Limbach's competitors in mechanical contracting now do 30% to 60% of their business in data centers. Limbach does none.
Its management said so out loud while the market was still treating the company as a data center winner. It spent the past three years building direct relationships with the owners of hospitals and factories instead, and this summer that business stopped growing.
"But a rising tide should lift all boats who are playing that tide."
Yaron Naymark pitched Limbach on this podcast in June 2023, watched it rise 6.5 times, sold most of it, and is buying it again after a 75% drawdown.
I listened to the full interview so you can skip it. 51 minutes of audio, 20 minutes of reading.
Here are the 14 takeaways that matter.
👤 Guest: Yaron Naymark of 1 Main Capital, a long-biased investment partnership, who pitched Limbach on this show in June 2023 and watched it rise 6.5 times before selling most of it
🎙️ Host: Andrew Walker, the founder of Yet Another Value Blog, who interviews investors about the individual stocks they own
📰 Published: 1 September 2026 on YouTube (Yet Another Value Podcast) and on the show's own feed
🔴 YouTube | 🟣 Apple Podcasts | 🔗 Show notes | ⏱️ 51 min | ✅ Time saved: 31 min
Key Takeaways
Limbach's peers do 30% to 60% of their work in data centers and Limbach does none of it Naymark called it a mistake in hindsight, made while the company was busy avoiding general contracting risk
A 30% fall in EBITDA on a 6% fall in revenue is a fixed-cost story, not a bidding story Management chose not to cut the cost base because it reads the slowdown as temporary
The stock fell 50% on a guidance cut of about 11%, and the shape of the guide is why Down 30 in the first half and up 20 in the second is a promise investors did not believe
Even at $65M of EBITDA, well under the $80M guide, he makes it $4 a share of free cash flow That is about 10 times earnings on a balance sheet with effectively no debt
The data center boom raises Limbach's labor costs whether or not it wins the work Technicians can earn $150,000 a year, and $20 more an hour moves them to a data center job
Buying private contractors at five to six times beats buying its own stock at six times The $50M buyback authorized in December 2025 is still untouched, and he expects it to stay that way
The CYMCOR deal is a $30M way into data centers through the back door Advising the builder is what gets you invited to bid on the build
He is buying businesses that will be neutral or better on AI and are not priced that way The other two examples in the portfolio are IWG and KKR
1. From Enron to a SPAC rollup
Naymark started with the company's history rather than the pitch, on the grounds that the last Limbach episode was three years ago and he could barely remember it himself.
Limbach is a mechanical, electrical and plumbing contractor that specializes on the mechanical side, mostly heating and cooling systems for buildings that cannot afford to lose them: hospitals and advanced manufacturing plants
The company is more than a century old. It was founded by a man named Limbach, sold, and ended up inside Enron; a private equity firm bought it out of the Enron bankruptcy and took it public in 2016 by merging it with a special purpose acquisition company
The listed company existed to buy small local contractors in a fragmented market at prices well below its own. Naymark said the small players traded at four to five times EBITDA then and have crept up to five or six times now
Before that could happen the company took on large new-construction jobs, wrote them down, lost money and went into COVID as a distressed equity
It dug out through earnings growth and free cash flow, paid down debt, and the board replaced the chief executive in early 2023, promoting the chief operating officer, Mike McCann
2. Owner-direct lifted margins
McCann's shift was away from working for general contractors on new buildings and toward working directly for the people who own the buildings — upgrades, repairs, retrofits, adding a wing to a hospital
Naymark said owner-direct work ties up less cash, earns higher margins and carries less risk of a single job blowing up
The mix went from roughly 80/20 in favor of general contracting when the company listed to roughly 75/25 in favor of owner-direct as of last year
EBITDA margins went from the low single digits to the low double digits over that stretch, with acquisitions layered on top
3. Zero data center exposure
The market briefly decided Limbach was a data center winner, and Naymark said the company had told everyone it was not. "Mind you, at the time, the company was not pursuing and was vocal about not benefiting significantly from data center business, but a lot of their competitors and other MEPs in the space were benefiting tremendously from data centers."
The work has been profitable for the peers who took it, because the customers buy speed rather than the lowest price. "Not that they're throwing money at every problem and not that they're not looking at what things cost but they really care about quality and speed."
He called the omission a mistake in hindsight. The company was singularly focused on staying out of general contracting trouble and on its owner-direct relationships, and missed the trend while it did
A lot of mechanical contractors now run 30%, 40%, 50% or 60% of their business through data centers, he said, and Limbach is at "we effectively have zero" If it wins a normal share, he put next year at $100 million or more of EBITDA, and $120 million organically, before any acquisitions
4. Demand hit an air pocket
The owner-direct business slowed in the first half. Naymark attributed it to tariffs and the trade war, Medicaid cuts to the healthcare vertical he traced to what he called the build back better bill, and this year's Iran war and higher oil prices, with discretionary projects paused or put on hold
Organic revenue fell about 5% to 6% and EBITDA fell about 30%, a gap he said is mostly fixed costs spread over less work rather than bad pricing
Management did not cut into that fixed cost base, because it read the slowdown as temporary and knows the costs are hard to add back in a growing market
Headline revenue rose while organic revenue fell, which Walker raised as a clarifying question before anything else. The bridge is the acquisition of Pioneer Power Pioneer Power carries lower margins than the core business, which management expects to raise over time
5. Why the stock fell 50%
Guidance for the year came down from $90 million of EBITDA to $80 million alongside the second-quarter results, and the stock fell about 50%
Naymark's explanation is the shape of the back-half guide rather than the size of the cut. "I think the stock's down 50 because of the way they guided the back half, which is down 30 in the first half, up 20 in the second half year-over-year seems unrealistic." "And for public market investors, it's very hard to own a stock where you think they might miss or guide down again."
With no annual guidance at all, he thinks the stock would be down about 30, in line with the earnings decline
There is effectively no debt, so the enterprise value is the market capitalization, which he said takes the leverage risk out of the fall
He thinks the full-year number is achievable and that any miss would be small
Walker said two of the three worst days the stock has ever had were this year's first and second quarter earnings, both down 30%
6. Double-dipping the same name
Walker opened with a general question about buying a stock a second time, and disclosed his own record on it: the names he has done best and worst on are the ones he sold and bought back, because a new risk can appear between the sale and the repurchase and he comes back to it with his first-time-around view.
Naymark said the thesis is the same one as 2023 — margin expansion, then multiple expansion, then capital allocation — at a slightly higher multiple and on higher margins than the first time
The end market is the part he keeps coming back to. "This is an end market that's not a melting ice cube end market. There's going to be a need for this service for the decades to come."
Private competitors and listed peers are seeing heavy demand, organic growth and margin expansion, and Limbach has labor that is in short supply The largest players are capacity-constrained on fabrication; Limbach has spare fabrication capacity to sell
The floor is what he leaned on, and he built it below the guidance. At $65 million of EBITDA rather than $80 million, less about $5 million of stock compensation and $5 million of capital spending, he gets $55 million of pretax earnings and $45 million of free cash flow. "That's four bucks a share and you're trading at 10 times that number today" A print like that would be "a disaster near term for the stock", he said, and he would still underwrite upside from it rather than downside
7. The bear case on margins
Walker put the short case to him directly: that management panicked during last summer's air pocket in orders, bid a pile of very low-margin work, and is burning through it now. The worry underneath it is that a team that did that either does not know what its own work is worth or will do it again. He said there was a short report on Value Investors Club he thought was very good.
Revenue guidance went the other way — a midpoint of about $750 million taken up to about $780 million at the second-quarter guide — while the EBITDA guide came down, which is the pattern the bears point at
Naymark's test is which way revenue moved. "I would be more concerned about the bookings that they took on over the last few quarters if we saw in the first half of this year revenue up, margins down substantially."
The company attributes most of the gross margin decline to three things: Pioneer Power now sitting in the consolidated numbers, fewer write-ups on jobs finishing this year than last, and the fixed-cost deleverage. It is guiding to significant gross margin expansion in the second half If that guide is met, he said, it undercuts the claim that the new work was knowingly taken at lower margins
On the fear that owner-direct is general contracting under another name, he said the label would not be the problem. General contracting stocks trade at 10 to 25 times EBITDA on the data center tailwind, so a Limbach that won that work and got called a general contractor could still re-rate
What he got wrong: as the business moved toward owner-direct he grew dismissive of weak bookings, on the view that short-duration work is won and burned inside a quarter and never shows up in backlog. "The bears turned out to be right over the short term."
His argument for management's alignment is McCann's own position. The chief executive worked his way up to the job, never took much cash compensation, and was worth about $40 million on paper when the stock was $150. "He didn't sell a single share."
8. A rising tide he isn't in
Walker's own preparation produced the starkest number in the episode. He said that over the past three years Comfort Systems, which trades as FIX, is up 800% and EMCOR, EME, is up 250%, against 16% for Limbach, and that the one-year gap is starker still His first note while prepping was that he did not understand why the business was not firing on all cylinders, and he had not realized Limbach had no data center work at all
He asked whether a rising tide should not lift Limbach anyway, since competitors chasing data centers leave fewer bidders on the healthcare work
Naymark said the tide only lifts the boats in it. "But a rising tide should lift all boats who are playing that tide." For everyone else the boom is a cost. "If you're not getting data center work, the demand from the data centers is pushing labor and material costs higher and making things more inflationary in nature." The core customer is being quoted higher prices and pushes back, so the costs arrive without the pricing power
Walker's theory was that owner-direct made it worse. "You know you'll hear about people making like $150,000 a year as an air conditioning technician." Technicians can leave for $20 more an hour on a data center job while Limbach sits inside long-term owner contracts it cannot reprice
Naymark's answer was volume: 10% to 15% organic growth would offset a lot of the inflation, and a 6% revenue decline is what makes it eat the margin instead
9. Management was surprised
Walker read the first-quarter and second-quarter calls and flipped through the fourth-quarter 2025 call. Guidance was reaffirmed in the first quarter and cut in the second, with 2026 called a reset year, three months apart.
He asked whether management was caught out, and Naymark said it was. "Yes I do think they were surprised." He recalled the company saying on the first-quarter call something like "We're comfortable with Q2 consensus estimates", which he said is what got him and other holders into trouble
He owned the stock into the fall and has been buying since. "I did own the stock going into the Q2 blow up." "I didn't just reinitiate on the down 50, but I have added to the position substantially in the last few weeks."
The mechanism was the burn rate, not the order book. "So they had bookings, they went into backlog, they expected that backlog to burn at normal burn rates and customers were dragging their feet, some voluntarily, some involuntarily." Voluntary delays were macro: tariffs, the possibility of war, so a project waits. Involuntary ones were owners who could not find electricians, which stalls the mechanical work behind them
He said the company has scrubbed the numbers and has decent visibility into the third quarter, less into the fourth, and that competitors both public and private have seen the same trend outside data centers over the last six to nine months, now normalizing
If the second half lands, he said, the first half was the anomaly and the stock goes back to where it was before the fall; if the company wins $100 million or $200 million of data center work for next year on top, it becomes a data center story with operating leverage, and there are scenarios where the stock doubles or triples in six to nine months
10. M&A beats the buyback
The company authorized a $50 million buyback in December 2025 and has not bought anything under it. Walker's first read was that management could see the slowdown coming; he then said Naymark's answer was the better one
Naymark expects the cash to go into acquisitions instead. "So I would be surprised if they're buying back stock. I think they're focused on acquisitions."
The arithmetic is that buying private mechanical contractors at five to six times EBITDA, with no capital spending attached, is worth more than buying the company's own stock at about six times, because there is no day-one spread either way but the acquisition adds scale, diversification and operating leverage, and with them a lower cost of capital He thinks the platform is worth well above six times, so deals done now create value for the day the multiple returns to eight, 10, 12 or 15 times
Every listed company should be ready to issue and to buy, whether or not it does either. "Like every public company should have an ATM ready to go and a buyback ready to go." An at-the-market program is standing authority to sell shares into the market, and he said filing one costs about a hundred dollars "All these meme stocks that didn't have ATMs and their stocks are screaming like we don't know how to issue shares."
On scale, Limbach does $750 million to $800 million of revenue. Comfort Systems does $11 billion or $12 billion, EMCOR is in the tens of billions, and private firms he has spoken to run $5 billion to $8 billion. He said Limbach is still small in a consolidating market
11. The math to $200 a share
Walker said one of Naymark's letters carried a $200 three-year price target, and asked him to walk through it. He framed the business the way a caller on the sponsor's platform had: construction does not go away, the owner relationships make the revenue close to recurring, and there is an acquisition engine on top.
The math is $10 a share of free cash flow by 2030 at 20 times it. "I thought at that time and I still think currently that you can get to $10 a share of free cash flow by 2030 through some organic growth and layering on acquisitions."
The stock is $40 today, as Walker noted straight after
Where the multiple comes from: "So like I said the peers are trading for 10 to 25 times EBITDA." Comfort Systems, a non-union shop with more scale and better margins, is above 20 times; EMCOR is around 15; Legence, taken public by Blackstone, is around 13; smaller and mid-size players are at 12 to 15
He allowed that the business could be worth 15 rather than 20, or 25 rather than 20, but said 20 is reasonable for a clean balance sheet in an end market benefiting from the data center build-out
12. Who could buy Limbach
Walker asked the reverse question: a peer buying Limbach gets fabrication capacity it can point at data centers, the synergies, and the arbitrage between the two multiples.
Naymark's condition for staying independent is execution, and he set the bar low. Low single-digit organic growth with flat to growing margins, earned by leveraging overheads rather than by raising gross margin, is enough to make Limbach the buyer rather than the target
If the execution problems continue, he thinks the company should be consolidated into a larger player. EMCOR and Legence are the two he named as plausible, and both are union shops, as Limbach is Comfort Systems is not — he put its union headcount at about five people in the entire company — which he said makes it an unlikely buyer
He believes EMCOR looked at Limbach more than a decade ago and did nothing, and would not rule out a second look
A hostile approach is the one route he ruled out. It is "hard to do non-friendly takeovers" in a business where the talent walks out of the door every evening, though he said a company that put itself up for sale would draw interest at a premium
Walker added the buyer's arithmetic: public company costs come out in a deal, so an acquirer bids off $90 million or $95 million of EBITDA rather than the $80 million guide, and at its own higher multiple
13. CYMCOR buys a way in
Alongside the second-quarter results Limbach bought a business that sells advice to data center builders. "It's a program management business that focuses on data centers." The company has done the same work in healthcare for years
Limbach paid $30 million for it and expects about $4 million a year from it, which Naymark said is a higher multiple than it pays for the mechanical contractors it usually buys
The reason to pay it is what the advisory work leads to. "So you advise the builder of the data center and that gives you a foot in the door to bid on the work that you're advising them on." In healthcare, "I think it's like a 20x pullthrough multiple is what they've seen historically"
If that repeats, he gets to a few hundred million of data center revenue, or 20% to 25% of the business, which he pointed out is at the low end of what the peers carry
He called the zero-to-hundreds-of-millions version "a dream case, but it's possible", and said the downside if none of it happens is valuation support, stabilizing end markets, costs to cut and a clean balance sheet
14. AI winners nobody prices
Walker changed the subject for the last ten minutes, to how Naymark invests in physical-world businesses while the AI trade runs without him. The two of them, he said, once joked on a message thread that they should have just bought a leveraged Micron fund.
He is buying businesses that will be neutral or better on AI and are not priced as either. "I've always tried avoiding things that are rapidly changing and hard to kind of think about what the business might look like 5 to 10 years out."
Limbach is the first of his three examples: peers pulled along by data center construction trade at much higher multiples, so he could get the faster growth and the re-rating together
IWG is the one the market has decided is an AI loser, on the reasoning that AI removes office jobs and office jobs are what fill offices. He thinks it ends up a beneficiary "And in a world where you're no longer growing your headcount over time and maybe even reducing it's hard to sign a 10-year lease if you don't have visibility into what your footprint's going to look like in 10 years." Short-term rental is a low single-digit share of office space today, and he expects it to rise
KKR is the third, reinitiated this year during the scare over private credit's exposure to software companies bought before AI. "And I think the firms that are best positioned to survive that are the ones with the longest track records and the most blue chip names that are likely going to be given a pass for a bad vintage or two because they have 10 vintages before that did very well" Mid-market firms with fewer vintages get less rope, so he expects them to consolidate and the largest managers to take the share
His growth case for the largest alternative managers is geographic and channel-based, not performance-based. The wealthy retail channel is only starting; institutions in Asia hold a mid-to-high single-digit allocation to alternatives against 25% to 50% in the United States, and KKR runs the largest Asian alternatives business; Europe is underpenetrated too; and KKR is still catching up in credit, infrastructure and real estate
He also expects them to take share from index funds. "Do you really want to own all the businesses that are AI losers?"
Walker's addition was data. "Now data is the new oil. People were saying that 10 years ago, so maybe there's nothing new, but they have extremely sophisticated, extremely unique data from decades of deals, diligence, owning companies, all that sort of stuff." He also gave the other side of the passive argument: avoiding the junk has been the case for active management for years, and the index has still been a hard benchmark to beat
Bonus Insights
Walker spent several minutes on the chairman rather than the chief executive. Josh Horowitz, whom he has met twice, is on his third chairmanship; the first, a dental services company Walker referred to by its ticker, sold to private equity at what he remembered as a large premium, another board he sat on sold, and BK Technologies has been among the best-performing small caps of the past year or so. Horowitz owns a decent amount of Limbach stock
The de-SPAC question came last. Walker's standing joke is that anything that came public through a SPAC eventually returns to the $10 deal price, and he noted Limbach listed in 2016, was at about $4 a share by 2019 and ran from there, so it may have made that trip already Naymark's answer was that there is plenty of SPAC wreckage but there are winners too, and he named Restaurant Brands and APi Group. "Absolutely. It had a lot of blowups along the way for sure."
This was Naymark's sixth appearance on the show, though both of them thought it was the fifth. He was wearing the podcast's own shirt, said he almost wore the hat as well, and was pleased to save the shipping cost by earning another one
Naymark's bottom line is that Limbach's earnings fell because a fixed cost base sat against a temporary drop in revenue rather than because the business broke, and that at about six times EBITDA with no debt he is paid to wait for the data center work the company is only now going after.
Products, Companies & Tools Mentioned
Limbach Holdings (The mechanical contractor the episode is about: heating and cooling work for hospitals and factories, now roughly 75% owner-direct, with no data center business and a stock down about 50% since the second quarter)
Comfort Systems USA and EMCOR Group (The two listed peers Walker said are up 800% and 250% over three years on data center work against 16% for Limbach; Naymark puts Comfort above 20 times EBITDA and EMCOR near 15, and named EMCOR as a plausible acquirer)
Legence (Newly public and, in his telling, at about 13 times EBITDA; the other union shop he thinks could consolidate Limbach)
Blackstone (Named as the owner that took Legence public)
Pioneer Power (The acquisition that lifted headline revenue while organic revenue fell, at lower margins than the core business)
CYMCOR (The data center program management business bought alongside the second-quarter results for $30 million, expected to contribute about $4 million a year)
KKR (Reinitiated this year on the view that the largest alternative managers get forgiven a bad vintage and take share from mid-market firms and from index funds)
IWG (Priced as an AI loser because AI removes office jobs; his case is that companies which cannot forecast headcount stop signing 10-year leases)
Books & Resources Mentioned
Yaron Naymark's first Limbach pitch on this podcast, June 2023 (The original version of the thesis, which this conversation is measured against throughout)
1 Main Capital's investor letters (Where the $200 three-year price target Walker asked him to justify was published)
Value Investors Club (Where the Limbach short report Walker called very good was posted late last year)
Limbach's first-quarter and second-quarter 2026 earnings calls (Walker read both, plus the fourth-quarter 2025 call, to test whether management was surprised by its own numbers)
If this was worth your time, send it to someone who follows the name.
Get the latest market chatter as it happens: