Sunlight Partners

Why is a 160-year-old boiler maker Babcock & Wilcox proposing to build a high-cost, high-emissions, 1980s style power plant for data centres? Sunlight Partners is short shares of $BW; See our report and disclosures here sunlight.partners/disclaimer…. 1/5

We're delighted to be featured in this week's Hedge Fund Alert, published by With Intelligence, by @SPGlobal. The piece covers our launch, our debut activist campaign targeting Babcock & Wilcox, and the philosophy behind our work: uncovering environmental risks the market has misunderstood. More to come. 🌱

Our first Impact Statement explains the thinking behind our Babcock & Wilcox ($BW) campaign, and how we assess market impact. You can read it in full here: mcusercontent.com/19133e49a0…

4/6 $BW’s share price implies they will make 9-10 boiler plants by 2046, which seems to only make sense if the gas turbine shortage persists for two decades.

2/6 CCGT plants don’t use $BW boilers, they use a heat recovery steam generator (HRSG) to make steam from gas turbine exhaust. In very simplified terms – a $BW boiler makes steam like a saucepan on a gas stove. An HRSG makes steam like a car radiator with the cap off – absorbing engine heat (not combustion) into a series of tubes filled with water. Here are pictures of both units:

1/6 One of the most confusing things to us about $BWs strategy (and why we are SHORT $BW) is their new plan to build 1980s steam boiler plants then transform them into modern CCGT plants at a later date. You can hear $BW CEO discuss the plan on @CNBC with @jimcramer here starting at 5:10: youtube.com/watch?v=Fpx28Uon…

6/6 If the gas turbine shortage normalizes – why would the market want to build a series of redundant $BW boiler plants sitting alongside abundant modern, efficient CCGT plants?

5/6 And to convert $BW's proposed plant to CCGT means getting a modern gas turbine – the exact piece of equipment that is supposedly in such chronic short supply that we’re rebooting 1980s boiler power.

3/6 Building a new CCGT plant around an archaic steam boiler plant means you could re-use the steam turbine (which $BW doesn’t make), but the boilers (that $BW makes) would likely be switched off or sparingly used due to cost.

There is a reason no US utility has built this style of gas steam plant in 41 years, and why no other data centers or boiler makers are proposing to build one now. It makes no sense commercially and is a huge backwards step environmentally for the US power sector. Optimistic analysts believe $BW can build 10 of these plants in the next 20 years; we doubt they'll build one. 5/5 sunlight.partners/disclaimer…

A new plan to add gas turbines to this archaic plant at a later date to create a “Combined-Cycle Plant” seems illogical – because Combined-Cycle plants don’t use boilers. Why is it even being considered? Related parties $RILY, $APLD, and $BW developed the project in a sharp pivot to "AI Power" in late 2025, fueled by circular investments, warrant incentives and hype. Until then, its core business was servicing remaining 1950s-60s era power plants for parts and maintenance. Its share price has since jumped 400%, and major investor $RILY has been dumping stock. 4/5

Beyond that, it would compete in the oversupplied North Dakota market, a net exporter of electricity that has abundant wind + battery storage and is building two large and modern Combined-Cycle gas plants already. The company claims it has a “Speed to Power” advantage because steam turbines are currently easier to acquire than modern gas turbines, but this is undermined by the realities of power plant permitting; Claims of starting operations by 2028 are absurdly optimist and twice as fast as permitting and construction timelines for recent (modern) gas plants in North Dakota. 3/5

The laws of thermodynamics mean $BW's 1980s gas steam tech (Rankine Cycle boiler plant) burns 40% more gas and emits 40-50% more CO2 to deliver the same power as modern Combined-Cycle gas plants. The proposed plant would be grid connected In North Dakota, costing $2.4 billion to build and potentially operating for 30-40 years. 2/5