Madison Air Is Being Punished for Winning: The Data Center Mix Tax Meets a De-Rating Tape

Orders +26%, backlog +119%, guidance up — and a margin line that fell because the fastest-growing business is the least profitable one. The market has spent 90 days selling exactly that supply chain.

MAIR · Earnings Call · 2026-11-05

The success-based mix tax

Madison Air Solutions is a recent IPO still teaching the market what it is — Jill Wyant opened the call by saying she'd "start with an overview of our business and strategy" in an effort to continue to educate investors. What the education revealed is an unusual tension: the company's greatest growth engine is also the thing compressing its margins, and the equity tape has already decided how it feels about that engine. Third-quarter net sales rose 25% to $1.127 billion, adjusted EBITDA rose 19%, and the margin printed 26.5%. The CFO was blunt about the cause: “Margins were 26.5%, with essentially all of the year-over-year decline driven by the success-based mix from outsized growth in Nortek Data Center Cooling” — JJ Foley, Chief Financial Officer · 2026-11-05. In the commercial segment, revenue grew 39% organically to $772 million while segment margin fell to 26% against 28% a year ago. The mechanics matter. success-based mix is not a euphemism for broad decay — it is arithmetic. “the data center business at its core right now, as it's early in the scaling journey, is lower margin” — JJ Foley, Chief Financial Officer · 2026-11-05. Strip Nortek Data Center Cooling out and, per management, company margins are up year to date. This is a portfolio being diluted upward by its best opportunity.

The tape disagrees with the backlog

Here is where the story gets genuinely interesting. Madison's data centers exposure is its number-one keyword by a wide margin, with record backlog of $3.2 billion (up 119%), a book-to-bill of 1.26x, and orders up 26% in the quarter and 33% year to date. Over 70% of that backlog converts in 2027 or later. And yet the global theme tape shows the entire AI data center supply chain in a violent 90-day drawdown: AI data centers is among the largest decliners across the whole market, with dozens of negative constituents — optics, memory, power, co-packaged anything. Data center revenue baskets, high bandwidth memory, co-packaged optics and PCIe Gen 6 names all sit deep in the red over both 90 and 30 days. So Madison is guiding up into a theme the market is actively de-rating. That divergence is the crux. It explains why an analyst opened the Q&A asking management to define "success-based mix" — the phrase is doing a lot of load-bearing work — and why the very next questioner went straight at whether data center margins actually inflect in 2027. The answer was a roadmap, not a number:

for the data center business in isolation, I would tell you we see a few points of margin expansion over the course of the next several quarters... The current margin dynamic is largely a function of program maturity.
JJ Foley, Chief Financial Officer · 2026-11-05

JJ Foley quantified the maturity curve: hours on new programs are already down roughly 25% from where they started, Oklahoma City was built slightly ahead of volume for absorption, and services follow the installed base. That last leg is the real long-term earnings power argument — install base plus aftermarket, currently only about 10% of total company revenue, growing at or above equipment.

The election, the moratorium, and what Madison says it isn't

Asked about data center moratoriums and the freshly-decided elections, Wyant gave the call's most defensible answer — a definition of what Madison deliberately is not:

We've not gone into the neoclouds. Frankly, we don't have the capacity, and we're not going to invest out in front of it. We've not pursued speculative demand.
Jill Wyant, President and Chief Executive Officer · 2026-11-05

She noted cooling equipment arrives at the tail end of construction, that more than 80% of the U.S. megawatt build-out underway is permitted or under construction, and that orders outside data centers still grew more than 20% with the majority of fifteen commercial end markets expanding. That diversification claim is the ballast under an otherwise concentrated narrative — and it is supported by the Big Ass Fans result, the strongest commercial grower this quarter, framed around a "Heat Tax" and a $6 billion market that is only 4% penetrated.

The quiet parts: capacity discipline, price-cost lag, and residential

What management chose not to do is also informative. Despite the growth, Wyant said the 2027 plan contemplates no massive capacity build: “we are not contemplating massive investments in capacity... We can accommodate the growth, the responsible growth, within the four walls of the capacity we have today.” — Jill Wyant, President and Chief Executive Officer · 2026-11-05 That is a grow-to-invest posture, not an invest-to-grow one — a genuinely differentiated claim for a company with this order book. On price cost, the language was carefully hedged: dollar pressure covered in year one, rate coverage in the following year, with a few on the longer cycle side of things that'll recover here as we get into 2027. That is a recovery deferred, not delivered. Residential is the overlooked engine. Net sales rose 3% to $356 million in a soft housing market, but adjusted EBITDA rose 12% to $105 million with margins up 225 basis points — driven by productivity, tariff refunds, and AprilAire synergies that Foley admitted are “over-delivering relative to the original underwriting case” — JJ Foley, Chief Financial Officer · 2026-11-05, with commercial revenue synergies outrunning cost synergies. A 29%+ margin business growing through penetration rather than housing starts is not a cyclical story.

ebm-papst: the leverage trade

The pending ebm-papst acquisition — roughly $5.4 billion enterprise value, funded with a ~$2.25 billion PIPE plus ~$2.8 billion of debt and cash — is where the two narratives collide. It takes combined sales to about $6.9 billion, expands the addressable market from $40 billion to over $70 billion, brings a 250-million-fan installed base, and carries a $160 million cost-synergy target exiting year three. Management says ebm-papst is tracking to ~$2.8 billion of 2026 revenue and ~$343 million of EBITDA, possibly with upside. But recall the August reaction: the equity fell 16% in a week on announcement. Investors disliked the check size. And now the company is asking them to believe in a deleveraging story — net leverage at 2.6x excluding PIPE proceeds, sub-2.5x by year-end — while simultaneously adding a multi-billion-dollar acquisition that will lever the balance sheet back up at close. "A clear path to organic delevering" and a large debt-funded deal in the same breath is a promise that needs 2027 to grade it. Meanwhile the guidance raise ($4.05–$4.125 billion sales, $1.04–$1.07 billion EBITDA, ~26% margin) and $298 million of year-to-date free cash flow at 135% net-income conversion are real. So is Wyant's closing frame: Return on Air as the repeatable playbook. The stock is unlikely to trade on the playbook. It will trade on whether Nortek margins inflect before the market finishes de-rating the theme that made them look so unattractive — and on whether Q4 book-to-bill, which Wyant guided to merely "in the range of one" from 1.3x this quarter, holds the story together.