on a yearly basis. The proposed new system would move to hourly accounting. Ninety-nine percent of companies agree with moving toward an hourly system.

There are two holdouts, and I’m going to name them because they’ve already been named in the press. I’ll call them Darth Amazon and Darth Meta, two Sith Lords who are doing everything they can to lower the standards. They’re backing an alternative system called Emissions First, which, frankly, is a joke.

They’ve been outvoted, put in the minority, and now they’ve threatened to take the issue to the US Senate. So this fight is going to turn very ugly. I hope reason prevails, but there are some tough times ahead. That’s my prediction.

[MB]: I’ve actually published and analyzed quite a bit about the new proposals. There are various groups that do accounting the old way, and they like the old way. There’s a spin-off from RMI, the nonprofit, called WattTime. There’s also another company, ReSurety, which does liability assessments for VPPs and VPPAs.

Their position has always been that global emissions are what count, so putting a solar farm in a high-coal region is better than putting one next to a data center if that data center happens to be in a lower-coal area. That’s a valid point.

But the argument right now is about additionality, temporality, and locality. You can think of it this way. Additionality is bringing more beer to the party you’re going to. Don’t drink the beer that’s already there. Temporality is bringing beer when the party is actually happening, not the next day. Locality is bringing beer to the house that’s having the party, not the house next door.

What that means is you have to put new generation near the demand area. The demand has to match the generation from the new asset, close in time and location. That’s a tough standard. It’s an expensive standard.

[LS]: It’s really two guys. I’m not going to name them. They know exactly who they are, set against everyone else. The fact that they’re willing to use strong-arm tactics, leveraging their lobbying power and going to the Senate to do what you might call skulduggery, that’s not how this industry or community works.

They’re in the minority. The answer can’t be “my way or the highway.” In this case, it’s going to be the highway. That’s my prediction.

Okay, Gerard.

[GR]: The last prediction comes back to Michael, and we broadly agree on it: the golden days of LNG are over. What I mean by that is we’re going to see a whole set of businesses get into financial difficulty because they’ll be selling at negative margins. At the same time, a lot of new LNG capacity is coming online, which means you’re also going to see asset problems tied to those investments.

There are a few reasons for this. First, global demand for gas has peaked. It may already have peaked in 2025, but if not, then in 2026. OECD demand is shrinking and continues to do so. Growth in China is down. What Michael said earlier about Pakistan rolling out solar and similar developments is only going to become more important.

This also means that emerging markets, as Pakistan illustrates, are becoming far more price-sensitive.

The final point I’d make is that we’ll see Russian gas come back into Europe, and that changes the entire dynamic. I’m saying that because I believe part of any peace agreement will involve Russian gas returning.

So the golden days of LNG are over, and I think we’ll start to see the restructuring of that system toward the end of 2026.

[MB]: Let’s talk about that LNG thing.

[GR]: Go for it. You talk about it because you brought it up, Michael. Go for it. Yeah, yeah.

[MB]: I think Gerard is right. I’ve been looking at global LNG demand and growth, especially North American growth in LNG exports. In Canada, we have two massive projects approved through Mark Carney’s Major Projects Office, including pipelines.

The implications of all this capacity coming online are very interesting for the United States, and they point to another economic problem. Joe Biden put a halt to new LNG export terminals because domestic gas prices were rising. They were becoming pegged to international prices instead of domestic ones. With massive amounts of export capacity coming online and bidding into spot markets, that will push domestic prices up.

At the same time, the cost of domestic production is rising. Just as shale oil is moving into more marginal resources, shale gas is hitting marginal sites as well. So domestic production costs are going up.

The result is exposure to international prices. Even if those prices are depressed, domestic energy prices in the United States are likely to spike because the system remains heavily gas-dependent. People are going to be in real financial trouble in the United States in the coming year.

[LS]: Gentlemen, if I summarize our six predictions:

Michael predicts that China’s battery systems will drop to $40 per kilowatt-hour. ‘

I predict a credit and grid crisis around new data centers. You asked me for a number, so I said 50% of announced data centers will never be built.

Gerard predicts lower wind and solar installations in 2026 compared to 2025.

Michael predicts 20 gigawatts of solar in Africa in 2026.

The fight over the revision of the Greenhouse Gas Protocol is going to turn ugly.

And finally, the LNG glut will create distressed assets, with flat demand and more supply arriving on the market.

Gentlemen, I’d like to remind you and our listeners that Ukraine is still at war. It’s been almost four years, and our hearts go out to the Ukrainians who are resisting these awful attacks every day. Our thoughts are with them.

Any last words, gentlemen?

[GR]: I’m just going to wish you both a great 2026, and the same to all our listeners. Thank you very much for listening. We’re looking forward to serving you over the next 12 months.