Legence Corp. (LGN) stock intrinsic value fair value estimate & owner earnings analysis

Legence designs, builds and maintains the heating, cooling, electrical and process systems inside complex buildings: data centers, hospitals, labs, universities and factories. It has two segments. Installation & Maintenance does the contracting and service work and is now about 84% of revenue. Engineering & Consulting does the design and energy-efficiency work. Blackstone assembled the company through acquisitions and took it public in September 2025. Blackstone's entities still hold Class B units that can be exchanged into Class A shares, so we count both classes together: about 108.1 million shares (76.9M Class A plus 31.2M Class B as of August 2026). What the business earns. Revenue grew from $1.62B in 2023 to $2.10B in 2024 and $2.55B in 2025. On January 2, 2026 the company acquired The Bowers Group, a Maryland mechanical contractor, for about $427M, paid with $283M of cash (partly from a new $200M term loan), $99M of stock and $45M due at the end of 2026. First-half 2026 revenue was $2.30B, roughly double the prior year, and Bowers supplied about $300M of second-quarter revenue alone. We measure owner earnings as operating cash flow, with working-capital swings removed, less stock compensation and the capital spending needed to maintain the business. On that basis owners earned $59M in 2023, $82M in 2024 and $66M in 2025. Before interest, which is how we value the business before paying its lenders, those figures are $121M, $168M and $163M. Two accounting judgements matter. First, 2025 stock compensation of $68M includes about $28M of payouts on pre-IPO partnership interests. These were funded by Blackstone's holding entities, not by company cash or new shares, so we exclude them from the normalized level. Ongoing company equity awards are running about $6M a quarter. Second, the large amortization of acquired customer relationships and backlog is added back as a sunk acquisition cost. Its economic decay is instead reflected in a base case that stops growing after year two. Recent results show this was the right choice: impairments of acquired businesses came to $5M in 2023, $18M in 2024 and $27M in 2025, plus $41M in the second quarter of 2026 on the sustainability consulting business. The starting level. The company before 2026 was a different, smaller business, so we start from 2025 rebuilt for today's shape. That means $163M before interest, plus about $28M of Blackstone-funded compensation, plus a conservative ~$50M for a full year of Bowers, giving roughly $240M. First-half 2026 cash generation ran well ahead of that pace, but half a year in a seasonal contracting business is not enough to set the level. Business cash flow fell about 3% in 2025, and owner earnings fell about 19% as IPO-related compensation rose. Our base case still has year one rising to $260M, for named reasons in the filings: Bowers closed January 2, 2026, and contracted backlog (remaining performance obligations) rose from $2.79B at December 31, 2025 to $4.02B at June 30, 2026, with most of it due to be worked off within twelve months. The cases. In the base case, business cash flow is $260M in year one and $280M in year two as the signed backlog converts. It then settles at $270M a year, and after year ten it fades to a sustained $250M. That later level rests on the recurring maintenance and service work on an installed base of mission-critical buildings. No market growth, share gains or margin expansion is credited. The business is worth about $4.31B. Net debt is about $1.09B: $1.04B of term loan, notes and finance leases, less roughly $190M of cash beyond operating needs, plus the $45M Bowers payment and about $200M as today's value of payments owed to the former owners under the tax receivable agreement. That leaves about $3.22B, or $29.78 a share. The bear case reflects the main concentration risk. About half to two-thirds of recent revenue comes from data center and technology customers. It also draws on the record of building-systems contractors in the 2008-2010 downturn, when the largest US peers saw revenue fall about 15-25% and operating profit fall by roughly half. We assume the data center build-out pauses in years three to five and business cash flow drops from $240M to about $110-140M for three years. It then recovers only to $180M, and the company keeps writing off acquisitions as it has every year. On that path the business is worth about $2.88B. It owes about $1.14B, including extra borrowing through the trough, leaving about $1.74B, or $16.05 a share. Even at the trough, cash flow before interest covers interest of roughly $60-65M a year, and the term loan does not mature until December 16, 2031, so the debt survives this case. In the bull case, the company's identified levers work: the backlog converts on schedule, and its demonstrated pattern of tuck-in acquisitions keeps adding service capacity. Business cash flow reaches about $350M by year three and holds there. The business is worth about $5.76B. After about $1.09B of net debt, that leaves $4.67B, or $43.15 a share. What has changed. The previous estimates were built on a different earnings measure that started from the company's adjusted EBITDA, so they are not comparable. Since then, the business closed Bowers, its contracted backlog grew by more than $1.2B in six months, and it took another impairment on its consulting arm. The company has also disclosed material weaknesses in its internal controls over IT systems and journal-entry segregation. That does not change the cash figures used here, but it lowers how much weight we can place on any single reported number. Margin-of-safety price. We are only moderately confident in the earnings level. The company has three years of history as a combined entity, it has doubled in size in eighteen months, and its owner earnings per share in the filings ($0.56-0.78) sit well below the rebuilt level of about $2.20 a share that the cases rely on. The worst plausible outcome is that hyperscale data center spending stops abruptly while fixed-price jobs are underway, losses emerge on large contracts, and the impairments continue. With leverage of about 1.5 times cash earnings, debt due in 2031, and maintenance work that continues regardless, the company would survive this, but per-share value could fall to the mid-teens. The things we are most likely to be wrong about are what Bowers really earns, how much of the data center work recurs, and how far the stock-based pay and tax receivable payments to the former owners keep draining shareholders. If all three prove wrong, the shares are worth about the bear figure. The company pays no dividend, so the whole return has to come from the business building value over three to five years. The margin-of-safety price is $14: from that price, earnings that only hold near today's rebuilt level still imply 20% a year, and that does not depend on the data center boom continuing. Above that price, the return would depend on the base case playing out.