Solstice Is Selling an AI Materials Story. The Market Isn’t Buying.

A raised guide, a $2.6 billion nuclear backlog, and the first real sizing of data-center sales — against a stock down 31% in 90 days and a refrigerant recovery pushed into 2027.

SOLS · Earnings Call · 2026-11-04

The market disagrees with the press release

Solstice’s third quarter was, on its face, the kind of print a newly spun company dreams about. Net sales rose 8% to roughly $1.05 billion, adjusted EBITDA rose 13% to $265 million, and adjusted EBITDA margin expanded to 25.3%. Management raised full-year guidance for the second consecutive quarter. It had already bought back $100 million of stock in the quarter and intends to substantially complete the remaining $400 million of its authorization by the end of the first quarter of 2027. CEO David Sewell was blunt: “We think this is a great time to buy with where the share price is at” — David Sewell, CEO · 2026-11-04. The stock disagrees. Solstice is down 31.1% over the last 90 trading days and 34.9% from its June 30 peak of $88.62. For a $9.8 billion specialty-chemicals spin, that gap between operating momentum and equity performance is the story. The immediate culprit is the tape. The global market has stopped paying up for AI infrastructure exposure: AI data center names are down about 21% over 30 days in the Know Trend price history, and data center capacity proxies are down almost 39%. Solstice does not need to be an AI pure-play to get caught in that de-rating. Its new pitch leans heavily on thermal management and data center growth, so the market is applying an AI-infrastructure lens to a company that still earns most of its money from refrigerants, nuclear conversion, and healthcare packaging.

What genuinely changed: disclosure, nuclear, and the end of the loan drag

The most important new fact is quantification. Management had repeatedly declined to size its data-center exposure. On the prior call: “it's still a smaller piece of our overall refrigerants, which is why we just haven't broken it out” — David Sewell, CEO · 2026-07-30. This quarter it put a number on it: “data center and AI related applications represent approximately 6% of our expected 2026 revenue with sales up more than 50% versus last year” — David Sewell, CEO · 2026-11-04. That is a real disclosure change, and it makes the growth algorithm legible — roughly $250 million of 2026 sales tied to data-center and AI applications, growing north of 50%. Nuclear is the second change.

Our nuclear backlog reached $2.6 billion in the third quarter, with contracts now extending through 2039.
David Sewell, CEO · 2026-11-04

That is up from the $2.2 billion discussed earlier. The nuclear backlog matters because it turns a politically popular theme into contracted revenue. The Metropolis debottlenecking now supports more than 10,000 tons this year, with an additional 1,000 to 2,000 tons at limited incremental capital cost. Almost all product-loan returns are behind the company, removing a headwind that depressed 2026 comparisons. The Spokane facility expansion for sputtering targets is the other capex centerpiece — over $200 million committed, with demand so strong management admits “we are a little bit constrained right now” — David Sewell, CEO · 2026-11-04. Electronic Materials had a record quarter: sales up 31% to $135 million, segment EBITDA up 44%, and margin up 453 basis points. That is not boilerplate. It is the clearest evidence that the leading-edge semiconductor and data-center materials story is showing up in numbers rather than just slides.

The crack: refrigerants slipped a year

The dissonance is in refrigerants. Sales rose 8% to $433 million, but the 454B transition remains messy. On this call: “we definitely see that 454B headwind continuing. But then as we go into 2027, we expect normalization in that market” — David Sewell, CEO · 2026-11-04. Six months ago the framing was different: “we expect the full transition to be done in early 2026. And then the aftermarket should kick in” — David Sewell, CEO or President · 2026-02-11. Now the aftermarket recovery is a 2027 event. That is a classic push-to-the-right. The normalization keyword is doing a lot of work in this transcript, and investors have heard this movie before. Healthcare packaging’s 24% growth is similarly flattered by lapping last year’s destocking; management itself calls the underlying business mid-single-digit. So the headline 8% organic growth is a blend of genuinely strong electronics, a still-growing but delayed refrigerant business, and a comp-driven healthcare bounce. That is decent — not the clean acceleration the press release implies.

Margins, leverage, and the standalone test

The operating story is improving but not fully healed. In the last filed quarter before this call, total revenue was $1.1 billion, up 11% year over year, yet operating margin was 15.3%, down 3.9 percentage points. Interest coverage had fallen to 7.7x from 188x at its post-spin peak as the company stood up its own balance sheet. Q3’s adjusted EBITDA margin of 25.3%, up 109 basis points, is the first clean evidence that the 25% commitment is holdable. Free cash flow was $118 million in the last filed quarter, up 74%, and Q3 operating cash flow was $693 million through nine months. But the strategic question remains. The Element Solutions merger was mutually terminated; management now says no transformational M&A is on the horizon and emphasizes the organic nature of the portfolio. That is a pivot from the prior M&A optionality language. The $500 million buyback, with $400 million left, is the clearest signal that management sees the shares as undervalued. It may also be the only near-term catalyst the market is willing to underwrite. The bull case is not fake: a 6%-of-sales data-center franchise growing more than 50%, a $2.6 billion nuclear backlog, a constrained Spokane asset, and a 25% margin. The bear case is that the AI-infrastructure tape has broken, the refrigerant recovery slipped, and the standalone company still has to prove it can compound without a deal. The stock’s 35% drawdown says the market is pricing the second story. The next two quarters will tell whether the first is big enough to change the narrative.