AI's insatiable appetite for electricity could revive a forsaken energy source
By Maksym Misichenko · CNBC ·
By Maksym Misichenko · CNBC ·
What AI agents think about this news
The panel consensus is that Babcock & Wilcox's (BW) exposure to a coal revival is speculative and risky, with utilities economically incentivized to build natural gas plants instead. The key opportunity lies in BW's natural gas capacity, but even that faces strong competition from GE Vernova. The single biggest risk flagged is the high short interest in Applied Digital, a key partner in BW's backlog, and the uncertainty surrounding Wright's emergency orders blocking coal plant closures.
Risk: High short interest in Applied Digital and uncertainty surrounding Wright's emergency orders
Opportunity: BW's natural gas capacity, if it can compete with GE Vernova
This analysis is generated by the StockScreener pipeline — four leading LLMs (Claude, GPT, Gemini, Grok) receive identical prompts with built-in anti-hallucination guards. Read methodology →
Coal could come back. That was my insight, or more, my instinct, after interviewing Kenny Young, the CEO and seven-year veteran of Babcock & Wilcox , the 160-year-old boiler manufacturer turned engineering and construction company. Young wanted to talk about the surging demand for power spurred by the data center boom. So did I. B & W has a $2.7 billion backlog, $2.4 billion of which is a deal with Base Electron, backed by Applied Digital , a company purpose-built to design digital infrastructure for high-performance computing. I want to write about this "Mad Money" encounter for a few reasons. First, to show you that the data center story is so much bigger than we imagine. Our thinking is constrained by a particular negative bias that says it all has to end, like the dot-com crash of the early aughts. That bias has kept people from making easy money, like the money you would have made by buying Babcock & Wilcox stock, which is up 244% this year alone. This $21 stock traded below $1 one year ago. Second, I am no groundbreaker here: obviously, I'm late to the party. But that has somehow meant nothing to so many of these companies — witness Micron , Intel , Sandisk — that I must acknowledge my timing. Some of you might consider it late, late, late, as I wrote about last week . Others argue, 'So what, it's the data center.'" Third, I want to point out that the power demands are so great that the once-forsaken energy source of coal is going to come back in a big way if the utilities don't stop President Donald Trump and the Department of Energy from forcing coal-based or coal-using companies to continue using it. The dirty fuel — at least relatively if not absolutely — accounted for 50% of U.S. power in 2007 and is now down to 15-17% of the grid's energy source. Even though it is down 40% from 2010, it still powers 173-190 gigawatts (GW). We may need 90-100 GWs of new energy if the data center buildout continues at this pace, so the idea of reviving coal, or at least not letting coal plants close, is hardly fanciful. I write that because while the environmental toll of coal has been obvious for generations, the president regards coal as a major resource and domestic national security weapon. Which brings me back to Babcock & Wilcox. Last Friday, B & W placed an offering of 10.8 million shares at $18.50, mostly to shore up its balance sheet and prepare for a major expansion. The stock had gone out at $21.22 the day before and finished trading at $21.85. So you could call the deal wildly successful, even as it was handled by B. Riley, a brokerage house under investigation by the Securities and Exchange Commission. I am disclosing this relationship because Wolfpack Research, a short-selling firm, has cited it as a negative for Applied Digital and, by extension, for B & W. What's considered wrong here? Mainly, Wes Cummings, CEO of Applied Digital, also served as the president of B. Riley Asset Management until February 2024. Short sellers have claimed that the $2.4 billion contract with Applied-backed Base Electron was used to pump up B & W's stock. B. Riley holds a substantial stake in B & W, so the increase in backlog was meaningful. I don't think it could have done that secondary all the way up here without it. I go into all of this not to discredit B & W, because it clearly has the technology needed to build plants, but because I don't want anyone to think that B & W has huge multiple-power contracts. The company would most likely not have raised the money so easily without the Base Electron contract, and we have no real assurance that Applied Digital will go through with its plans. Like many companies in this arena, Applied Digital loses significant money. So does B & W. Applied Digital was worth $1.5 billion a year ago. It is now worth $12 billion. Late, late, late. Importantly, 32% of Applied Digital's outstanding shares are sold short. That could be the usual skepticism and bias against so many of these data center "stories." Or it might be the tenuous relationship with the tarnished B. Riley. Applied Digital has a relationship with CoreWeave , which accounts for the bulk of Applied Digital's $16 billion backlog. It has another 15-year lease with an unnamed hyperscaler. That's enough to make Applied Digital "real" and therefore validate B & W's stock price increase. Another positive for B & W: it has its own proprietary capability for building natural gas power plants. Right now, GE Vernova , which the trust owns, is the principal builder of natural gas-fired plants. But GE Vernova has made a point of telling me that it is sold out near term and can't add more plants than it has now. In fact, that's the principal rap against GE Vernova. It's out of capacity. That makes B & W a good secondary call on natural gas plants, and the company assured me that it is not limited and has the capacity to build more. That could be a terrific part of the B & W story. Remember, however, that I am not a groundbreaker. All that I have told you is known to both B & W aficionados and vocal short sellers. So, if you buy B & W stock, you have to do so knowing about the controversy surrounding the company, despite its ties to the totally legit but heavily indebted CoreWeave. Perhaps, given the secondary and the B. Riley overhang, the easy money has been made here. But perhaps for another reason — coal — it hasn't. Which is what got me most excited about this story: B & W's exposure to coal. While this giant Base Electron-Applied Digital contract for 1.2 gigawatts is right in the wheelhouse of the natural-gas-powering data center story, B & W is basically a coal play. B & W has a hand in building and maintaining coal plants worldwide. It's the best at what it does. Except that what it does is being phased out worldwide, especially here, where the country — including electric utilities themselves — has tried mightily to wean itself off coal. Until 2025, when President Trump came to office. Trump's a huge believer in coal. A year ago, he signed an executive order called "Reinvigorating America's Beautiful Clean Coal Industry." For one moment, forget the Orwellian nature of the order, No. 14241, and accept that we have a lot of coal plants that would otherwise be closing after several decades of executive orders and agency prosecutions against coal because of the inherent pollution it causes. It is true, though, that these plants are vital to the baseload of many utilities around the country. They are also integral to providing back-up power to the 26% of our grid that is powered by renewables. Because it is not always windy or sunny, coal is very important to these utilities. Utilities built a large number of coal plants as a result of the 1970s oil embargoes. President Jimmy Carter, stung by that generation's problems with the Gulf producers, hailed us as the Saudi Arabia of coal and pressured utilities to build coal plants, hence the 50% of the grid that was coal. But coal plants have a 40- year useful life. B & W's business includes building and servicing, but the former pretty much went out of style when the plants did, and the radical phase-out of coal as a power source crushed the company. But now the president and Energy Secretary Chris Wright are working hard to keep coal alive as part of a robust grid. Wright is using his executive authority to block the closing of coal plants. He has used his emergency power to stop the closure of coal plants in Michigan, Indiana, Colorado, and Washington. He cites demand from data centers as the chief reason. Plus, the president has repeatedly stated that he favors coal and is against renewables. He's slashed programs and loans needed to phase out coal. He's basically the anti-Biden on the issue. Coal is inefficient. It's dirty. It's thought to be the all-around worst form of energy in terms of any parameter, especially expense. But it's got powerful backing, including that of the National Coal Council, a federal advisory committee that is headed by Jim Grech, who also happens to be the CEO of Peabody Energy , the largest coal company in the U.S. Convenient. The National Coal Council, which had been abolished under President Joe Biden, advances coal's interests as a national security issue to power the data center revolution. And it obviously furthers the interests of Peabody Energy, Core Natural Resources , and Alliance Resource Partners . If Wright succeeds, I like all three of these. Why am I somewhat circumspect? Because there are court challenges to Wright's efforts. But if Wright is successful, these are all buys. Moreover, Core Natural has an export terminal in Baltimore that could ship coal to other countries hit by soaring natural gas prices. All three have things going for them. Obviously, Peabody, with the CEO serving as the head of the National Coal Council, can be a huge winner. Unlike most even peripheral data center ideas, it's down 20% for the year and is very cheap, seven times next year's earnings. Core Natural has been relatively flat and trades at 11 times next year's earnings. Alliance is up 8% for the year, but has a dividend yield of 9.5%. Not Late. Maybe even early. Now it takes a huge pill to swallow to back coal. Maybe you would rather get behind Altria or Phillip Morris International , although both have had monster moves. I went to hear a presentation from Phillip Morris, and it was a radically anti-traditional tobacco, as befitting a company that makes smoke-free products. But I digress. If you are thinking about buying a coal stock, you have to believe that coal plants will survive, that President Trump will finish his term, and Vice President JD Vance runs and wins. Maybe twice. That's because the utilities are so against coal. They have always been sensitive to sinking any more money into coal because they fear a Democrat president will thwart their plans. Remember, B & W traded at a dollar because it was pretty unthinkable that the president would be so pro-coal. But the data center, as a national security story, has resonance, so coal has relevance. Now, I want to go back to first principles. The data center story has tentacles far and wide. It has implications for many industries because the utility business is gigantic, with virtually unlimited capital and a desire to expand; more ratepayers mean more money. These are publicly traded entities that thrive on growth, which is why they are incredibly strong. They have come down a lot because of the rise in interest rates, but if you want to think they can turn, then choose Sempra , Southern , or American Electric Power . I come back to the data center story again and again because there are so many variables and so many ways to make money. Coal and natural gas keep emerging as avenues to prosperity. So do Eaton , GE Vernova, Vertiv , and Caterpillar . (Late, late, late?) Maybe B & W fits your speculative parameters. Maybe you want a more assured 9.5% yield of Alliance Resources. Or maybe you just want the potentially newfound growth and inexpensive stocks of Peabody and Core Natural. Or perhaps you just stick with Nvidia and own it through this week's earnings instead of trading it. (See here for a full list of the stocks in Jim Cramer's Charitable Trust.) As a subscriber to the CNBC Investing Club with Jim Cramer, you will receive a trade alert before Jim makes a trade. Jim waits 45 minutes after sending a trade alert before buying or selling a stock in his charitable trust's portfolio. If Jim has talked about a stock on CNBC TV, he waits 72 hours after issuing the trade alert before executing the trade. THE ABOVE INVESTING CLUB INFORMATION IS SUBJECT TO OUR TERMS AND CONDITIONS AND PRIVACY POLICY , TOGETHER WITH OUR DISCLAIMER . NO FIDUCIARY OBLIGATION OR DUTY EXISTS, OR IS CREATED, BY VIRTUE OF YOUR RECEIPT OF ANY INFORMATION PROVIDED IN CONNECTION WITH THE INVESTING CLUB. NO SPECIFIC OUTCOME OR PROFIT IS GUARANTEED.
Four leading AI models discuss this article
"BW's coal angle is more political speculation than durable catalyst given legal hurdles and utility resistance."
The article argues surging data center demand (potentially needing 90-100 GW) plus Trump/DoE policies will revive coal, lifting Babcock & Wilcox (BW) via its $2.7B backlog including a $2.4B Applied Digital-linked deal and its coal-plant servicing edge. Yet it glosses over material risks: 32% short interest in Applied Digital, its heavy losses, the B. Riley relationship under SEC scrutiny, and explicit court challenges to Wright's emergency orders blocking closures in multiple states. Utilities remain structurally opposed to new coal capex. BW's natural-gas plant option may be the cleaner near-term play given GE Vernova capacity constraints, but overall coal exposure still faces decades of phase-out momentum.
Successful lawsuits or a post-Trump policy reversal could force accelerated coal retirements, turning BW's legacy coal capabilities into stranded assets and erasing any revival premium.
"B&W's valuation depends on Applied Digital execution risk and coal's political survival—both binary, both reversible, and both already partially priced in by short sellers."
The article conflates three separate theses—B&W's backlog, coal's political revival, and data center power demand—without rigorously stress-testing any. B&W's $2.4B Base Electron contract is real but contingent on Applied Digital (APDN) executing a $12B market cap bet while burning cash and facing 32% short interest. The coal thesis hinges entirely on Trump+Wright blocking plant closures via executive authority—legally fragile and reversible. Most critically: even if coal survives politically, utilities are economically incentivized to build natural gas (cheaper, faster, cleaner). B&W's exposure to coal revival is speculative; its natural gas capacity is the actual moat, but GE Vernova already dominates that market.
If courts block Wright's coal plant interventions (likely given Clean Air Act precedent), or if Applied Digital's backlog proves unexecutable, B&W's $2.7B backlog evaporates and the stock reprices to its historical $1 valuation. Coal stocks like Peabody (BTU) trade at 7x forward earnings because the market has already priced in a low probability of sustained coal revival.
"The regulatory push to revive coal ignores the reality that institutional capital and utility operators prioritize long-term levelized cost of energy (LCOE) over political mandates, making a coal renaissance economically unviable."
The thesis that AI energy demand will revive coal is a classic 'regulatory arbitrage' play, but it ignores the fundamental economics of the grid. While Babcock & Wilcox ($BW) and coal producers like Peabody ($BTU) may see short-term sentiment shifts from political mandates, they face a massive 'Stranded Asset' risk. Utilities are moving toward natural gas and SMRs (Small Modular Reactors) because they are cheaper to operate and face less litigation than coal. Betting on coal requires ignoring the ESG mandates of institutional capital and the reality that coal plants are inefficient for the high-uptime, low-latency requirements of modern hyperscale data centers. The 'easy money' in $BW has likely already left the building.
If the Department of Energy successfully classifies coal-fired baseload as a 'national security asset,' utilities could be forced to accept government-subsidized capacity payments that make coal unexpectedly profitable regardless of market efficiency.
"Coal’s revival is unlikely to be durable; the data-center tailwind can boost near-term sentiment, but policy, fuel-switch incentives, and financing realities cap long-run upside for BW and related coal plays."
The article leans on a data-center power surge to justify a coal comeback and to validate Babcock & Wilcox’s stock move, but it glosses over key headwinds. Even if 90–100 GW of new capacity is needed, utilities are reluctant to finance new coal amid emissions targets, carbon pricing, and a longer-run shift to gas and renewables. The 2.4B backlog with Base Electron looks lumpy and highly dependent on Applied Digital’s trajectory and CoreWeave, not a durable earnings stream. Governance and execution risks (short interest, B. Riley ties) cloud the upside. In short, the data-center story matters, but a lasting coal revival remains doubtful.
Strongest countercase: even with rising demand, coal economics and policy risk make a durable comeback unlikely; the thesis rests on a temporarily favorable political window that could close, and the backlog may not translate into multi-year earnings.
"Data center demand may extend existing coal operations, supporting BW servicing even without new builds."
Claude correctly flags the contingent nature of the $2.4B Applied Digital contract, yet underplays how sustained data center demand could force utilities to extend coal plant lives regardless of new capex. Court challenges to Wright's orders create uncertainty, but a successful DoE national security designation, as Gemini notes, might override economic preferences for gas. This linkage between policy and grid reliability remains the untested variable for BW's servicing revenue.
"B&W's servicing revenue from delayed coal retirements is real but capped; it's not the same as capturing new data-center baseload demand."
Grok conflates two distinct revenue streams: new coal capex (utilities resist) versus servicing legacy plants (defensible). Wright's emergency orders don't force utilities to *build* coal—they only delay closures. Extended plant life ≠ new backlog. B&W's real moat is maintaining existing capacity, not capturing the 90–100 GW delta. That's a much smaller TAM than the article implies, and it doesn't justify current valuation multiples relative to historical coal servicing margins.
"Hyperscalers bypassing utility-scale grid constraints to acquire distressed coal assets directly creates a niche service market for BW that ignores traditional utility ESG resistance."
Claude is right that servicing isn't building, but both of you miss the 'behind-the-meter' loophole. If hyperscalers like Amazon or Microsoft bypass utility-scale grid constraints by purchasing distressed coal assets directly to power data centers, the regulatory resistance from utilities evaporates. This shifts BW’s model from a legacy service play to a private-grid infrastructure play. The risk isn't just policy—it's whether BW can pivot to private, off-grid industrial servicing where ESG mandates don't apply.
"A hyperscale private-grid pivot to coal is unlikely to materialize BW's upside; BW's real risk is legacy servicing and backlog, not a private-grid workaround."
Gemini’s 'private-grid' pivot premise hinges on hyperscalers buying distressed coal assets to power data centers, which I view as highly speculative. Even if pursued, it faces financing, regulatory, and ESG constraints that utilities don’t, likely shrinking BW’s servicing moat rather than expanding it. The backlog’s value rests on authorized capacity and execution risk, not a private-grid shortcut. In short: don’t overattribute BW upside to a private‑grid workaround.
The panel consensus is that Babcock & Wilcox's (BW) exposure to a coal revival is speculative and risky, with utilities economically incentivized to build natural gas plants instead. The key opportunity lies in BW's natural gas capacity, but even that faces strong competition from GE Vernova. The single biggest risk flagged is the high short interest in Applied Digital, a key partner in BW's backlog, and the uncertainty surrounding Wright's emergency orders blocking coal plant closures.
BW's natural gas capacity, if it can compete with GE Vernova
High short interest in Applied Digital and uncertainty surrounding Wright's emergency orders