Madison Air and it's Big Ass Fans

Sup, BIG DAWGS, it's me, Gabe Azar, Head of Burrito Rolling, intern at Azar Capital Group, and Head of Research and Chief Morale Officer of Azar Research Collective. Your third-grade teacher's favorite investor. Currently writing to you nerds from my L desk. Today, I am writing about Madison Air. Now sit back and enjoy my crappy writing. And remember, buy low and sell high, my friends.

The boys over at Azar Research Collective are an honest bunch, so let's be honest: Madison Air is a company in our wheelhouse. This report would not have been written and shipped for much longer if Madison didn’t own Big Ass Fans. But here we are, getting ready to rip a report on Madison Air. Ask 100 people if they are fans of Big Ass Fans, and they will all probably say “yes,” but probably only 8 of them would know that it's the name of a company. 

Companies like Madison Air and their subsidiaries are interesting because nobody thinks about these fans and cooling units until they break or people start complaining about how hot or cold it is. But this is something you only know or think about if you’ve worked in one of those situations where the AC is broken, and you are relying on small hand fans or gusts of wind to keep you cool.  

What fascinates me with this AI capex industrial reshoring wave is that there are many downstream effects that this spend is having on the rest of the economy that people have yet to realize. I am sure data centers and data ranches have always been customers of Madison Air and their competitors, but now, as the build-out ramps up their amount of spend the hyperscalers are spending with Madison Air and other industries across the supply chain is rapidly increasing. This opens whole new revenue sources for many companies that analysts have yet to add to their models. While the initial sound of increased defense and infrastructure spending by the government may sound scary, the downstream effects are what built this country. 

I am trying something a bit new where I riff a little before spewing nonsense about the company. Sorta like a cold intro, with some more personal thoughts and feelings on the company and what's happening within the industry and game before we laser in on Madison Air or the company we are diving in on. Despite selling them to Big Ass Fans, Madison Air is betting heavily on commercial sales. Madison Air currently has a brand portfolio of 33 names, giving them massive exposure to both residential and commercial customers. 

These customers operate in advanced manufacturing, data centers, education facilities, healthcare and life science labs, hospitality and retail buildings, industrial units, government buildings, logistics centers, and your local recreation center. Madison Air’s management noted that they are seeing slowed growth in their residential unit while commercial is seeing strong growth. Either way, we are getting ahead of ourselves; this is still the riff section. I also got word that Madison Air was Chicago’s largest IPO of the year (so far), so hell yeah. Let's rip.

Madison Air Solution Corporation is an air quality solutions company based out of Chicago. The company serves two customer segments: commercial facilities and residential single/multi unit homes. Its commercial segment sells custom-engineered air handling, air, liquid and hybrid cooling, air movement and heat, humidity control, and energy efficiency systems into data centers, healthcare, cleanrooms, advanced manufacturing, logistics, and education; this segment did nearly $2.5 billion in sales in the past twelve months. The residential unit did just under $1.4 billion, selling ventilation, purification, humidification, dehumidification, and sensor units through its AprilAire and Broan-NuTone brands. Currently, Madison Air employs almost 10,000 people and derives nearly 95% of its revenue from North American operations, with 50% from replacements and upgrades, 40% from new builds, and the remaining 10% from services and aftermarket sales.

Larry Gies founded Madison Air in 1994. While the History of Madison Air (The Public Company) would only be a several-page book, the history of its core assets and the private equity fund would be a thousand-page epic, as many of them have been through hell; some of them have now been through three leveraged owners in the past twenty years. In 2009, Nortek, which also owned Broan-NuTone and Reznor, filed for Chapter 11, erasing nearly $1.3 billion in debt after the housing crash. Nortek was acquired in 2016 by Melrose Industries, which operated the company and sold its Air Management unit in 2021 to Madison Air for $3.6 billion. In 2021, the company acquired its holy grail (my opinion),  Big Ass Fans, for $1.1 billion before going on a nice spree, acquiring AcoustiFLO, AprilAire, and divesting Nortek’s HVAC unit. One of today’s core differentiators in these business revenue models is the commercial mix. 

Madison Industries is the private equity / permanent capital industrial enterprise that Larry founded. Since its founding, they have scaled businesses across filtration, safety, and the air business. The firm's model is buying founder-led and owned businesses that are leaders in their market niche. This is how Madison Air was forged, eventually taking it public, raising $2.6 billion; proceeds of the IPO went to creditors. You can view this IPO more as a deleveraging event than a traditional exit. Madison’s mission is to “make the world safer, healthier, and more product through more productive through better air”, whatever the hell that means. Madison believes that air quality is a productivity asset with a measurable ROI, or as they call it, “Return On aIr.” Madison has a simple four-point strategy: focus on growing end markets, outgrow customers through channel and product innovation, an 80/20 operating model on cost and simplification, and allocate capital towards bolt-on M&A deals. 

Madison Air believes they will eat well from the newfound data ranch demand, noting that there is roughly $28 billion TAM that has been unlocked since 2021. We believe TAM is nothing more than a number made up by the marketing nerds to raise hype in shareholders and investors. The bad boys over at Azar Research Collective and Azar Capital Group care about the Total Addressable Problem: what is the size of the problem Madison is trying to solve, and within that problem lies the capital they desire. Madison Air Industries has the backlog, orders, and organic growth that is coming with this wave, which is an excellent sign as it shows Madison is putting in the work. Madison Air has nearly 600 R&D employees, showing that they are heavily invested in improving their products as the demand for tier 1 cooling blows up. 

Madison’s core business offering is custom air solutions for commercial buildings, where an engineer would need to go around writing the specs of the building and the customer demand before building out their unit. The residential unit is more of a plug-and-play, as homes are more similar and require less customization than the specific needs of a hospital, laboratory, or cleanroom. The data center came at an excellent time for Madison as they recently acquired Nortek Air Management solutions in 2021 just before the AI hype wave. Despite the new business opportunities that are coming from Data Centers, analysts note that its data center cooling segment is rapidly growing and now represents nearly 13% of the company's total revenue. This is not something we are too worried about; this is a new and rapidly growing segment for the industry, so this sort of growth will likely be followed by a plateau and slight decline as demand normalizes in several years. 

Madison Air is seeing a slight decrease in organic residential sales, most likely due to economic concerns. We think that the commercial customers in the long term are worth more to Madison due to the highly technical nature of the products they install on commercial builds vs a single-family home. This is off the top riff, but we think that there could even be value in selling off its residential unit. This would load Madison Air’s war chest with capital for R&D, capex investments, and continued M&A. Doing this would also enable them to focus on the higher-value customers and the emerging data center economy, and maximize their market share early in the game rather than having to fight as equals with the rest of the big dawgs that also operate both residential and commercial units. 

Madison currently has a large backlog, something that the CFO stated: “We are delighted with our orders and backlog performance in the second quarter” in their most recent earnings call. This was said after being asked about their pricing power, and margins looked like within this backlog. This is an important and valid question, despite companies never disclosing margins on a by order case. Management stated earlier that the data center growth is just beginning its margin journey, although they guided their commercial unit to have margins near 28% by the third quarter. This is an excellent sign that they are investing heavily currently in building capacity to fulfil demand they’ve already won. So once the builds are done, and they ship units, margins will be ripping (in an industrial sense). 

We think four currently smaller pieces of the revenue pie and technology could matter a lot more in a few years due to Madison's emerging appearance within the commercial space. First is liquid cooling; AI hardware produces more heat per rack than only air can “remove”. So now Madison Air sells a coolant distribution unit, something that a company can easily begin to start using. Next is their services and aftermarket revenue, currently 10% of the company's revenue, so a fairly decent amount. This is higher-margin revenue, as the issue is more digital than mechanical, as the company built out a mobile app for parts which has seen massive amounts of traction for self-service channels. 

Third is sensors and digital controls; Madison Air's entire commercial thesis is that air produces measurable results. Proving those results requires monitoring, and if that data is any good, their customer may upgrade. This shifts Broan and AprilAire from equipment dealers to platform sellers.  Lastly is FrostShield, a patent-pending defrost technology that extends the operating range of air source pumps and removes the need for the unit to enter a defrost mode. FrostShield is sold under the Addison brand, which is dedicated to outdoor air systems. Management thinks this technology will increase Addison's sales and overall momentum. 

As Madison Air recently just IPO’d, there are some traditional boring “catalysts” like the IPO lock-up period. Although more interesting is hyperspace capital spending guidance, which will send a durable demand the company can begin to rely on for supply chain forecasting. Hopefully by the end of the year Madison Air is able to report its guided 28% margins within its commercial unit and bring net leverage to below 2.5x. 

The business of air, like an onion, is layered with each layer or segment having different returns. The bottom or most inner layer of the onion or industry is heating and cooling. Think of furnaces, air conditions, rooftop units for smaller buildings, this layer is fairly price competitive. Madison Air left this segment of the market in 2024 when they sold Nortek Global HVAC. The next layer is more engineered and focused on ventilation and residential air quality with things like bathroom fans, change hoods, whole home humidifiers. In this segment brand awareness and compliance matter a ton as much of the business is tied to housing. So the customer is buying peace of mind. The third layer is the most valuable and mission critical segment. This includes hospital cleanrooms that require controlled pressure with a set number of air changes per hour, data centers need heat removed continuously or the gear will fail. All this customer cares about is uptime, engineering decides the winner not the price. Since 2023 demand has been rolling towards the outer layer due to AI computing cooling demands.

The marketing and third party research bro’s valued the data center liquid cooling market at just under $5 billion in 2025 and are expecting it to reach $27 billion by 2027. This is a massive jump, something we are unsure of as the broader data center cooling market is estimated to compound at nearly 13%. You can treat those numbers as a truly best case scenario on creatine, while the demand spike is real, I do not think the communities of our nation are enabling the data ranches to be built out at that rate. Madison Air’s own management expects that total data center to reach nearly $40 billion in the coming years. They also have a massive backlog within this unit that is growing 133% You, its confirmed orders are more than double than what the company has capacity for. In Madison Air’s 10-Q they even note that their backlog has grown from $1.23 billion to $2.86 billion since June 2025. 

In the mission critical segment purchasing decisions are not made over a single cold call, they run through several parties. Often starting with what the building owner wants and requires, then consults an engineer to write the specs, then it's the mechanical contractors turn to price the installation, then it's the manufacturer turn to install. The power sits with whomever the engineer trusts and has done business with in the past. An engineer who has done twenty projects with Nortek Air Handlers is a lot less likely to switch to a competitor on their 21st project to save a couple bucks. The Hyperscalers have changed this dynamic and have gone directly to the manufactures to co-developed equipment rather than having the engineers do it all. This enables them to go faster, deep, and iterate faster but it also means one customer holds the leverage. 

The data center cooling demand is currently split into three customer buckets, each has its own growth rate, margin profile, and fragility. The newest and fastest growing customer is the hyperscaler, they buy large facilities, scale programs, co-develop gear, negotiate pricing, and order a ton of products. Next is Colocation customers, meaning a multi tenant building owner renting out space to others. This is a smaller segment that requires more standard specs and a broader vendor list, these customers are focused on price which is bad for long term value but good for product mix diversification. The third and shrinking pool of the three is enterprises running their own server rooms from either their office or from the folks from the previous sentence. Management noted that they are focused on the hyperscalers and colocation customers as this strategy is chasing stronger demand. 

This demand is mostly coming from AI workloads, reshoring, building code changes, and aging buildings that need new / upgraded shit. AI workloads generate more heat per square foot than anything else. This is where the liquid cooling method was invented as air cooling before extremely different past 50 to 80 kilowatts per rack and AI clusters are way beyond that. New American manufacturing construction has hit recent historical highs.  The demand for air products has been high because of the demand for sterile pharmaceutical facilities, semiconductor fabs, and other types of manufacturing facilities. There is a whole fly wheel here as they first buy the large air units then buy the industrial sized fans and they eventually become a long term customer. Building codes are also updating as old buildings age and new buildings are built, ventilation standards have been juiced up after the COVID pandemic. 

Madison Air has also begun to receive tariff refund payments thanks to the Supreme Court's decision in February which invalidated certain tariffs. Madison Air is noting these refunds as a reduction in cost of goods sold which slightly improves margins. Madison Air’s CFO noted on their most recent earnings call that the tariff drag is mostly impacting the companies canadian plants that ship into the United States. This is the most obvious reason for the non-US revenues seeing a 42% YoY decline. Nearly 95% of all Madison Air comes from within the United States, due to their strong market presence. A competitor importing cooling equipment pays taxes on the entire product while Madison Air and other domestic manufactures just need to pay the taxes on the important raw materials. 

You can only do so much with air. As chips get denser, there is only so much you heat you can remove before it gets impractical and that's where the liquid cooling method came to be. So the boys are now working on how they bring the cooling closer to the chip, first to the rack then to the heat exchangers, then to the processor. One of Madison Air’s new flagship products, Statepoint, is an indirect evaporate system that rejects heat from the building through evaporation rather than mechanical refrigeration. Madison Air teamed up with the Facebook infrastructure team to cook up that idea. Side note, that is an excellent win because they got to use the Facebook/Meta resources on a problem they can now solve for the rest of their customers. 

Nortek reported that it can cut data center water use by up to 90% against chilled water designs. Jill Wyant is so bullish on this technology stating in a recent earnings call that “this idea that air cooling is going to disappear is overstated. There is just a lot of thermal energy in a data center that needs to be rejected." This is an interesting note because Madison Air still sells both solutions, but data ranches have four layers of air cooling demands. The first layer is the building plant layer, which rejects heat from the facility and evaporative systems. Then the room and row layer which moves cooled air through the hall into CRAH units. The third layer is within the rack, liquid is carried to the rack from coolant distribution units. The final layer is at the chip layer, which Madison Air currently does not have a product offering for. 

There are several waves of new insurgents or competitors acquiring different layers of the cooling stack. A company called Liquidstack is going after the immersion cooling layer, which goes after the CRAC manufactures. Isotope, threatens the people who use air cooled technology within the server rack layer with their precision immersion cooling technology. Then there are companies like ZutaCore, CoolIT, and Accelsius that are working on chip layer cooling technology. ZutaCore is redefining cooling technology with its waterless, direct to chip cooling solution that was built for high performance workloads and is already being used by Nvidia, AMD, Dell, and Intel. ZutaCore is definitely a company worth following. Although consolidation between the new players has already begun, with Schnieder buying Motivair, Johnson Controls investing in Accelsius and KKR taking a stake in CoolIT. ZutaCore though has just recently raised a venture round which makes them a more valid option for Madison Air to look to acquire them. On the residential front, it's a bit quieter with smart home platforms and heat pump adoption leading the way. 

Fuel The Keyboard

This blog runs on caffeine and 10-k’s. Send a cup of joe my way to keep the words flowing! Support me here.

Madison Air is not the largest company in any of the market segments they serve, but they do have the capital, product line, and backlog to become a large player within the next few years. Vertiv is a full-stack threat for Madison Air as they are currently the leader in selling thermal and power equipment globally, plus they hold the largest share of liquid cooling. Schneider Electric has also been buying into the Data Center infrastructure segment with its recent purchase of MotivAir (bought 75% in 2025 and has plans to buy the other 25% in 2028).  Munters is Madison Air's closest comparison; they are a Swedish company that specializes in evaporative and indirect cooling. Munters competes for the same hyperscaler accounts Madison Air does with its StatePoint product. Trane, Carrier, Johnson Controls, and Daikin all sell conventional chiller plants into buildings. They usually go into business when the building occupant is looking for an easy plug-and-play workhorse. Vertiv leads the liquid cooling market with an estimated 11% market share. But Madison Air’s residential brands Broan-Nutone and both lead the market and have been category leaders for decades. 

Madison Air’s strongest moat is that when they win a deal on a data center, hospital, or cleanroom the engineer designs the system around Nortek dimensions and performance. This creates strong lock in as then customers will continue to buy Nortek services and the engineer gets more practice installing Nortek products. Madison Air / Nortek developed Statepoint with Facebook's (Meta) infrastructure organization several years ago. This product now gains massive inquiries about using Nortek’s laboratory and engineering team to help create their own cooling components. At this level Nortek is no longer just selling a product to a company, they are partnering with them and custom building that product, this creates extremely high switching costs for the hyperscaler. This is worth noting as Madison Air is seeing a large portion of their revenue coming from a single customer, this advantage doesn’t gradually end. It compounds as other hyperscalers see and replicate what's working. 

Madison Air’s flagship residential brands Broan and Nutone both date back to the 1930s. These businesses have been through it, they have been through every housing cycle, have had multiple owners, own bankruptcy, and still remains a go to name for electricians. Madison Air’s other residential brand names have also been around the block and have compounded sales at nearly 8% since 2007. Depending on the customer segment, Madison Air’s product line may have network effects characteristics at large enough install units plus trained engineers that have experience with Madison Air. Despite its size, Madison Air’s cost structure is not as strong as its competitors due to its 80/20 operating model which is an advantage in some areas but Trane and Carrier are able to buy raw products more cheaperly. 

Where Madison Air loses and their moat is shallowest is the large combo meals. When a customer wants power, cooling, and services under a single contract they often call Vertiv or Schneider instead of Madison. Madison products are often more expensive, better fit for customer engineering needs. On the pro side, barriers to entry and switching costs are increasing due to new entrants needing the engineering know-how, code and safety certifications, bonding capacity for mission critical projects, and a list of references they can use to acquire business. Switching costs are higher for commercial clients than a homeowner who needs a new bathroom fan. 

Financial Analysis (Written by my AI analyst)

Second quarter net sales were $991.3 million, up 20.9 percent from a year earlier. On a pro forma basis, which adjusted last year to include AprilAire as if the company had always owned it, growth was 14 percent, split into 11 points of volume and 3 points of price. First half sales were $1,915.0 million. Gross margin was 38.0 percent, down 30 basis points. Adjusted EBITDA was $265.8 million at a 26.8 percent margin, down 70 basis points pro forma but up 155 basis points from the first quarter. The blended margin hides the story, which lives in the segments and points in two opposite directions. Commercial margin fell 299 basis points. Residential margin rose 423 basis points. The growing segment gave up profitability and the shrinking segment gained it.

Total debt at June 30 was $3,053.7 million: $1,345.6 million of term loans at SOFR plus 1.75 percent, $1,035.0 million of senior notes at 5.9 percent due 2029, and $700.0 million of secured notes at 4.1 percent due 2028. Against $261.8 million of cash, net debt is $2,824.6 million. A detail nobody mentions: the company holds $1,200.0 million of notional interest rate swaps and collars against $1,345.6 million of floating rate term debt. Roughly 89 percent of the floating exposure is hedged, which means the headline SOFR linked coupon behaves much more like a fixed one. Anyone modelling rate sensitivity on the term loan is modelling a risk the company has largely already bought insurance against.

The asset side deserves attention it rarely gets. Goodwill of $3,314.8 million plus other intangibles of $3,146.0 million total $6.46 billion, which is 77 percent of the company's $8.35 billion of total assets. Goodwill and intangibles are what a company records when it pays more for an acquisition than the hard assets are worth. Here they are the accumulated residue of roughly $8 billion of deals. Property, plant, and equipment is only $361.5 million net. This company owns far more purchased reputation than it owns steel and buildings. Two other items matter. Raw material inventory rose 24 percent since December to $266.8 million, a deliberate pre-buy against tariffs and the production ramp. And the AprilAire seller earned out $87.0 million moved into current liabilities, so that cash leaves around May 2027.

First half operating cash flow was $156.4 million, slightly below last year's $160.4 million despite much higher earnings. The gap is working capital, which consumed $136.0 million, mostly receivables as Commercial sales grew. That is a normal problem for a fast growing manufacturer. Product ships before customers pay, so cash gets tied up in receivables. It becomes a real problem only if collections slow or growth stops abruptly. Capital spending was $16.5 million, or 0.9 percent of sales, leaving free cash flow of $140.0 million and conversion of 123.3 percent of net income. Trailing twelve month free cash flow was $429.8 million on an 11.5 percent margin. Free cash flow yield on the roughly $15.9 billion market value is about 2.7 percent.

The IPO raised $2.58 billion. The company used it to pay off $2.6 billion of term loans, then repriced what remained, cutting the rate on the surviving $1.35 billion by a full percentage point in June. The savings land in the second half. The company paid $180.7 million of cash interest in the first six months, when the full pre IPO debt was outstanding and the payoff itself created one time costs. Full year guidance is roughly $240 million. That leaves about $60 million for the back half, a third of the first half rate. Every dollar saved drops to pre tax income. Across roughly 507 million shares, it is real earnings growth that nothing in the trailing numbers reflects yet.

A pension settlement is coming. In June 2026 the company agreed terms to resolve a large part of its United States pension plans through lump sum distributions and annuity purchases, expected to settle later in 2026. The projected benefit obligation is $39.0 million and roughly $36.7 million of plan assets were converted to cash in June. On settlement the company will reclassify previously unrecognized actuarial gains from other comprehensive income into earnings. There will be a one time income statement item in the second half, and its sign is favorable. 

The stated order is organic investment first, balance sheet second, acquisitions third. No dividend, no buyback. The long term net leverage target is below 2.5 times, reachable organically by year end, and management has said it will exceed the target for the right acquisition and pay down quickly afterward. One constraint appears on no slide. As part of the reorganization around the IPO, the company cannot conduct primary follow-on equity offerings for two years, until roughly April 2028. It cannot issue shares to fund a deal. Every acquisition between now and then must be paid for with free cash flow or borrowed money, which makes the revolving credit facility rather than the share count the thing to watch.

The trailing price to earnings ratio of roughly 148 times is an artifact rather than a valuation. Depreciation and amortization of $107.6 million in the first half, most of it amortization of intangibles created by acquisitions, sits between the cash the business generates and the profit it reports. A screener sees a triple digit multiple. Someone reading the cash flow statement sees roughly 25 times adjusted earnings. Return on invested capital is the honest problem. At roughly 7 to 8 percent it sits at or below most estimates of the cost of capital. The operating businesses earn well on the equipment they actually use. The prices Madison paid to assemble them have not yet cleared the hurdle. That changes how the investment case has to work. It cannot rest on the returns as they stand. It requires growth and deleveraging to improve them, which is a higher bar than saying the ratios look attractive.

Full year sales guidance rose $75 million at the midpoint to $3,825 to $3,925 million. Adjusted EBITDA guidance did not move, staying at $1,020 to $1,065 million. More revenue producing the same profit means lower margins, an implied reduction of roughly 50 basis points at the midpoint. Management assigned the proof to a third quarter guided at 27 to 28 percent margins with further improvement in the fourth, against a third quarter 2025 revenue base of $898 million.

A separate reading problem runs through every disclosure. The company reports growth four different ways. Reported compares this year to last as filed. Pro forma adjustments last year as if AprilAire had always been owned. Combined adds AprilAire's standalone results to last year without other adjustments and does not comply with GAAP. Organic strips out acquisitions and currency entirely. Residential is the clearest illustration. Reported growth was 16.2 percent. Combined growth was 2 percent. Organic growth was negative 4.8 percent. All three appear in the same disclosure package and all three are accurate. A reader who does not know which language a number is spoken in does not know the number.

Fuel The Keyboard

This blog runs on caffeine and 10-k’s. Send a cup of joe my way to keep the words flowing! Support me here.

Madison Air currently has an extremely strong commercial organic growth engine, running up nearly 22% in the most recent quarter. This was led by air, liquid, hybrid cooling, customer air handling and air movement solutions. While Madison Air did see a massive increase in data centers of 50%, they did increase their prices slightly and saw an increase in demand. Madison Air’s backlog also grew just over 20%, to over $2.87 in confirmed orders. Currently nearly all of Madison Air's revenue is coming from North America, although StatePoint installations have begun in Ireland, South Korea, and India. While International revenue probably is less than the company's Red Bull bill, it will likely become a larger portion of the pie in the future.

 Madison Air’s organic growth stems from its residential channels which is mostly replacement and service demand, while its commercial segment works with larger customers building customized units for their facilities. Madison Air’s largest channel has become data center cooling as the hyperscaler capital spending guidance reaches the trillions, this will quickly become the company’s largest source of revenue. Replacement and repair revenue is another large slice of the pie, as units get older the more likely they are to be replaced and repaired to adhere to any new building codes or demands. On the residential front, April Aire has a massive opportunity, management noted that 92% of American homes have no air quality system at all. Each percent that goes down is the equivalent to millions of potential service visits per year. 

On the inorganic side of things. Well. It should be obvious as Madison Air has been built on roll ups and bolt ons. No need to run through them all because that would take a few thousand words, but the company has completed nearly $10 billion in deals. This includes Bigg Ass Fans, Nortek Air, AcoustiFlor, AprilAire, and most recently EBM-Papst. They recently sold Nortek Global HVAC, which shows their willingness to sell an asset that is no longer adding value to their thesis. One core advantage to Madison Air’s acquisition strategy is that more than 80% of its bolt ons have been private deals, sourced proprietarily with no auction process. Madison Air has done an excellent job at integrating their acquisitions into current operations, with AprilAire contributing to $213 million in sales. Madison Air is now focused on acquiring either technology platforms, increased channel access, and new markets that add services and aftermarket exposure.

Madison Air is currently investing in capacity expansion and recently in March they signed a seven year deal on a production facility with an option to extend the lease. Management has historically been pro asset light to protect their free cash flow, with the increased demand coming from the data center and AprilAire, they thought it was right to invest in expansion. Madison Air currently has over 600 employees on their Research and Development team, with their working increasingly showing results through the CDU range, increased FrostShield and product adoption, as well as continued Healthy Air Systems releases plus plug and play launches like their new humidifier than can cover a 6,200 sq/ft unit using drastically less water than before. Nortek Data Center Cooling has also moved its headquarters closer to the soon to be data center corridor. And they have deployment plans to work with Meta in Ireland, Empyrion in Seoul, Digital Edge in Manila and RackBank in India. These are small revenue projects but they are more about Madison Air’s ability to service international markets. 

Nearly all of Madison Air’s supply chain and manufacturing base is domestic, with $302 million out of $360 million of the company’s property and equipment sitting within the United States. Its main foreign imports are steel,aluminum, and copper all of which now carry the Section 232 Duties. The company hedges its steel with commodity futures, running a book of foreign exchange contracts (type shit). They also have built up a warchest of raw material inventory, which is insurance against any continued tariffs or fuel uncertainty. Management has also stated that its increased demand for data center cooling products has straightened their supply chain, especially within their electrical components. Hyperscalers are also coming to Madison Air with Jill Wyant (CEO) saying on a recent earnings call “... Came to us recently and said we want to use your labs and your very talented lab team … qualify additional sources of supply.” This is an extremely telling statement for the quality of Madison Air’s products and R&D team but also Hyperscalers are relying on Madison Air for their own supply chain decisions. 

Then there is the digital portion of Madison Air’s supply chain, which has less public facing documentation around it as most people don’t think of supply chains as being a digital thing. Madison Air noted they have had troubles in the past with the general design and effectiveness of their IT infrastructure.  Madison Air has been seeing an increase in their digital orders with it reaching nearly 30% of the self service revenue. This revenue is higher margin as these customers are cheaper to service, loyal, and produce usage data, which helps their aftermarket flywheel increase efficiency. Luckily for Madison Air this does not need a massive capex investment to fix, just a few lethal software engineers. 

Madison Air is exposed to many execution risks across its business lines, especially from its slowed residential business, customer concentration, and increasing revenue backlog. Madison Air recently reported it has a backlog of $2.87 billion, while this is excellent news on paper. What does it actually mean, is this high margin revenue, are these locked in deals or can their customer take their business elsewhere if time to shipment is delayed? Management does get ahead of this and notes that backlog revenue is defined as confirmed orders that have not been shipped. Madison Air also saw a slight fall in margins in their commercial unit from its data center unit, management noting that as data center business matures margins will stabilize and grow. Although the counterpoint to this is that hyperscalers are known to negotiate harder, require more expensive customer projects, and require products fast. One of their customers also currently accounts for nearly 13% of net sales, while Madison Air does not name said company, we can only assume it is Meta/Facebook due to their previous engagements in the data center realm. 

A real deceleration risk for Madison Air, is its data center business. Hyperscaler spending has gone up and to the right for the past few years but projects are being put on pause thanks to idiot politicians and other morons who are anti data center built out.  These decels should be locked up. Management has signalated this this slow down and pause might start showing up in the fourth quarter of 2026. Another risk for Madison Air is that a stalled data center build out means it would take them longer to improve margins on that segment of their business. 

Non-US revenue also took a major hit, shrinking from $77 million to $44 million. The company did not explain why but one can assume it can be tracked back to tariffs and shipping costs.  Next is their valuation, Madison Air is currently being valued for its future growth, if the cowards of Washington and your local government slow more projects this could cause Madison’s valuation to take a hit. Lastly is the decrease in growth within its residential unit by almost 5%. This was driven by volume declines in professional distribution channels for ventilation caused by price increases and also decreased unit sales.

Tariff refunds are helping cushion margins in the short term but this will eventually end.  Competitive pressure from Vertiv, Schneider, Johnson Controls, and Trane are tightening as they all are entering the liquid cooling business and attaching services to broader electrical bundles. Something that Madison is currently unable to match. Regulatory risks are also increasingly something Madison Air has to worry about relating to their data center segment, refrigerant rules, water restrictions, and product permitting. While Madison Air has invested heavily to follow and lead in product innovation with its StatePoint product, any rollback in regulation or energy code could remove any product advantages they had. 

The AI spending is both a durable advantage for Madison Air and a potential weakness. If the data center business is scaled at a rate faster than the rest of their product segments, when the music slows, Madison Air could see a significant drop in revenue. Hyperscalers have planned their spend out years in advance, Madison Air must do the same with their components, so they aren’t caught subsidizing their segments from the data center build out. AI spending is still in its early days and over funding infrastructure is something that the history books have written about many times. 

On the management and governance side of things. Activist investors might want to sit this one out, the parent company controls 95% of the voting power through its Class B shares. The board is led by the founder and his loyalists. At the end of the lock up period Madison Air's parent entity will still own a significant chunk of the company, though since this is a permanent capital vehicle, there is no telling on what they’ll do. Azar Research Collective does not believe insider selling is a useful “metric” or something even worth noting. I am guessing insiders will be selling as they now want some liquidity to buy some fancy items. 

Madison Air’s parent has a significant economic and cost interest currently with 64.7% economic interest and 95.2% voting power. The company currently has a board of just four members, one of them which is Larry Gies, the founder of Madison Air. He has been running Madison Air since day one. He holds an accounting degree from Illinois, where the business school now is named after his family, Larry also holds a management degree from Northwestern's Kellogg. 

Jill Wyant, joined Madison Air in 2021 as president and CEO. Before joining Madison she ran Ecolab as EVP and President of Global Regional and Healthcare. She spent over a decade at Ecolab climbing the company’s ladder from strategic planning to running a $5 billion operation with 25,000 employees. She also currently sits on the board of Dow and serves as Vice Chairman of the U.S. Chamber of Commerce. Under her leadership Madison Air’s revenue has grown nearly sixfold. Her knowledge of ball is high with the “Return on Air” marketing language.  

JJ Foley is the company’s CFO and has been with the company since 2021. His entire work history before joining Madison Air was spent at GE, starting as a member of the GE financial management program and leaving as a Senior Director of Investor Relations. This resume matches what Madison Air needs, the skills of someone with an auditor background but also the knowledge of a seasoned investor relations professional. 

Unlike many private equity operators that roll up industrials, Madison Air has historically left the operator who built the asset to operate the asset after they close the acquisition. They also give that operator / founder equity which helps retain them. Karin Overstreet is currently the president of Nortek Data Center Cooling, currently the most valuable asset in Madison’s portfolio. Karin Started her career at Honeywell where she was a supply chain manager, then jumped from HP to Lockleed, GE, Stanley Black & Decker to IDEX corporation where she was the Global president. Jatin Khanpara is the Chief Technology Advisor to Madison Air, he has been transforming the air industry for over 30 years and holds over 7 patents. Starting his career as a Program Manager at Whirlpool Corp where he spent 14 years leaving the Head of Global operating platforms.  

Despite being a very old company, Madison Air is a young public company. The IPO lockup expires in mid October, with their third quarter results coming in a few weeks after in November. This should be Madison Air’s first real test as they are committed to 28% margins within their commercial unit. Also in October, other hyperscalers will be dropping their spending guidance, which will set the demand for the thermal supply chain. Management expects to release their first annual report by the first quarter of 2027. The good thing though is that more than half of their $2.87 billion backlog is expected to convert to real revenue in 2027, this will help them cut their leverage down from nearly 6x to their 2.8x target. 

Underneath these headlines will be their backlog conversion, residential units either bottoms through the AprilAire penetration or keeps losing volume.This backlog revenue is mostly data center related, while Madison Air is diversifying into the data center cooling segment they may be locking themselves into a corner of the market. Tariff refunds have also recently been cushioning margins so when those end, will Madison be able to keep strong margins. 

Madison is less than a year into its public life and is currently only covered by analysts who anchored its roadshow. They have two strong maturing businesses in industries that are going nowhere. Despite the current residential slowdown, Madison Air thinks AprilAire will turn things around. The bad boys over at Azar Research Collective like Big Ass Fans and think investors should follow their commercial and residential margins, backlog conversion from data centers and non centers, hyperscaler spending guidance, book to bill, and any other acquisitions that will strengthen current market share.  

Jill Wyant said on their recent earnings call “Air is essential infrastructure in the markets we serve.” This sentence is pure marketing but also fact.  

Thank you for your attention to this matter. See you later, space cowboy.

Fuel The Keyboard

This blog runs on caffeine and 10-k’s. Send a cup of joe my way to keep the words flowing! Support me here.

Disclosure

Educational analysis. Not investment advice. No recommendation made or implied. These views are my own and have not been influenced by friends, family, or enemies of the state. This content is for informational and educational purposes only and does not constitute financial, investment, legal, or professional advice. Past performance is not indicative of future results, and the author assumes no liability for any investment decisions made based on this content.