Life Time Group Holdings, Inc. (NYSE: LTH) — The Buildings Are Funding the Growth

Report date: 26 July 2026 · Price reference: $42.33 (close, 24 July 2026) Coverage status: Initiating coverage Standing disclaimer: Sections 1–15 below contain no buy/sell recommendation and no price target. The single exception is the Claude's Take block immediately below, which is deliberately fenced off. This article is general information, not investment advice.

Timing note. Life Time reports Q2 2026 on 30 July 2026, four days after this report date. Everything here is written on information available through 24 July 2026; the most recent reported quarter is Q1 2026 (released 5 May 2026).

⚡ Claude’s Take

This block is the author’s own subjective opinion, and it is the only place in this article where a position is taken. It is general information and not investment advice. The analysis in Sections 1–15 below carries no position and no price target.

Verdict: AVOID at $42 — accumulate on weakness in the high-$20s to low-$30s. A genuinely good business at a price that requires its own decelerating algorithm to stop decelerating. Not a short.

Life Time is a better operator than its critics allow and a worse compounder than its price implies. The advantage is real but narrow and local: a $60–90M, ~97,000-square-foot box with aquatics, childcare, racquet courts and a spa is a de facto local monopoly in an affluent ten-minute drive-time ring, takes two to four years to entitle and build and three to four more to ramp, and the family programming attached to it — swim progressions, summer camps, pickleball leagues — creates genuine habit-based stickiness. Four straight years of double-digit comparable-center growth and 330bp of margin expansion are not achievable in a contestable market. Management is executing well.

The problem is what sits underneath it. Life Time earns 5.2% on invested capital (≈6.6% stripping goodwill and idle construction-in-progress; ≈7.9% even granting a fully-matured portfolio) against a cost of capital built from its own capital structure at roughly 8.1%. Planet Fitness earns 11.7%. Life Time has never cleared its cost of capital in its entire public life — in its best year ever. Reported FY2025 “free cash flow” of $206M dissolves on inspection: management’s definition adds back the proceeds of selling its own buildings. Strip the $227M of sale-leaseback proceeds and the $94M of one-time COVID employee-retention credits and a Zurich insurance settlement sitting inside operating cash flow, and the underlying business consumed about $115M of cash in a record year. Since the IPO, cumulative organic free cash flow is roughly −$945M against $983M of sale-leaseback proceeds — the deficit and the asset sales match to within 4%. In 2026 management guides capex of $1.15–1.21B against adjusted EBITDA of $925–940M, and has already raised the sale-leaseback target from $300M to $400M to bridge it. Meanwhile the algorithm is decelerating in a straight line — comparable-center revenue 13.5% → 12.9% → 11.2% → 10.6% → 9.9% → 8.6%, guided to 6.3–7.3% — and management’s own new disclosure decomposes Q1’s 8.6% as mix +3.5pp, price +3.0pp, in-center +2.3pp, volume −0.2pp. Pure price is running at roughly inflation. “Mix” is the replacement of cheap members with expensive ones, and the cheap cohort is nearly exhausted at 3.4% of dues revenue.

Two facts settle the moat question for me. Center memberships at end-2025 (822,380) were still 3.7% below end-2019 (853,748) — with 29.5% more clubs. And in 2020, the one natural experiment available, Life Time lost 41.3% of its center memberships while Planet Fitness lost roughly 10%. The $15/month operator demonstrated more customer captivity than the $230/month one, because engagement-based stickiness does not survive an interruption in access. The only attrition rate Life Time has ever published — 16.1% in the seasonally best half of 2019 — implies roughly industry-average churn, and the company has not disclosed the metric once since going public while claiming record retention on four consecutive calls.

Framing: quality compounder at the expensive end of its own range — not a momentum crowd, not a falling knife. The stock sits 1.2% below its all-time high after a 55% three-month advance, and the factor model finds no momentum loading at all: this is an idiosyncratic operational re-rating (≈85% stock-specific variance) driven by de-levering to 1.6x and the clearing of a multi-year sponsor overhang. But the re-rating has been paid entirely through the multiple. At 2.99x book and 3.11x sales, Life Time has never been more expensive in its public history (99.3rd and 95.9th percentiles respectively; ignore the 22.8th-percentile P/E, which is a depressed-denominator artifact of the 2021–23 loss years). The tape now underwrites continued execution rather than a closing valuation gap.

The entry zone I would want is roughly $28–32 — about 2.0–2.2x book, or 20–22x normalized earnings (FY2025 normalized diluted EPS was $1.27, not the $1.66 reported or the $1.44 company-adjusted; FY2026E charging stock compensation is ~$1.50). That zone is not arbitrary: it is where the company itself repurchased 2,192,500 shares at $28.60 in May 2026, and where the sponsors cleared 43M shares at $29.50–$30.40 through 2025. The best-informed seller and the best-informed buyer both transacted there within the last fifteen months. Since the IPO, insiders and sponsors have sold roughly 79.5M shares for ~$2.27bn against about $4.0M of open-market buying — better than 500 to 1 — and no insider has bought a single share since 31 May 2024, through a 150% advance. In fairness, most of that selling is fund-life mechanics rather than a valuation view: the sponsors’ weighted-average realised price was ~$28.50, they made only ~2.9x over eleven years, the $18 IPO buyer has out-earned them, and LNK Partners has never sold a share. But nobody on the inside is signalling that $42 is cheap. At $42.33 you are paying a full multiple for a business whose returns on capital do not yet cover their cost.

Conviction: medium. Flips bullish if organic free cash flow — excluding sale-leaseback proceeds and one-timers — turns durably positive while comparable-center revenue holds at or above 7%, i.e. the “>$400M FCF by 2030” plan starts arriving early. Flips bearish if comparable-center growth breaks below the 6.3% guidance floor, or if membership units go negative excluding the deliberately-shrunk qualified-medical cohort — either would prove the price lever is spent.

Tag: a real business financed like a real-estate fund.

📈 Stock Price Action — Five-Year Event Map

Life Time has been public for 4.8 years, not five — the IPO priced in October 2021, so this map covers the company’s entire listed history. The arc is a full round trip and then some: a $17.75 first close (7 Oct 2021), a 58% collapse to an all-time low of $9.22 (4 Nov 2022), and a grinding four-year recovery to an all-time closing high of $42.84 (16 Jul 2026). The stock last closed at $42.33 (24 Jul 2026) — 1.2% off the all-time closing high and 3.7% off the $43.96 intraday high — against a 52-week range of $24.14–$43.96. Total return since the IPO is +138.5%; the last three months alone contributed +55.3%.

1. The IPO window (Oct–Nov 2021). Life Time listed at a $17.75 first close and ran 25% into a 26 November peak of $22.14 on reopening optimism for premium, high-ticket fitness. That peak stood as the high-water mark for the next 32 months.

2. The rate shock and the all-time low (Nov 2021–Nov 2022). The stock fell 58.4% to $9.22 on 4 November 2022 — the deepest drawdown of its public life, and it did not reclaim the November-2021 peak until 2 August 2024. Interpretation: a levered, capex-heavy, discretionary-spend operator was exactly the profile the 2022 rate shock punished hardest; the August-2022 CFO transition added governance noise into the trough.

3. The recovery leg (Nov 2022–Apr 2023). A 125% rally off the low. The Q3-2022 print drove +18.4% in a single day on 10 November; the UnitedHealthcare Renew Active expansion followed on 16 November; and the preliminary FY2022 results 8-K on 9 January 2023 added +13.1% in a day. Three discrete, dateable catalysts account for most of the leg.

4. The capex drawdown (Apr–Oct 2023). A 43% give-back. The Q2-2023 print missed and the stock fell 13.6% on 25 July; the Q3-2023 print fell another 15.2% on 25 October despite 17.9% revenue growth, as the market objected to spending on the premium experience and to a sequential membership decline. Interpretation: the second leg down was compounded by rising rates making sale-leaseback proceeds — the company’s principal growth-funding channel — materially less attractive. This is the single most instructive episode in the file: it is the market pricing the financing model, not the operating model.

5. The de-levering re-rating (Nov 2023–Oct 2024). The stock roughly doubled. The FY2023 print added 10.4% on 28 February 2024; Q2-2024 reported net-debt leverage down to 3.0x with positive free cash flow; and the October 2024 refinancing — an upsized $500M of 6.000% senior secured notes due 2031 plus a revolver increase to $650M — removed the 2026 maturity wall. A 12.0M-share offering at $21.75 on 12 August 2024 was absorbed with only a brief dip.

6. The momentum leg (Dec 2024–Feb 2025). A 50% run to $33.23 on 19 February 2025, catalysed by the preliminary FY2024 8-K and the FY2024 results. Interpretation: this is where the deleveraging story became consensus.

7. The supply overhang (Feb–Oct 2025). The stock fell 26% to $24.59 across eight months in which the sponsors sold 43 million shares — 23.0M at $30.40 (priced 27 Feb) and 20.0M at $29.50 (priced 5 Jun). The Q2-2025 print took 9.7% off on 5 August. Interpretation: the drawdown is largely supply-explained rather than fundamentals-explained; operating results kept beating throughout, and both offerings priced below the February high, capping the stock near the sponsors’ clearing price.

8. The overhang clears (Nov 2025–Jul 2026). A 72% advance to an all-time high. Q3-2025 was a beat-and-raise; the preliminary FY2025 8-K landed 22 January 2026; the company closed $200M of sale-leasebacks on 30 April 2026; and on 5–6 May 2026 the stock gained 11.8% then 14.3% back-to-back on a Q1-2026 beat-and-raise delivered alongside a 2,192,500-share private repurchase at $28.60 from Leonard Green, TPG and Partners Group plus an 8,770,000-share sponsor sale. Interpretation: the market read the combined print-and-repurchase as retirement of the multi-year sponsor overhang, and the two events compounded into the largest two-day move of the stock’s public life.

1. Executive Summary

Life Time Group Holdings operates 190 large-format “athletic country clubs” (18.4M indoor sq ft, 31 states and one Canadian province) at an average of roughly $230 per member per month — a price point four to eight times the US industry average of ~$69. It is a genuinely differentiated business in a structurally bad industry, and the distinction between those two facts is the whole report.

What is working. Revenue grew 14.3% in FY2025 to $2,995.3M and 11.7% in Q1 2026 to $788.7M. Adjusted EBITDA margin has expanded 170bp in FY2025 to 27.5% and reached 28.7% in Q1 2026. Average revenue per center membership has risen from $2,810 (FY2023) to $3,531 (FY2025), +25.7% in two years — and it has risen while average visits per membership increased to 12.5 per month. Raising price 10%+ per year without losing usage is the single strongest piece of evidence that a real local franchise exists. Net leverage has fallen from 2.28x (YE2024) to 1.6x (YE2025); the company earned a BB rating in Q2 2025, a year ahead of plan; and the multi-year sponsor overhang was retired in May 2026.

What is not. Three things, and they compound.

First, the returns do not cover the cost of capital. Return on invested capital was 2.84% (FY2023), 4.06% (FY2024) and 5.22% (FY2025) — reconstructed independently from the statements at 4.85%, or ~6.6% stripping the 2015-LBO goodwill and $555M of idle construction-in-progress. A WACC built from Life Time’s own capital structure is roughly 8.1%. Greenwald’s threshold for the presence of a competitive advantage is a sustained 15–25%; 6–8% signals its absence. Planet Fitness earns 11.7%. The trend is genuinely improving; the level has never been acceptable.

Second, the reported cash generation is not what it appears. Management’s own “free cash flow” definition adds back proceeds from selling its own real estate. FY2025’s headline $206.5M comprises $227.4M of sale-leaseback proceeds against a true CFO-less-capex figure of −$21.0M — and that −$21.0M itself flatters the picture, because operating cash flow contained $94.2M of one-time COVID employee-retention credits and Zurich insurance proceeds. On a recurring basis the business consumed roughly $115M of cash in a record year. Life Time has generated positive CFO-less-capex in one of the last five years, and that year (FY2024) was the trough of its growth-capex cycle. Cumulatively since the IPO, organic free cash flow is approximately −$945M against $983M of sale-leaseback proceeds. The match is not coincidental — it is the business model.

Third, the growth algorithm is decelerating in a straight line and its composition is deteriorating. Comparable-center revenue has gone 13.5% → 12.9% → 11.2% → 10.6% → 9.9% → 8.6% across six quarters, guided to 6.3–7.3% for FY2026 against a 6–8% long-term target. More importantly, management’s new Q1 2026 disclosure decomposes the 8.6% as: mix +3.5pp, price +3.0pp, in-center +2.3pp, volume −0.2pp. “Mix” is the replacement of cheap members with expensive ones — and the largest cheap cohort, third-party-insurance “qualified medical” memberships, is already down to 3.4% of dues revenue after falling 14.9% year-on-year. That lever is finite by construction. Center memberships per club have fallen from 4,537 (FY2024) to 4,351 (FY2025) to ~4,410 in Q1 2026 versus 4,591 a year earlier. Taking the longest available view: center memberships at end-2025 were 3.7% below their end-2019 level despite a 29.5% larger club estate, and center revenue per indoor square foot has risen 25.3% over six years — roughly cumulative CPI. In real terms, revenue productivity per square foot of real estate is approximately flat since 2019.

The 2026 commitment. Management guides FY2026 capex of $1,145–1,205M (growth $875–915M, maintenance $140–150M, modernization/technology $130–140M) against adjusted EBITDA of $925–940M. Capex exceeds EBITDA by roughly $240M at the midpoint, and management has raised its sale-leaseback target from a $300M minimum to ~$400M to bridge it — explicitly, in its own words, “supporting our ongoing focus on generating annual positive free cash flow.” More than half of 2026 growth capex is for clubs opening in 2027 and beyond. The 2026 class averages ~94,000 sq ft versus ~66,000 in 2025, is predominantly ground-up development, and is planned at 3,500–4,000 memberships per club versus 4,400–4,600 today.

Where the market is. At $42.33 the stock sits 1.2% below an all-time high after a 55% three-month advance, on 2.99x book (99.3rd percentile of its own history) and 3.11x sales (95.9th percentile). Enterprise value is ~$10.8bn on financial debt or ~$13.5bn including $2.65bn of lease liabilities. That is ~11.6x FY2026E adjusted EBITDA, or ~28x FY2026E earnings once stock compensation is charged rather than added back.

The synthesis. This is not a fraud, not a melting ice cube, and not a short. It is a good operator with a genuine local moat, run by a founder who has built the same product for 34 years, executing a strategy that consumes more capital than it produces and earns a sub-WACC return on the capital it consumes — priced, for the first time in its public life, as though neither of those things were true. The bull case requires the deceleration to stop and the 2030 free-cash-flow inflection to arrive on schedule. The bear case requires only that the current trajectory continue.

2. Business Overview

2.1 What the company actually sells

Life Time sells a subscription to a physical place, and then sells more things inside that place. The physical place is deliberately extreme in scale: the average center exceeds 96,000 indoor square feet (18.4M sq ft across 190 centers at 31 March 2026), and a typical build includes indoor and outdoor aquatics, a full basketball court, tennis and pickleball courts, a climbing wall, a full-service spa (LifeSpa), a café (LifeCafe), childcare (Kids Academy), group fitness studios, and increasingly a co-working floor (Life Time Work) and a longevity clinic (MIORA). Roughly 220–260 employees staff each center; the company employed over 44,000 people at year-end 2025, of whom over 33,000 are part-time and over 11,100 are certified fitness professionals.

This is not a gym. The closest structural analogue is a suburban country club that happens to be open to anyone who will pay $230 a month, and the competitive consequence of that positioning runs through the entire report.

2.2 Revenue segmentation

The mix is stable and heavily recurring: roughly 73% of center revenue is monthly subscription dues, the remainder is discretionary in-center spend — personal training (the largest single component), spa, café, kids’ programming, racquet lessons and retail. “Other revenue” captures Life Time Work, Life Time Living rents, media and athletic events.

The recurring share is genuinely high-quality. Dues are billed monthly by electronic funds transfer, cancellation is generally month-to-month after an initial period, and management reports that 2025 was the best retention year in the company’s 34-year history. That last claim is unverifiable — Life Time has never disclosed a numerical churn or attrition rate across any of the six earnings calls reviewed, despite claiming record retention on four of them. That is a material disclosure gap in the single most important operating statistic for a subscription business, and it is noted here as an Open Question rather than accepted.

2.3 The operating KPIs that matter

Read across that table and the business model becomes legible. Between FY2023 and FY2025 Life Time added 18 centers (+10.5%) and 59,164 center memberships (+7.7%) — meaning memberships per center fell 2.5% — while average revenue per membership rose 25.7%. In Q1 2026 the pattern intensified: memberships per center of 4,410 against 4,591 a year earlier, down 3.9%, with revenue per membership up 10.2%.

This is a business that has decided, deliberately, to serve fewer people at a much higher price. Management is explicit about it. Akradi on the Q4 2025 call: “Parking lots are packed, people are coming in… we basically want to optimize the membership so that we are making sure the customer experience in no shape or form deteriorates… you can have maybe a fewer less members for that optimal deal. Therefore, the only way you can do that is really raise the membership prices. And we are doing that really to protect the customer experience.”

Taken at face value that is a defensible, even admirable, strategy — protect the product, charge for it. The analytical question is whether it is a strategy or a constraint, and Section 5 addresses that directly.

2.4 The adjacencies

Life Time is building several businesses inside its clubs. In order of current materiality:

• Dynamic Personal Training (DPT) — the real in-center engine. Sessions grew ~18% in each of the last two years; comparable PT revenue in Q4 2024 was “nearly triple” the prior year; trainer headcount is up low double digits. Management notes some trainers are “booked solid,” which is a capacity constraint, not a demand problem.
• MIORA — longevity, hormone and GLP-1 clinics. Grew from 2 locations in mid-2025 to 7–8 by February 2026. Management’s ambition is one in every club eventually (“just like we have personal training in every club”), sized at “at least 50% of our revenue of our personal training.” It is early, and management concedes execution friction: “we’ve had obviously some challenges with some of those openings with some knick-knack things left over on construction or permits,” plus HIPAA and regulatory complexity the core club business never carried.
• LTH nutritional supplements — +31% y/y, with an Akradi aspiration of “a billion-dollar revenue business in the years to come.” That aspiration is unsupported by any disclosed base. Management was candidly negative on the digital channel: “it’s mediocre. It’s so-so.”
• Life Time Digital / L•AI•C (“Lacy”) — grew from 1.7M to ~3.3M subscribers. These are free, non-paying, non-club accounts, and management is explicit that monetization is “not in the near term.” It should be valued at zero today and treated as a funnel.
• Life Time Work and Life Time Living — co-working and luxury residential attached to clubs. Living is 6–7 locations. Notably, Akradi concedes the returns are inferior to the core: “even if we are 25% superior to the apartment business… it’s still extremely below the IRR of our club operations. Therefore, what we are working very, very diligently on is having a different vehicle that basically does not use Life Time’s money to build it.” Management deserves credit for saying so plainly, and for pushing the Paradise Valley project to outside capital.
• Racquet sports — 800+ dedicated pickleball courts and 5.9M participations in 2025. Strategically this matters more than its revenue line, because leagues are a retention mechanism (Section 4).

Verdict on the business overview. Life Time is a high-quality, genuinely recurring, genuinely differentiated consumer subscription business wrapped around a very capital-intensive physical asset. The revenue is real, the mix is stable, and the adjacencies are mostly sensible extensions of an owned customer relationship rather than diversification for its own sake. The single most important disclosure gap — churn — is conspicuous, and management’s refusal to quantify retention while repeatedly claiming it is at a record should be held against the file, not waved through.

3. Industry Dynamics

3.1 The industry is larger and healthier than at any point in its history

Demand is not the problem in this industry, and the post-COVID recovery is complete. For scale, Life Time’s 888,000 memberships are about 1.1% of the 81M US member base, while its FY2026 revenue guidance of $3.32–3.35bn is roughly 7% of the ~$47bn pool — a six-fold revenue-share-to-member-share ratio that is the arithmetic definition of the premium niche.

3.2 The profit pool is barbelling, and the middle is disappearing

The mid-tier collapse is the most important structural fact for Life Time, and it is unambiguously good: LA Fitness is closing boxes across Texas, Ohio, Virginia, Minnesota and California and converting others to its Club Studio format; 24 Hour Fitness never recovered its pre-bankruptcy footprint. Capacity is being removed from the segment directly beneath Life Time. Meanwhile at-home digital — the substitution threat that dominated the 2020–21 debate — is in outright decline, with Peloton subscriptions down 8% year-on-year to 2.66M in a fifth consecutive year of revenue decline. The pandemic-era bear case that technology would disintermediate the physical club has been decisively falsified.

3.3 The capital cycle — three speeds in one industry

Applying Marathon’s supply-side lens produces an unusually clear picture, because the industry’s tiers are at completely different points of the same cycle.

Boutique and Pilates: boom rolling into bust. This is the textbook case. Bodybar is targeting 70+ openings in 2026, JETSET has awarded 350+ territories, Pilates Addiction is targeting 100+, and solidcore is going from 175 to 250 studios while exploring a sale — against US reformer-Pilates industry revenue that declined 0.8% in 2026 on studio-count growth of just 0.2%. Xponential Fitness closed 140 units against 341 gross openings in 2025 (47 closures in Q4 alone), guided 2026 revenue ~16% below 2025, settled with the FTC for $17M — the largest franchise-case consumer redress in history — plus ~$22.75M in a franchisee class action, and fell 48%. Low barriers, cheap buildout, abundant franchise capital, no scale economies: capital arrived and returns left.

HVLP: late boom, first crack visible. 2026 supply additions across Planet Fitness (180–190 new clubs on a 2,896 base), Crunch (~100/year after Leonard Green took majority control from TPG Growth in April 2025), EoS (225+ open or committed, targeting 250 by 2030), Chuze (doubling in 5–6 years) and VASA total roughly 330–350 new boxes. Then, in Q1 2026, Planet Fitness missed its membership targets and cut same-club-sales guidance from 4–5% to approximately 1% — while holding unit guidance constant. The stock fell 31%. That is Marathon’s signature exactly: supply plans held fixed as same-store demand rolls over, each operator responding to its own demand signal without pricing in the other four building in the same trade areas.

Mid-tier: bust and exit. Good for whoever survives.

Premium: mid-boom, and capital is arriving. Equinox raised ~$1.8bn in March 2024 (Sixth Street, Silver Lake, Ares, HPS, L Catterton and the Related principals), refinancing $1.2bn of COVID-era debt and funding growth against 107 clubs with a 25+ location pipeline — roughly 20%+ footprint growth. Bay Club acquired 425 Fitness in Seattle and is expanding west. LA Fitness’s Club Studio is going from 15 to 50 in twelve months. Ultra-luxury entrants like Continuum in New York are pricing at $10,000/month. Life Time itself is accelerating to 12–14 clubs a year with a 2026 class of ~1.2M square feet.

That is a genuine yellow flag. But the absolute numbers differ by an order of magnitude: roughly 40–60 new premium large-format boxes in 2026, versus 330–350 HVLP and 500+ boutique studios. The binding constraint on premium supply — entitled 100,000 sq ft sites in affluent MSAs plus $50–70M of gross capital per box — is not one that cheap franchise debt can relax. Scored against Marathon’s warning-sign checklist, the industry lights four of five (M&A at premium valuations, sponsor-backed unit-growth arms race, expanding franchise lending, rising sell-side attention); the absent one is a rash of sector IPOs, and Xponential’s collapse has effectively closed that window. For premium specifically the score is two or three of five.

3.4 Does the HVLP land grab threaten a $230/month operator?

Tested directly, and the answer is largely no — the customer is different. HFA’s visitation data show luxury visit frequency rising in the same year HVLP supply surged. Life Time took 10.2% ARPM growth in Q1 2026 with total visits up (32M versus 30M) in the same quarter Planet Fitness cut same-club sales to ~1%. Lincoln International’s Kyle Perreira described the dynamic plainly: “the mid-tiers dropped out and the consumer either went down to HVLPs… or consumers… go to the Life Times, Equinoxes and Bay Clubs.”

The real premium-tier competitive risk is not Planet Fitness. It is Equinox, Bay Club and Club Studio building in the same affluent trade areas, and trade-down inside Life Time’s own base in an affluent-wealth shock.

3.5 Demand drivers

GLP-1 drugs are net accretive, and the mechanism specifically favours premium operators. A William Blair survey of 300 GLP-1 users found gym membership rose 3pp to 35% after initiation (versus +2pp to 24% in the 2024 survey), 72% reported working out more often (versus 60%), and among users under 45 reported club membership jumped to 62% from 43%. Seventeen operators now run GLP-1 programmes against a user base projected to reach 30M Americans by 2030. The clinical logic is clean: GLP-1s cause lean-mass loss, so protocols push users toward supervised resistance training — precisely what a $230/month full-service club sells, and can monetise twice through dues plus personal training plus MIORA. Caveat: n=300, sponsored, self-reported; directional only. The 2023-era bear case that a weight-loss pill would empty the gyms is not supported by the available evidence.

Affluent-household concentration is both the tailwind and the concentrated risk. Households earning $250k+ now drive 49.7% of all US consumer spending, up from 36% thirty years ago, and the top quintile generated more than 60% of discretionary spending growth. Bain projects luxury experiences growing 3–7% in 2026 versus goods at 1–4%. Life Time is levered simultaneously to the strong leg of the K-shaped consumer and to the goods-to-experiences rotation. That is a real structural tailwind — and it means the business is a bet on affluent-household wealth effects (equities and home equity), not on the median consumer. A 30% equity drawdown is a more relevant risk factor here than a recession in wages.

Family, youth and racquet demand is where the durability lives. Pickleball facility membership grew 21.3% year-on-year to 7.6M; participation among ages 6–12 doubled to 2.2M and teen participation rose 157% over two years; the US youth sports market is roughly $40bn annually at ~$1,000 per athlete. Adults 65+ were the fastest-growing membership demographic at +8.6%, which maps directly onto Life Time’s ARORA programming. Meanwhile traditional HIIT and stand-alone cardio — the products most easily substituted — continue to soften.

3.6 Regulation

The FTC’s Negative Option (“click-to-cancel”) Rule was vacated by the Eighth Circuit in July 2025, days before enforcement, on the procedural ground that the FTC failed to conduct the required preliminary regulatory analysis. On 11 March 2026 the FTC issued an Advance Notice of Proposed Rulemaking asking whether and how to amend the rule; comments closed 13 April 2026. The vacatur did not end exposure — ROSCA and Section 5 of the FTC Act remain, and the Commission has a live complaint against Fitness International (LA Fitness) alleging consumers could cancel only in person or by certified mail at their own expense.

The important analytical point is that cancellation friction is asymmetrically an HVLP issue. That model’s economics depend on a large never-visits cohort, which industry data show has already collapsed to an all-time-low 4.6%. Life Time’s members visit roughly 39 times per quarter; the company does not need friction to retain a member. A restored click-to-cancel rule compresses HVLP lifetime value considerably more than premium lifetime value.

State health-club statutes impose a mild but genuinely pro-incumbent fixed cost. Life Time’s own 10-K describes requirements to “include certain terms in our membership contracts, including the right to cancel a membership, in most cases, within three to 10 days after joining; escrow funds received from pre-opening sales or post a bond or proof of financial responsibility; and adhere to price or financing limitations.” Illinois requires registration and a bond for contracts longer than a year; Washington requires prepayments held in trust; Virginia releases escrow only after 30 days of operation; New Jersey caps contracts at three years.

Labour is the material recurring cost pressure. With 44,000+ employees and 220–260 per centre, and 22 states raising minimum wages in 2026 (California to $16.90, Connecticut $16.94, New York City and surrounding counties to $17.00, with 13 states and 44 localities indexing to CPI), wage inflation is running 2.5–3% on Life Time’s centre labour. This too is asymmetric: labour is a far larger share of a much smaller revenue-per-member at the $15–30/month tier.

Finally, MIORA imports genuinely new regulatory surface — prescribing, telehealth, state medical-practice rules and HIPAA — that the core club business never carried. Small today at 7–8 locations; worth monitoring as it scales.

3.7 Verdict — structurally bad industry, genuinely defensible local niche

At the aggregate level this is a bad industry and the record proves it. Run Greenwald’s tests honestly: you cannot count the top firms on one hand (55,000+ facilities, no operator above ~10% revenue share); entry and exit are constant; no operator has held stable share over five to eight years; and the commodity tier has no identifiable source of advantage. The 2020 graveyard — 24 Hour Fitness, Gold’s Gym, Town Sports, YouFit — plus Xponential live in 2026 is the empirical proof. High fixed costs, high capital intensity, low switching costs, no barriers to entry.

But Greenwald’s own instruction is to think local. The relevant market is not “US fitness”; it is a ten-to-fifteen-minute drive-time ring around an affluent suburb. That is where the only genuine advantage available in this industry can exist — economies of scale combined with customer captivity, the strongest of the three genuine advantage types. Section 4 tests whether Life Time actually has it.

The honest structural bear is not about competition; it is about the capital cycle. Life Time’s growth is a spread trade: build a box at $50–70M gross, recycle most of that capital to net-lease buyers at a cap rate it does not control, keep the operating spread. Single-tenant net-lease cap rates compressed to 6.80% in Q1 2026 but ticked wider in Q2 2026 (retail +5bp to 6.60%, industrial +10bp to 7.25%). Rising capital intensity, plus rent expense up $34.2M in 2025, plus a deliberate shift to larger and costlier ground-up boxes, is precisely the asset-growth pattern Marathon associates with poor forward returns even in good businesses.

Net: poor at the industry level, good-but-deteriorating-at-the-margin in the premium sub-segment. The industry does not kill Life Time. A cap-rate backup, a construction-cost overrun, or an affluent-household wealth shock would.

4. Competitive Position

4.1 Naming the mechanism in Greenwald’s taxonomy

Three genuine competitive advantages exist: supply/cost, demand/captivity, and economies of scale combined with captivity. Life Time’s case must be made in the third, and it must be made locally.

Supply or cost advantage — none. Labour (44,000 employees), suburban real estate, fitness equipment, energy and construction are commodity inputs purchased in competitive markets. Life Time’s in-house architecture and design capability is real and produces genuine build efficiency, but Greenwald is explicit that operational effectiveness of this kind is emulable and is not a barrier. No patents, no proprietary process, no privileged resource access. Dismissed.

Brand and intangibles — a real demand generator, not a standalone moat. The test is unambiguous: does the brand permit price increases without volume loss? Section 4.2 runs it properly and the answer is a partial pass at best.

Network effects — claimed implicitly, not real. The community framing (51,800 events in 2025, leagues, LT Games) implies network value. Economically it is a local density effect that saturates at a few thousand members per club. There is no cross-club or cross-user value that rises with Life Time’s national scale. Multi-club access for relocating members and brand recognition in site selection are conveniences. Label: speculative.

Local economies of scale plus partial captivity — the real advantage. The mechanism is concrete. A ~97,000 sq ft facility costs $60–90M gross to build (Life Time’s Brea Mall club cost $90M for 123,000 sq ft), sits on a scarce entitled prime site in an affluent MSA, takes two to four years to entitle and construct and three to four more to ramp, carries roughly $2.53M/year of rent per leased centre and 220–260 employees, and serves about 4,350 memberships. A second premium full-service operator entering the same five-mile ring splits demand and neither covers fixed cost. That is precisely Greenwald’s local-scale-plus-captivity combination, the strongest and most durable of the three types.

The evidence it is real: comparable-centre revenue of +15.3%, +12.2% and +11.1% for three consecutive years (+8.6% in Q1 2026), and adjusted EBITDA margin from 24.2% to 25.8% to 27.5%. You cannot post those numbers in a genuinely contestable market.

The qualifier that matters most: this is roughly 189 separate local franchises, not one national moat — and Greenwald’s own warning applies directly. Market growth is the enemy of economies-of-scale advantages. Life Time’s stated expansion is toward denser urban and vertical formats — New York, Boston, Chicago, coastal California — catchments that do support multiple premium operators and where Equinox already sits. The moat is strongest exactly where Life Time is growing least.

4.2 The pricing-power test, run properly

Management’s new Q1 2026 disclosure makes this testable for the first time. Against +8.6% comparable-centre revenue:

Pure list-price power is running at roughly 3.0% — approximately inflation. The other 5.8 points are mix upgrade and share-of-wallet extraction. Both are genuine value creation, and mix migration in particular is skilful management. Neither is brand pricing power in Greenwald’s sense.

The six-year view sharpens it. Effective dues per member have risen roughly 81% since 2019 (~10.4% CAGR) — alongside a 3.7% decline in absolute centre memberships despite a 29.5% larger club estate, with real revenue per square foot flat. Management is candid that volume was deliberately traded: “our newer centers are typically reaching their desired utilization and revenue with fewer memberships and in certain centers we are limiting qualified memberships.”

That is a defensible strategy. It is not the “raise price, keep everyone” signature of a true intangible moat. And it has a defined end: management guides pure price to a sustainable 2–3% long term, which means the 6–8% long-term comp target requires mix and in-centre attach to keep delivering 4–5 points indefinitely.

4.3 The captivity test — situational, and weaker than the narrative

The evidence for captivity is genuine. Visits per membership run about 149 per year and total visits rose 18.4% over two years (103M → 122M) against 7.8% membership growth — usage intensity is rising, which is the opposite of a business quietly losing its customer. Roughly 60% of memberships are couples or family. Junior add-ons run $30–100/month. Cancelling a Life Time membership means disrupting a swim progression, a summer camp booking, a pickleball league and childcare arrangements — several people’s calendars, not one person’s. The on-hold tier is an explicitly engineered switching cost: $15/month preserves the right to convert back without paying a new enrolment fee, and 50,556 memberships (5.8% of total) sit parked there. Life Time is paying to keep a re-entry ramp open, which is a sophisticated retention mechanic.

The evidence against is decisive, and it comes from the company’s own filings.

The IPO S-1 contains the only attrition figure Life Time has ever published: “our attrition rate for the first six months of 2020 was approximately 29.1% compared to 16.1% during the first six months of 2019.” The 10-K separately notes attrition is elevated in Q3 and Q4 “as the summer pool season ends and we enter the holiday season.” So 16.1% describes the seasonally best half — implying roughly 30%+ gross annual attrition in a normal pre-COVID year, which is approximately the industry average (HFA benchmarking implies ~33.6%; IHRSA ~28.6%).

Life Time has disclosed no attrition or retention rate in any 10-K since going public. Asked directly on the Q1 2026 call, the CEO answered qualitatively: “Retention is absolutely great… the more they use the club, the less they are likely to want to drop out.” That is a hypothesis, not evidence, and the company controls the data that would settle it. Claiming record retention on four consecutive calls while declining to quantify it is a disclosure choice, and it should be weighed as one.

Then there is the 2020 natural experiment, which is the single hardest piece of evidence in this file.

Planet Fitness peaked at 15.5M members in Q1 2020 and was at roughly 14M by October — about −10%, against Life Time’s −41.3%. The $15/month operator demonstrated materially more customer captivity than the $230/month one. The mechanism is intuitive once stated: a $15 membership sits below the psychological friction cost of cancelling, so people forget about it; a $230 membership sits far above it, so people act. Life Time’s captivity is engagement-based, and engagement-based captivity does not survive an interruption in access. Rent, meanwhile, rose 12.2% through a year of closures. The decremental margin was roughly 52.6%.

Structurally, nothing is contractually sticky: memberships are month-to-month, there is no annual commitment, no equipment purchase, no data lock-in, and state law mandates three-to-ten-day cancellation windows.

Verdict on captivity: situational. High for a family with three children in swim team and racquet leagues; near zero for a single adult using the weight floor. Life Time’s entire strategy — the mix shift toward couples and families, the deliberate run-off of qualified-medical memberships — is an attempt to re-weight the base toward the sticky cohort. That is coherent and so far successful. It is not the same thing as a moat.

4.4 Direct comparison — Life Time versus Planet Fitness

Life Time extracts roughly 29 times the revenue per member that the Planet Fitness system does — and earns less than half the return on capital. The premium positioning is entirely real at the revenue line and evaporates at the capital line. Planet Fitness converts unit growth into royalty with no balance sheet; Life Time converts unit growth into $60–90M of gross capital plus a perpetual rent claim. That single contrast is the whole difference between the two businesses, and it is the reason a superior product has produced an inferior investment return on capital employed.

The rest of the field is instructive in a different way. Equinox — the closest premium peer, 107–115 clubs with 25–40 more planned — required a ~$1.8bn rescue-style private raise in March 2024 led by Sixth Street and Silver Lake with Ares, HPS, L Catterton and the Related principals, to refinance $1.2bn of pandemic debt and fund growth. That is simultaneously a relative-strength point for Life Time (which self-funded through sale-leasebacks) and evidence that premium big-box economics are hard for everyone. Bay Club is acquiring regionally (425 Fitness, Seattle, June 2025). At the other end, Orangetheory recorded 26 confirmed US closures plus a full UK exit between August 2025 and May 2026, largely from a franchisee Chapter 7; Xponential guided 2026 revenue 16% below 2025.

Capital is leaving boutique and entering big-box premium. Near term that is a tailwind for Life Time — visible in in-centre revenue +15.1% and DPT sessions +18% as boutique refugees convert. Medium term it is a supply warning for precisely the format Life Time, Equinox and Bay Club are all committing capital to at once.

4.5 The two formal Greenwald tests

Market-share stability — passes locally, inconclusive nationally. Life Time’s share of US health-club industry revenue moved from roughly 5.4% (FY2019) to roughly 6.4% (FY2025) — a sub-two-point move over six years, nominally inside Greenwald’s “formidable barriers” band. But the share gain was purchased, not won: it cost roughly $2.1bn of net capex over 2023–2025. Share bought with capital is not evidence of a barrier. Locally the position looks stable and dominant within its format, but Life Time does not disclose trade-area competitive overlap — the one datapoint that would settle it.

Return on invested capital — fails, decisively and repeatedly.

Greenwald’s standard: sustained after-tax ROIC of 15–25% indicates advantages are present; 6–8% indicates they are absent. Life Time has never cleared the bar in its entire public record, in its best year ever, on any measurement basis.

The fair counterpoint deserves to be stated at full strength. ROIC is on a strong upward trend (2.8% → 5.2% in two years). Twenty-nine of 189 clubs — 15% of the estate — are still in their three-to-four-year ramp. A further $555.1M of construction-in-progress is capital earning nothing at all. Grant roughly $100M of incremental EBIT from the ramping cohort reaching maturity, on the ex-goodwill, ex-CIP base, and you reach approximately 7.9%. That is the most generous defensible figure. It is still single digits, and it is still below the ~8.1% cost of capital.

There is also a measurement subtlety that cuts against the bull. Incremental ROIC over FY2023–FY2025 computes to roughly 27%, which looks superb — but it is flattered because each sale-leaseback removes property, plant and equipment from invested capital while adding a matching lease liability, holding the denominator roughly flat. Measured from FY2019 instead, over six years, $1,637M of incremental capital produced $233M of incremental NOPAT: 14.2% incremental ROIC. Respectable, above the cost of capital, and the strongest single number in the bull case — but on a base that still averages 5%.

4.6 Verdict on competitive position — narrow, local, and being structurally traded away

Life Time has a genuine competitive advantage. It is local economies of scale combined with partial, mix-dependent customer captivity, and it exists at the level of the individual suburban club rather than the enterprise. That advantage is why Life Time takes 10–12% annual price increases with visits rising in the same year Planet Fitness cut same-club sales to ~1%.

But it fails the two tests that decide the question — the ROIC test outright, and the pricing-power test on a clean reading of the company’s own decomposition. And it is being converted, deliberately and continuously, into something less valuable: roughly 71% of clubs are now leased (84% of those opened since 2015), rent has risen from 8.7% to 11.3% of revenue, and 55 properties sit in 15 master leases with only 14 lessors, meaning an underperforming club cannot be handed back individually without risking the whole basket. Each sale-leaseback converts a scarce owned asset that captures the local scarcity rent into a leased operating spread in which the landlord captures the rent and Life Time retains a senior fixed obligation that does not flex in a downturn — as 2020 demonstrated, when rent rose 12.2% through a year of closures.

This is a good business with a real local barrier, run by people who are executing well on mix, engagement and margin. It is not a compounder at the enterprise level until returns on capital clear their cost — and nothing in the 2023–2025 trend suggests that happens soon.

5. Growth History and Forward Opportunities

5.1 The historical record, disaggregated

Growth is entirely organic — Life Time has made no material acquisitions, and its only meaningful disposal was the triathlon events business (a $4.9M gain in FY2023). That is a point in its favour: the record is a record of building, not of buying growth.

The decomposition matters more than the total. Between FY2023 and FY2025 revenue grew 35.1%. Of that, roughly ten percentage points came from adding 18 clubs; the remainder came from raising the price of the existing base. Membership units contributed almost nothing: 763,216 to 822,380 is +7.7% over two years, against a centre count up 10.5%. In Q1 2026 the company’s own decomposition put volume at −0.2pp of an 8.6pp comp.

The deceleration is monotonic and now management-guided. Comparable-centre revenue has printed 13.5%, 12.9%, 11.2%, 10.6%, 9.9% and 8.6% across the last six quarters. FY2026 guidance is 6.3–7.3%, against a long-term target of 6–8%. Management has told the market the algorithm steps down; the question is whether 6–8% is a floor or a waypoint.

5.2 Quality assessment of the growth

Three tests, and Life Time passes one and a half.

Is it profitable growth? Yes, at the P&L. Adjusted EBITDA margin has expanded 330bp in two years to 27.5%, and centre operations expense has fallen from 64.0% to 52.4% of revenue since 2019. Operating leverage is genuine and impressive.

Is it cash-generative growth? No. Positive CFO-less-capex in one of five years. The growth consumes cash and is financed by asset sales.

Does it earn its cost of capital? No, at 5.2% consolidated. On new units management claims in excess of 30% cash-on-cash — a claim Section 7 shows cannot be reconciled to disclosed portfolio arithmetic.

A further quality concern: the composition of comp growth is deteriorating even as the headline holds up. Half of Q1 2026’s comp came from mix — replacing cheap members with expensive ones. The largest cheap cohort, third-party-insurance qualified-medical memberships, fell 14.9% year-on-year to just 3.4% of dues revenue, guided to ~3% by year-end. That lever is arithmetically almost spent: it cannot contribute more than roughly three more points ever, and once it laps, reported volume must either turn positive on its own merits or the comp steps down again.

The legacy-to-rack-rate gap is the healthier lever and is genuinely durable. Legacy members pay roughly $30/month below rack rate; the aggregate pool has held at $17–20M per month ($19.5M exactly, per the CFO) for several years, and roughly two-thirds of members still pay below rack. Crucially, the pool is not draining — it is replenished each time rack rate rises. The CFO is explicit: “I don’t see a world where it’s ever closed. I mean, that’s part of the kind of the retention play, having members pay under the rack rate.” That is an honest and unusually useful disclosure: it means the price lever is a perpetual 2–3% annuity rather than a one-time catch-up, but also that it will never deliver more than 2–3%.

5.3 Forward opportunities

New clubs — the primary engine, now accelerating and getting more capital-intensive. The plan is 12–14 openings a year from 2026, up from a prior 10–12 baseline, with up to 28 across 2026–2027. The 2026 class is roughly 1.2M square feet — nearly double the 2024 and 2025 classes — at an average ~94,000–95,000 sq ft versus ~66,000 in 2025, with 11 of 14 large-format ground-up developments. Management sizes the domestic opportunity at 450–500 clubs and keeps 85–100 deals in the pipeline.

Two features of this plan deserve emphasis. First, more than half of 2026 growth capex is for clubs opening in 2027 and beyond — the spend is not buying near-term EBITDA. Second, the new clubs are deliberately planned for fewer members: 3,500–4,000 memberships versus 4,400–4,600 in the existing estate. Higher capital cost per box, fewer members per box, at a higher price per member. Whether that arithmetic improves or degrades unit returns depends entirely on whether the price premium more than offsets the density loss, and Life Time does not disclose enough to verify it.

In-centre revenue — the highest-quality growth available. Dynamic Personal Training is the real engine: 220,000+ sessions a month, +18% for two consecutive years, with trainer headcount up low double digits and some trainers “booked solid” (a capacity constraint, not a demand problem). Dynamic Stretch runs 20,000 sessions/month, +34%. This growth requires almost no incremental capital — it monetises floor space and staff already in place — and is the single most attractive part of the forward story.

MIORA (longevity and GLP-1 clinics) — 2 locations in mid-2025 to 7–8 by February 2026, with an ambition of one per club and a sizing of “at least 50% of our revenue of our personal training.” Genuinely interesting, genuinely early, and it imports new regulatory surface (prescribing, telehealth, HIPAA) the core business never carried. Management concedes execution friction on openings. Treat as option value, not base case.

Supplements, digital and real estate adjacencies — discount heavily. The LTH supplement line grew 31% but management calls the digital channel “mediocre. It’s so-so,” and the “billion-dollar revenue business” aspiration is unsupported by any disclosed base. Life Time Digital’s ~3.3M subscribers are free, non-paying accounts with monetisation explicitly “not in the near term” — value them at zero and treat them as a funnel. Life Time Living’s returns are, by the CEO’s own account, “extremely below the IRR of our club operations,” which is why he is pushing it to third-party capital. Credit management for saying so plainly.

The honest summary: non-club revenue has been stuck at 2.8–2.9% of total for three consecutive years. The ecosystem is a retention garnish and a real-estate deal-sourcing apparatus — management says as much, noting that Life Time Living “is generating interest from new property developers and presenting opportunities for new center development and deal terms that were not previously available to us.” It is not a diversifying revenue stream.

5.4 Verdict on growth — real, high-margin, decelerating, and cash-consumptive

The growth is organic, the margin expansion behind it is genuine, and the in-centre attach story is high quality and capital-light. But the engine has shifted almost entirely from volume to price and mix; the mix component is nearly exhausted; pure price is running at inflation; management guides the comp down to 6.3–7.3%; and the unit-growth leg is getting more expensive per box while planning for fewer members per box. Growth that consumes $945M of cash over five years and earns 5% on capital is low-quality growth at the enterprise level regardless of how good the margin trend looks.

6. Financial Quality

6.1 The income statement, and what is actually in it

The operating-leverage story is genuine and is the strongest part of the file. Centre operations expense has fallen from 64.0% of revenue in 2019 to 52.4% in Q1 2026. Adjusted EBITDA margin has gone from 24.2% to 27.5% in two years and reached 28.7% in Q1 2026. That is not accounting; that is a fixed-cost base being covered by a rising average revenue per member.

But two lines require normalisation before any earnings-based conclusion, and they run in opposite directions across time.

(a) Sale-leaseback gains sit inside operating income — and in FY2022 they were most of it. Life Time books gains and losses on sale-leaseback transactions within “Other operating expense,” i.e. above the operating-income line:

In FY2022, 88% of reported operating income was a sale-leaseback gain. Anyone anchoring on FY2022 as a base year is anchoring on a property disposal. The good news for the bull case is that this distortion has since become immaterial: the 2023-to-2025 margin expansion — 10.9% to 15.8% normalised — is real operating improvement, not asset sales. That deserves to be stated clearly, because it is the single most common bear error on this name.

(b) FY2025 net income is materially flattered below the operating line. The FY2025 income statement shows interest expense, net of just $82.3M — against $148.1M in FY2024 — and an “Other income” line of $94.2M where FY2023 and FY2024 both showed zero. Both are explicable and both matter.

The interest decline is real and durable: the November 2024 refinancing replaced 5.750% and 8.000% notes with a $1,000M term loan B and $500M of 6.000% secured notes, April 2025 swaps fixed the full $997.5M term-loan notional at 3.409% (a 5.409% effective rate), and FY2024 had carried $13.8M of extinguishment charges inside interest expense. Credit management for the balance-sheet work; the BB rating earned in Q2 2025 came a year ahead of plan.

The $94.2M is not durable at all. The 10-K MD&A is explicit: “related to $54.6 million in net cash proceeds received in connection with employee retention credits under the CARES Act… and a $39.6 million payment by Zurich in partial satisfaction of legal claims.” A COVID-era payroll tax credit collected four years late, and an insurance settlement from litigation the Minnesota Court of Appeals reversed in August 2025. Add a $12.6M discrete tax benefit from the CEO’s exercise of expiring options, and the normalisation is:

Roughly 24% of FY2025 GAAP EPS is non-recurring. The cross-check is tight: the company’s own Adjusted net income of $325.5M, less its $51.8M stock-compensation add-back tax-effected at 24.28%, equals $286.3M — within 0.3% of the $285.6M above. The difference between $1.27 and the company’s $1.44 is entirely the stock-compensation add-back, which is a real cost to shareholders and should not be added back.

One further quarter-level distortion worth flagging for anyone comparing Q1 2026 to Q1 2025: Q1 2025’s effective tax rate was 7.8% ($6.4M on $82.5M pre-tax) because of the same CEO option-exercise windfall, versus 26.1% in Q1 2026. Q1 2025’s reported $0.34 was really about $0.27. Q1 2026’s $0.39 is clean — which means the underlying year-on-year growth is better than the optics, a rare instance where the distortion favours the bull.

6.2 Cash generation — the crux of the entire report

This table is the report. Read the two bolded rows against each other.

Life Time has generated positive cash from operations less capital expenditure in one of the last five years, and that year — FY2024, at +$50.6M — was the trough of its growth-capex cycle. Cumulatively over five years the business has consumed $944.5M more cash than it produced, and has received $982.5M from selling its own buildings. The deficit and the asset sales match to within 4%. That is not a coincidence; it is the financing structure of the growth plan.

The FY2025 number deserves one further adjustment. The −$21.0M of true free cash flow includes the $94.2M of employee-retention credits and Zurich proceeds sitting inside operating cash flow. On a recurring basis, FY2025 organic free cash flow was approximately −$115M — in the best operating year in the company’s history, on record revenue, record adjusted EBITDA and record margins.

Management’s non-GAAP “free cash flow” adds back sale-leaseback proceeds. This is the most aggressive presentation in the filing, and it is not a technicality: a sale-leaseback is a financing, not an operating inflow. Each dollar received buys a perpetual, escalating rent obligation that Life Time discounts at 9.28% over a weighted-average 16.5-year term. Calling the proceeds “free cash flow” while expensing the rent below it counts the benefit once and defers the cost — and the CFO’s own framing on the Q1 2026 call makes the dependency explicit: the April sale-leasebacks were “supporting our ongoing focus on generating annual positive free cash flow.”

What the mature estate actually throws off. The fairest read of underlying cash generation strips growth capex, which is discretionary: FY2025 operating cash flow of $870.5M less maintenance capex of $125.8M and modernisation/technology capex of $109.2M gives $635.5M, or 21.2% of revenue — before the $94.2M of one-timers, so call it roughly $541M recurring. That is what the 189-club estate produces before deciding to build anything. It is a real number and it is a good one. The question the valuation must answer is what that stream is worth given that management intends to spend $875–915M a year on growth against it.

6.3 Balance sheet — 79% of the fixed obligations are leases

Funded debt stands at $1,525.4M face: a $992.5M term loan B (SOFR + 2.00%, swapped to 3.409% for a 5.409% effective rate, maturing November 2031), $500M of 6.000% senior secured notes (November 2031), and $29.4M of mortgages at a weighted-average 5.47%. The $650M revolver is undrawn with $618.2M available to September 2029. Total liquidity is $823.0M. Maturities are trivial until the cliff: $21.8M in 2026, $27.8M in 2027, $10.1M in 2028, $10.1M in 2029, $10.2M in 2030, and $1,445.1M thereafter. This is competent liability management and there is no near-term refinancing risk.

The leases are the real balance sheet. Lease liabilities total $2,686.1M (of which $2,634.7M operating), at a weighted-average remaining term of 16.5 years and a weighted-average discount rate of 9.28%. Undiscounted future lease payments are $5,249.7M, plus a further $545.8M of signed-but-not-commenced leases — $5,795.5M of contracted future rent.

Management markets the 1.60x figure and has a stated ceiling of 2.0x. On a rent-adjusted basis — the only basis on which a 71%-leased operator can be honestly compared to an owned-estate operator — leverage is roughly 3.45x, more than two turns higher. Neither number is alarming for a business with this cash conversion; the point is that the headline understates the fixed-claim burden by about 2.2 turns, and that the gap widens with every sale-leaseback.

Concentration risk in the lease book is underappreciated. Of 185 leased properties, 55 sit under just 15 master leases with only 14 lessors. A master-lease basket means an underperforming club cannot be handed back individually without jeopardising the whole basket. The operational flexibility a lease is supposed to purchase has, for those 55 properties, been contracted away. Goodwill and intangibles of $1,416.2M represent 45% of book equity — a 2015 LBO artifact carrying real impairment risk in a downturn.

Real estate. 189 centres: 55 owned (including ground leases), 134 leased (71%), with 84% of centres opened since 2015 leased. Management asserts roughly $3.5bn of owned real-estate market value against $1.5bn of debt. That figure is management’s estimate, repeated without support, and is not reconcilable from the filings. It is flagged as an Open Question, not adopted.

6.4 Unit economics

Centre-level four-wall economics for FY2025: centre revenue $2,908.7M less centre operations expense $1,568.6M less rent $339.2M gives $1,000.9M, or $5.30M per club — a 34.4% four-wall margin. Before rent, contribution is $7.09M per club (46.1%). These are strong unit-level numbers and they explain why management is enthusiastic about building more.

They also explain the gap that Section 7 has to resolve. Management’s stated unit model is net invested capital of $25–30M per new location — defined as gross invested capital, net of construction reimbursements, less net proceeds from sale-leaseback transactions — earning a portfolio cash-on-cash return “in excess of 30%” after a three-to-four-year ramp. On $27.5M of net invested capital, $5.30M of four-wall EBITDA is 19.3%, before allocating any share of $583.8M of corporate overhead. Even granting that the 29 ramping clubs run at ~60% of mature economics — implying mature four-wall EBITDA of roughly $5.64M — the figure is 20.5%. The claim and the disclosed arithmetic differ by roughly ten percentage points, and the reconciliation is not in any public document.

6.5 Returns on capital — the verdict-determining metric

Reconstructed independently rather than accepted from an aggregator: normalised FY2025 EBIT of $474.3M, taxed at the 24.28% effective rate, gives NOPAT of $359.1M. Invested capital — equity plus funded debt plus lease liabilities less cash — is $7,115.3M at year-end and $6,866.0M on average.

Against an estimated WACC of ~8.1% — built as 60% equity at 9.5%, 13% funded debt at 4.1% after tax, and 27% capitalised leases at 7.0% after tax. Two independent methods converge on the same answer, and the gap is negative on every basis except the most generous hypothetical.

The single most important sentence in this report: in the best year of its corporate history — record revenue, record margins, record EBITDA, its lowest-ever leverage and an investment-grade-adjacent credit rating — Life Time earned less on its capital than that capital costs.

6.6 Verdict on financial quality — excellent operations, unproven economics

The operating improvement is real, large and management-driven: centre operations expense down 1,160bp as a share of revenue since 2019, adjusted EBITDA margin up 330bp in two years, interest expense structurally reduced, leverage halved, a BB rating a year early. Nothing here suggests an accounting problem, and the FY2022 sale-leaseback distortion — which some bear cases still cite — has genuinely washed out of the numbers.

But the economics do not clear the cost of capital, and they never have. The business has produced positive cash from operations less capex in one of five years. Growth is entirely price and mix with volume per club shrinking about 4%. Roughly a quarter of FY2025 earnings is non-recurring COVID-legacy income. And the reported “free cash flow” that anchors the bull case is, by the company’s own construction, substantially the proceeds of selling the assets it just built.

7. Capital Allocation

The one-sentence version: Life Time spends every dollar it earns and then some on building clubs, funds the shortfall by selling the buildings it just built at roughly 7.9% cap rates while it can borrow secured at 5.4%, reports the sale proceeds as “free cash flow,” measures its leverage on a basis that excludes the $2.65bn of lease obligations those sales created, and pays management 100% on Adjusted EBITDA — a metric that rises with every club opened regardless of the capital consumed.

7.1 The arc: 2015 LBO → 2021 IPO → 2026 sponsor exit

Leonard Green & Partners and TPG (with LNK Partners and later Partners Group) took Life Time private on 10 June 2015 at $72.10 a share, an enterprise value of “more than $4 billion,” on an aggregate equity check of roughly $1.37bn at an implied ~$10.00 per share basis. Two details matter. There was no dividend recapitalisation and no real-estate carve-out — the 2014–15 REIT-spin plan pushed by activists was abandoned in favour of the whole-company sale, and the owned real estate is precisely what made ~$2.6bn of LBO debt financeable. And between 2015 and the IPO, cash flowed into the company: $108.7M in 2019, $90.0M in 2020, and a $101.5M 12% COVID rescue loan in June 2020 that converted at roughly $16.24.

The IPO is widely mischaracterised and the correction matters. The October 2021 offering priced at $18.00 — the bottom of the $18–21 range — and was 39.0 million shares, 100% primary, with zero secondary. Including the partial greenshoe, the company raised $730.4M gross and $701.4M net, and applied $575.7M to term-loan paydown within a week. No sponsor sold a share. They bought $204.9M of it, taking 11,385,038 shares at $18.00 commission-free. The IPO was a de-levering plus a sponsor top-up, not an exit.

The exit came later, patiently, over nine transactions:

Roughly $2.60bn of liquidity, of which only ~$1.44bn came through marketed deals; ~$1.10bn went off-market through Rule 144 blocks, a negotiated private placement and the company’s own buyback. Ownership has gone from LGP 30.4% / TPG 22.3% at the IPO to LGP 4.95% / TPG ~3.5% as of June 2026 — 81% liquidated on both sides. TPG’s 13D/A filed 26 May 2026 states it has “ceased to be the beneficial owner of more than five percent.”

Three observations, and they cut against the lazy bear reading.

First, the sponsors’ weighted-average realised exit price was roughly $28.50–$28.76 against $42.33 today. Every sponsor sale was below the current price, by 30% to 95%. Gross MOIC was approximately 2.9x for both LGP and TPG over eleven years — an ~11.7% IRR, which for a 2015-vintage buyout is mediocre. Meanwhile the $18.00 IPO buyer has earned +135%, a ~19.5% CAGR over 4.8 years. The public shareholder out-earned the sponsors. Read the selling as supply and fund-life mechanics, not as a valuation signal.

Second, LNK Partners has never sold a single share. The same 10,501,477 shares appear in the 2022, 2023, 2024 and 2026 proxies, now worth roughly $444M on a ~$116M cost — approximately 3.8x. The sponsor without a fund clock made the most money. That is a quiet vote of confidence and the single most bullish item in the ownership record.

Third, and against the above: the last open-market purchase by any insider was 31 May 2024. Across the whole public life, insiders and sponsors have sold roughly 79.5M shares for ~$2,266M against about $4.0M of discretionary open-market buying — better than 500 to 1 — and there has been zero insider buying through 26 months in which the stock rose 150%. In 2026 the wider board and C-suite turned discretionary sellers for the first time (Lasher at $32.42, Almendares at $33.59 representing 52% of her holding, plus two EVPs; the highest insider sale on record is $38.65 in June 2026). No insider is signalling that $42 is cheap.

The founder’s record is the most nuanced piece. Bahram Akradi accounts for 97% of all insider buying — 273,411 shares for $3.90M at an average of $14.26, bought with his own cash across twelve open-market purchases in 2022–24 when the stock was cut in half. That is genuine conviction. His only sale, 5,000,000 shares at $30.13 on 27 February 2025, was mechanically driven: he net-exercised a 9,388,000-share founder option struck at $10.00 and expiring in October 2025, with 2,999,361 shares withheld at $31.30 — exactly $93.88M, precisely the strike cost. He retains roughly 16.3M shares (7.3%, ~$690M) and zero options. But the sale was discretionary rather than 10b5-1, and his total economic exposure including the retired option is down about 25% from the IPO.

7.2 Where the capital actually goes

The answer to “how is growth funded?” is unambiguous: by selling the real estate. Over FY2022–FY2025, operating cash flow of $2,109.6M covered 78.0% of capex of $2,705.2M. The $595.6M shortfall — $656.8M through Q1 2026 — was plugged by $908.5M of sale-leaseback proceeds. Over the same window net debt fell $475.0M, which is less than the sale-leaseback proceeds received: the buildings paid for the growth and the deleveraging. Strip the sale-leasebacks and Life Time would have needed roughly $600M more borrowing since the IPO, and reported leverage would be about 2.3x rather than 1.6x.

Only $124.0M of primary equity has ever been raised post-IPO — 6.0M shares at $21.75 inside the August 2024 follow-on. The stock is $42.33 today; the one time the company sold equity, it sold near the lows.

Gross growth capex per opening has risen from $40.9M in 2022 to $65.6M in 2025 — a 60% increase in three years, driven by the shift to larger ground-up formats and construction inflation. Net of sale-leaseback proceeds the figure ties precisely to management’s disclosed “$25–30M net invested capital per new location” ($25.3M on the FY2022–25 average). The disclosure is honest; it simply describes a number that only exists because the real estate is sold.

Marathon’s asset-growth anomaly is lit brightly. Total assets grew from $6,018M (2020) to $8,008M (2025), +33%. Capex is running at 3.01x depreciation. Construction in progress rose 148% to $555.1M. Free cash flow fell to $206.5M from $273.6M in a record EBITDA year. Rising capex-to-depreciation, a widening gap between reported earnings and true free cash flow, and a large secondary-issuance programme are three of Marathon’s four classic late-cycle warning signals, all present simultaneously.

The fair mitigant: this supply increase is company-specific, not sector-wide. Life Time is adding 5–6% of units a year into a genuinely underserved premium niche still posting double-digit comps. There is no visible flood of competing $65M athletic country clubs. This is not 2005 homebuilding.

7.3 The sale-leaseback programme — the arbitrage runs the wrong way

Three trends, all deteriorating. Price-to-book has collapsed from 1.31x (2022) to 1.04x (2025). Cumulative gains of roughly $97M are essentially all the 2022 vintage. And Life Time’s own disclosed weighted-average operating-lease discount rate has climbed from 7.97% in 2021 to 9.28% in 2025 — the terms are getting worse, not better.

The implied cap rate is the key derived number, and it contradicts management. Solving the FY2025 $222.9M operating-lease right-of-use addition as a 25-year, 2%-escalating annuity at Life Time’s own 9.28% discount rate gives year-one rent of roughly $19.75M on a $249.9M sale price — an implied initial cap rate near 7.9%. That is an estimate, not a disclosure, and it is flagged as such. But it sits well above the “6.5 to 6.7, 6.8 — it won’t touch 7” the CEO has repeatedly claimed, and the one transaction where the rent is disclosed — the 2024 related-party deal, $3.5M of rent on a $40.0M property — implies 8.75%.

Set that against Life Time’s actual cost of debt: a term loan swapped to 5.409% all-in and secured notes at 6.000%. The company is selling buildings at roughly 7.9% to fund growth it could finance with secured debt at 5.4%, and permanently forfeiting the freehold residual on 25-year leases with four to six five-year renewal options. Part of that gap is genuine term premium — a 25-year obligation against 2031 notes — so this is not a pure funding penalty. But on the pure financing arithmetic it is negative arbitrage, and it creates value only if the redeployed capital returns well above 8% after all corporate costs. The consolidated evidence says it currently does not.

What the programme does and does not flatter. It is not an earnings-flattering device in the way most bear cases assume: gains are excluded from Adjusted EBITDA, and the resulting rent is deducted, so each sale-leaseback mechanically lowers reported Adjusted EBITDA while raising cash. Credit where it is due. What it flatters is precisely the two things management is judged on — the company-defined free cash flow, which adds the proceeds straight back, and the reported net-debt leverage ratio, which excludes operating leases entirely.

And the advertised return has been quietly walked down and re-labelled. Reading five consecutive 10-Ks side by side:

The number fell from over 40% to over 30%, and the metric migrated from a return-on-capital measure to a cash-on-cash measure — pre-tax, pre-depreciation, pre-maintenance-capex, pre-overhead. Neither change was announced.

Even the walked-down target does not reconcile. FY2025 club-level EBITDA of $1,000.9M across 184 average clubs is $5.44M per club. On $27.5M of net invested capital that is 19.8% — and to reach 30% a club would need $8.25M of EBITDA, 52% above the current portfolio average, before a dollar of corporate overhead. Management’s own wording is “across our portfolio,” so the portfolio comparison is the fair one.

The leakage between $5.44M of club EBITDA and a mid-single-digit consolidated return is fully traceable: roughly $260M a year of corporate overhead the unit metric ignores; $1,235M of goodwill and $181M of intangibles from the 2015 LBO; 29 clubs open under three years plus 17 under construction earning nothing on capital already spent; and roughly $235M a year of maintenance and modernisation capex the “cash-on-cash” figure excludes entirely.

Verdict on the programme: competently executed, genuinely accretive at the unit level, and value-neutral at best at the enterprise level — with every observable term (cap rate, price-to-book, discount rate, cost per club) moving the wrong way.

7.4 Balance-sheet policy — the one unambiguous success

Net debt fell from $2,241M (2020) to $1,795M (2022) to $1,320M (2025); reported net-debt/Adjusted-EBITDA from 3.6x to 2.3x to 1.6x. The November 2024 refinancing was well executed: a $1,000M term loan and $500M of 6.000% secured notes due 2031 replaced $925M of 5.750% secured and $475M of 8.000% unsecured notes, at a cost of an $11.1M extinguishment loss. The revolver was lifted to $650M through September 2029; the term-loan margin was cut twice to SOFR + 2.00%; April 2025 swaps fixed the entire $997.5M notional at 3.409% for a 5.409% all-in rate; S&P upgraded on 18 June 2025, triggering a further 25bp step-down. Interest paid fell from $136.9M in 2024 to $85.4M in 2025.

This is genuinely good work and it is the honest core of the re-rating.

But the headline 1.6x excludes $2,648M of lease liabilities. Lease-inclusive net debt to Adjusted EBITDA is 4.79x; on an EBITDAR basis, 3.39x. The “dramatic improvement” management cites is substantially an artefact of moving debt-like obligations into leases via the very sale-leasebacks that fund the growth — which matters enormously, because that improvement was used as the stated reason to delete the leverage metric from the incentive plan.

7.5 Shareholder returns

No dividend has ever been paid, and the credit agreement and notes indenture restrict distributions.

A $500M repurchase authorisation was approved on 19 February 2026 — the first in the company’s history. Execution to date is roughly $73M, about 15% of the authorisation: ~0.4M shares for $10.7M at roughly $26.75 in Q1 2026, plus 2,192,500 shares at $28.60 for $62.7M in May 2026, purchased privately from Leonard Green, TPG and Partners Group concurrently with those sponsors placing 8,770,000 shares with Atairos at the same price.

That May transaction deserves real credit: the company bought stock at $28.60 that trades at $42.33 ten weeks later, executed off-market at roughly an 11% discount to the tape, while helping clear the sponsor overhang that had capped the stock for eighteen months. It is the best capital-allocation decision in the file. Two caveats keep it honest: it was a related-party purchase from affiliates with board designees, and the company deployed only 15% of the authorisation before the stock ran 48%.

It is worth noting the reversal that preceded it. Asked directly about buybacks in May 2025, Akradi said: “Definitely the latter [hold cash]… We’re not Amazon. We’re not Apple. We’re not JPMorgan.” By November 2025 it was “all options are on the table”; by February 2026 there was a $500M authorisation. Management changed its mind quickly and, as it turned out, correctly.

Dilution. Share count has risen from 193.06M (end-2021) to 222.70M (March 2026) — +15.4%, about 3.3% a year. Stock-based compensation was $51.8M in FY2025, falling as a share of revenue from 2.26% to 1.73% — modest and improving. One arresting comparison, though: FY2025 stock compensation of $51.8M exceeded true free cash flow of −$21.0M. Only the sale-leaseback-inclusive figure covers it.

7.6 Incentives — the explanation for everything above

The 2025 short-term incentive programme was based entirely on Adjusted EBITDA, with tranches at $740M / $760M / $780M against actual delivery of $825.2M — every named executive vested at maximum. Half of the long-term award is a three-year performance stock unit based on Adjusted EBITDA in 2025, 2026 and 2027; the 2025 tranche achieved 200% and the 2024-granted tranche 250%. The 2026 short-term programme is, in the proxy’s own words, “again based entirely on our Adjusted EBITDA for 2026.”

There is no return-on-capital metric, no free-cash-flow metric and no per-share metric anywhere in the 2025 plan. And the one balance-sheet-discipline metric was removed. The proxy is explicit: “The Company’s leverage ratio, which was utilized in the 2024 bonus program in addition to Adjusted EBITDA, improved dramatically during 2024 and the Compensation Committee and the Board of Directors determined that metric was not required again in 2025.” As defined — net debt over Adjusted EBITDA, excluding operating leases — that metric could be improved simply by selling buildings and leasing them back. It was retired the moment it was achieved.

This is textbook empire-building incentive design. A company whose entire capital-allocation question is “does the marginal club earn its cost of capital?” pays management purely on the size of Adjusted EBITDA — a metric that rises with every club opened, is indifferent to the capital consumed, and is struck before depreciation, before stock compensation, and before any return test. In Greenwald’s terms, management is paid for size in a business whose entire question is barriers and returns. In Marathon’s terms, it is precisely the incentive that drives the destructive half of the capital cycle.

Credit where due: 2026 introduces the first external metric. Performance stock units are now “equally split between (1) an internal metric of Adjusted EBITDA for 2026, 2027 and 2028 … and (2) a new external metric based on relative total stockholder return, measuring the Company’s three-year stock performance against the Russell 2000 Index.” That is a genuine improvement and the first sign the board has noticed. But relative TSR is still not a return-on-capital test, and it covers only about a quarter of total long-term incentive value.

Pay quantum. Akradi’s FY2025 total compensation was $15,199,475 against a median employee of $9,816 — a CEO pay ratio of 1,548:1, inflated by a 33,000-strong part-time workforce but arresting nonetheless. Perquisites include personal use of company aircraft, a company car with tax gross-up, and $42,000 of personal administrative support with a $35,205 tax gross-up. The quantum is defensible for a founder-CEO of a $9bn company; the structure is not.

7.7 Governance and related-party transactions

One genuinely clean item. The anti-pledging policy is absolute: officers, directors and employees are “prohibited from … pledging Life Time stock in any circumstance, including by purchasing Life Time stock on margin or holding Life Time stock in a margin account.” Zero pledges appear across the entire 256-filing Section 16 corpus. Akradi holds roughly $690M of unpledged stock.

The structural negatives are real and were built for a controlled company that is no longer controlled. The board is classified into three staggered classes and has been expressly retained as such. Director elections use plurality, not majority, voting. A two-thirds supermajority is required to remove directors for cause or amend key charter provisions. Life Time was a “controlled company” until 5 June 2025 and reached full independent-committee compliance only on 4 June 2026 — Akradi chaired the Nominating and Corporate Governance Committee until 17 September 2025. After the 2025 sponsor selling, LGP retains the right to nominate only one director and TPG and Partners Group none, yet the board declined to request resignations from four sponsor-designated directors; two have since resigned voluntarily. The board is now 12 members, 10 independent, and does maintain a standing Capital Allocation Committee. Akradi retains the right to nominate himself plus one additional director for as long as he is CEO.

The related-party real estate is the sharpest blemish. Life Time pays roughly $11.5M a year of rent to entities affiliated with the CEO and/or director Lasher, arising from seven sale-leaseback and lease arrangements struck between 2003 and 2024 — including an LLC in which Akradi owns 33% ($1.3M/yr), an LLC jointly owned by Akradi and Lasher ($2.4M/yr), and a partnership 100% owned by Akradi ($1.0M/yr). Initial terms run 20–25 years with four to six five-year renewal options.

The 2024 transaction is the one to flag. Life Time sold a property for $40.0M gross to an entity part-owned by the CEO and a director, against an estimated fair value of $40.3M, and recognised a $17.2M loss. The reason is disclosed: the company accounts for related-party sale-leasebacks “at their contractually stated terms” and does not adjust for off-market terms. The resulting rent is $3.5M a year on a $40.0M property — 8.75%, the highest implied yield in the entire programme. The $17.2M loss is the accounting system’s recognition that the lease was struck on above-market terms. Value moved from Life Time shareholders to an entity part-owned by the CEO and a director.

It was disclosed and Audit-Committee-supervised, and the sale price itself was at fair value. But the proxy discloses no competing bids, no process and no rationale, and it creates a reporting artifact worth knowing: the FY2025 MD&A describes a “$2.6M net sale-leaseback gain” for FY2024 while Note 10 shows $19.8M — the bridge is the $17.2M related-party loss in Note 14, and the MD&A wording obscures it. Separately, Meghan Akradi, the CEO’s daughter, is Vice President of Real Estate and Location Intelligence, with compensation rising from $0.2M (2023) to $0.506M (2025).

7.8 Verdict on capital allocation — competent, not intelligent, and the incentive guarantees it continues

The defence is real. The balance sheet has been genuinely repaired — leverage from 3.6x to 1.6x, duration to 2031, the term loan swapped fixed at 5.409%, an S&P upgrade, interest paid down 38% in a year. Stock compensation intensity is falling. Insider ownership is meaningful and entirely unpledged. LNK’s untouched 4.7% stake is a real vote of confidence. The May 2026 buyback at $28.60 was excellent. And the sponsors’ exit was fund-life mechanics, not a valuation verdict — the IPO buyer compounded at ~19.5% while LGP and TPG limped to ~2.9x over eleven years.

But the central fact is that this business does not fund its own growth. Operating cash flow has covered 78% of capex since the IPO; the shortfall was plugged by $908M of sale-leasebacks. In a record year with 11% comps, true free cash flow was −$21M, and the $206M reported as “free cash flow” was the proceeds of selling seven buildings. Management then measures leverage on a basis that excludes the $2.65bn of lease liabilities those sales created — 1.6x reported against 4.79x lease-inclusive — and used that improvement to justify deleting the leverage metric from the incentive plan.

What remains is a compensation structure that is 100% Adjusted EBITDA, an advertised unit return that fell from “over 40%” to “over 30%” while being re-labelled from return-on-capital to cash-on-cash and which still does not reconcile to the portfolio’s actual 19–20%, a gross cost per new club up 60% in three years, sale-leaseback pricing down from 1.31x book to 1.04x, and a company selling buildings at ~7.9% to avoid borrowing at 5.4%.

This is a good business being run for growth rather than for returns, financed by monetising the very asset that made it defensible, and directed by an incentive scheme that pays for exactly that. Nothing is hidden — every figure above comes from the filings — but very little is presented in the form an owner of capital would choose. The 2026 relative-TSR addition is the first sign the board has noticed. Until a return-on-capital or per-share metric enters the plan, Adjusted EBITDA growth at Life Time is a measure of how much capital was consumed, not of how well it was deployed.

8. Changes and Headwinds — Last Two Years

8.1 The material events timeline

8.2 What actually changed in the business

The balance sheet was fixed, and this was the dominant positive of the period. Net leverage went from 2.28x at end-2024 to 1.6x at end-2025. Interest expense fell from $148.1M to $82.3M. The term loan is fixed at 5.409% effective to 2031, the maturity wall was pushed out, and the credit rating reached BB a year early. Whatever else is arguable about Life Time, the financial-risk profile is materially better than it was two years ago, and management executed that deliberately and well.

The sponsor overhang was retired. Forty-three million shares cleared through 2025 at $29.50–$30.40, capping the stock for eight months, and the final 8.77M block moved to Atairos in May 2026 alongside the company’s own repurchase. The stock’s 72% advance since November 2025 is substantially the removal of that supply, not a change in fundamentals.

The growth plan was materially enlarged — and made more capital-intensive. The unit target went from 10–12 clubs a year to 12–14, with up to 28 across 2026–2027. Average new-club size went from ~66,000 sq ft (2025 class) to ~94,000–95,000 sq ft (2026 class), with 11 of 14 large-format ground-up builds and total 2026 square footage nearly double the 2024 and 2025 classes. Growth capex guidance rose to $875–915M. Q1 2026 capex alone was $260M, up 82% year-on-year.

The financing of that plan became more dependent on asset sales, not less. The 2026 sale-leaseback target was raised intra-year from “a minimum of $300 million” (February) to “approximately $400 million” (May), explicitly to support “our ongoing focus on generating annual positive free cash flow.” Against $875–915M of growth capex, ~$400M of sale-leasebacks is roughly 45% of the growth programme.

Capital returns started. The first buyback authorisation and the first execution both fell in this window, and the execution was excellent.

8.3 The headwinds

Deceleration is the principal one, and it is now management-guided. Comparable-centre revenue: 13.5% → 12.9% → 11.2% → 10.6% → 9.9% → 8.6%, guided to 6.3–7.3% for FY2026 against a 6–8% long-term target. Membership growth is guided to +0.5–1.0% in Q2, +1.0–1.5% in Q3 and +2.0–3.0% in Q4 (+3.5–5% excluding the deliberately-shrunk qualified-medical cohort).

Non-recurring income will not repeat. The $54.6M employee-retention credit, the $39.6M Zurich settlement and the $12.6M discrete option-exercise tax benefit are all in the FY2025 base and none recur. FY2026 GAAP net income is guided to $340–345M, below FY2025’s $373.7M, on revenue up 11% — the guidance itself concedes the point.

Cost inflation is persistent and asymmetric. Centre labour is running 2.5–3%; 22 states raised minimum wages in 2026. Construction-cost inflation directly raises gross invested capital per club, and Life Time’s own 10-K warns that higher gross invested capital “require[s] increases in the value of sale-leaseback transactions or higher operating profits per center to produce our targeted rate of return.”

The cap-rate exposure is real and outside management’s control. Single-tenant net-lease cap rates compressed to 6.80% in Q1 2026 but ticked wider in Q2 2026 (retail +5bp to 6.60%, industrial +10bp to 7.25%). Each 100bp of widening reduces proceeds per property by roughly 13–15% and raises net invested capital correspondingly. Management dismisses this — “if it was 25 basis higher or lower, it virtually has zero impact on the total economics of the business” — which is true at 25bp and not true at 150bp.

Premium-segment supply is arriving simultaneously from three directions. Equinox raised ~$1.8bn and is building 25–40 clubs; Bay Club is acquiring; LA Fitness’s Club Studio is going from 15 to 50. Life Time’s push into denser urban and vertical formats moves it toward catchments that support multiple premium operators.

A deferred exposure worth naming: the swaps expire in April 2028. The term loan is fixed only until then. The refinancing rate at that point is the largest unmodelled item in the capital structure.

8.4 Verdict — the thesis is stronger financially and weaker operationally than two years ago

The balance sheet, the credit rating, the maturity profile, the overhang and the start of capital returns are all genuine improvements, and they explain the re-rating honestly. But every operating trend that matters — comparable growth, membership units per club, the composition of comp growth, capital intensity, and dependence on asset sales — moved the wrong way over the same period. The market has repriced the balance sheet. It has not yet repriced the algorithm.

9. Risk Analysis

The two that matter. Risks 1 and 2 are the thesis. Everything else is either manageable, asymmetrically favourable to Life Time relative to the value tier, or remote. The genuine tail risk is the combination of 3 and 4 — a cap-rate backup coinciding with an affluent-consumer shock would simultaneously raise the cost of funding growth and reduce the cash flow available to fund it, in a business with $339M of contractual rent, $2.69bn of lease liabilities and a 52.6% observed decremental margin. That is the scenario in which the 2022 experience (−58%) repeats.

What is genuinely low-risk here deserves saying: this is not a balance-sheet story. Liquidity is $823M, there is no meaningful maturity before 2031, the debt is effectively fixed, the assets are real and saleable into a deep bid, and the equity is not impaired in any plausible scenario short of an extended discretionary-spending depression. The risk in Life Time is to the multiple, not to the solvency.

10. Valuation Discussion

No price target and no recommendation appear in this section. The purpose is to establish what the market is currently underwriting and to test whether that set of expectations is reasonable.

10.1 Where the stock trades — the live-price rebuild

Aggregator enterprise values for Life Time are badly stale, because they mark to the last fiscal period end (31 March 2026, when the stock was near $27) rather than to the live price. Rebuilt from the balance sheet at the 24 July 2026 close of $42.33:

The spread between 25.5x and 33.3x on the same fiscal year is the entire quality-of-earnings question in one line. The company’s own “adjusted” figure of $1.44 correctly removes the COVID credits and the Zurich settlement but adds back $51.8M of stock compensation, which is a real cost. $1.27 is the honest normalised number, and 33.3x is the honest trailing multiple.

10.2 Own-history context — the highest-signal datum

Discard the P/E percentile. Life Time’s public history contains GAAP losses in 2020–22 and near-zero EPS in 2023, so the trailing-P/E distribution is dominated by a depressed denominator; a 22.8th-percentile reading is an artifact, not cheapness. This is a known failure mode for own-history percentiles on post-loss cyclicals and it applies squarely here.

Read price-to-book and price-to-sales, which are undistorted. On both, Life Time has never been more expensive in its public life. One honest caveat: that public life is only 4.8 years and begins at the October 2021 IPO, so this is “richest since IPO,” not a decade percentile. It nonetheless spans a full boom-bust-boom cycle, which is more than most four-year windows contain.

10.3 Comparables — and why they must be handled carefully

There is no clean public premium-fitness comparable. Equinox, Bay Club, Invited, Crunch, EoS, Chuze and VASA are all private; Xponential is distressed; Planet Fitness is a franchisor with an entirely different capital model. The comp set must therefore be read as directional, not determinative.

Life Time trades at roughly 13.3x EV/EBITDA against Planet Fitness at 14.5x — a modest discount for a business with two-fifths the margin, three times the capital intensity, and less than half the return on capital. On EV/EBIT, which penalises Life Time’s heavier depreciation, it trades at a premium. The factor model reinforces the point from a different direction: Life Time’s statistically nearest neighbours are not fitness operators at all but real-estate services, commercial-mortgage REITs and healthcare services (LabCorp, TransUnion, FirstService, BXMT, Colliers), and its Real Estate factor loading exceeds its Consumer Discretionary loading in every model specification. The market is already pricing this partly as a real-estate business. It is not yet pricing it as one with a 5% return on capital.

10.4 Embedded expectations — what the price requires

Work backwards from $42.33. At a $9.42bn market capitalisation and $10.80bn enterprise value against FY2026E adjusted EBITDA of $932M, the market is underwriting some combination of the following. Each is stated with the evidence for and against.

(1) That comparable-centre growth stabilises at or above the 6–8% long-term target rather than continuing to decelerate. For: the legacy-to-rack-rate gap is a genuine perpetual annuity — roughly two-thirds of members still pay below rack, the aggregate pool has held at $17–20M a month for years, and it replenishes with each rack increase. In-centre attach through Dynamic Personal Training is growing 18% with capacity constraints, not demand constraints. Against: comps have fallen every single quarter for six quarters, from 13.5% to 8.6%; pure price is only 3.0 points and management guides it to a sustainable 2–3%; the mix lever that supplied 3.5 of the 8.6 points is nearly exhausted at 3.4% of dues revenue; and volume is already negative.

(2) That the 12–14-clubs-a-year programme earns well above the cost of capital. For: club-level four-wall EBITDA of $5.30–5.44M on net invested capital of $25–30M is 19–20%, which is a real return, and new clubs now reach contribution-margin positive within one to two months. Against: the advertised return was walked down from over 40% to over 30% and re-labelled to cash-on-cash; the disclosed arithmetic supports 19–20%, not 30%+; gross capex per opening has risen 60% in three years to $65.6M; and more than half of 2026 growth capex buys 2027-and-beyond openings.

(3) That free cash flow inflects positively without asset sales. For: management’s long-range plan projects more than $400M of free cash flow by roughly 2030, at which point Akradi says “I don’t want to sell any of my real estate.” The ramping cohort — 29 clubs under three years plus 17 under construction — represents real embedded earnings on capital already spent. Against: CFO-less-capex has been positive in one of five years; the FY2026 plan spends $1.15–1.21bn against $932M of adjusted EBITDA and raises the sale-leaseback target to ~$400M; and each sale-leaseback adds perpetual rent that makes the subsequent inflection harder.

(4) That the sale-leaseback bid stays open at current cap rates. For: the US market did 714 transactions and ~$14.4bn in 2025, and management calls it “robustly open.” Against: net-lease cap rates widened in Q2 2026 after compressing in Q1; Life Time’s own lease discount rate has risen 131bp since 2021; price-to-book on disposals has fallen from 1.31x to 1.04x. The 2023 drawdown is the live demonstration of what happens when this channel tightens.

A simple reverse test. To justify $10.80bn of enterprise value at a 10x exit multiple on unlevered free cash flow, Life Time needs roughly $1.08bn of sustainable unlevered free cash flow. Against FY2026E adjusted EBITDA of $932M less maintenance and modernisation capex of ~$280M less cash taxes, the current estate produces on the order of $450–500M. The price therefore requires the growth programme to roughly double sustainable free cash flow — which is exactly management’s 2030 plan, taken at face value and discounted only modestly for execution risk.

10.5 Scenario analysis

Applied to FY2028, three years out, with explicit assumptions.

*Equity value per share deducts funded debt net of cash, assumes modest deleveraging and a broadly flat share count, and excludes lease liabilities from the bridge (consistent with using a post-rent EBITDA). These are scenario outputs, not forecasts, and each rests on the stated assumptions rather than on a view about the share price.

The distribution is roughly symmetric in percentage terms around today’s $42.33 — about −22% to +63% — which is a worse skew than it looks, because the bear case does not require anything to go wrong. It requires only that the six-quarter deceleration continue at its observed rate.

10.6 What multiple is defensible

An 11–12x forward EV/EBITDA multiple is not obviously wrong for a business growing revenue 11% with expanding margins and a repaired balance sheet. The difficulty is that EBITDA is a poor proxy for economics here: it is struck before the depreciation on a $6.3bn property base, before the stock compensation the company adds back, and before the roughly $1.2bn of annual capital the business must consume to keep the algorithm running. On the measures that capture capital intensity — EV/EBIT of 21.2x, a P/E of 33.3x on normalised earnings, 2.99x book at the 99th percentile of its own history, and no free cash flow to speak of — the stock is priced as a compounder by a market that has, correctly, repriced the balance sheet and has not yet repriced the returns on capital.

Verdict on valuation. At $42.33 the market is underwriting the successful execution of a 450-to-500-club build-out at returns Life Time has not yet demonstrated at the enterprise level, funded by a channel whose terms are deteriorating, on an algorithm that has decelerated in each of the last six quarters. That is not an absurd bet — the operator is good and the local franchise is real. It is simply a bet with no margin of safety in the price.

11. Variant Perception

11.1 The consensus belief

Consensus holds that Life Time is a de-levered, self-funding premium compounder: a differentiated format with pricing power, expanding margins, leverage cut from 3.6x to 1.6x, an investment-grade-adjacent rating, positive and growing free cash flow, a newly-started buyback, and a long runway to 450–500 clubs. The stock’s 72% advance since November 2025 and its position 1.2% below an all-time high express that view. Short interest is 6.2% of float — moderate, though it has risen 19% month-on-month into the high, so a dissenting minority is building.

11.2 The strongest bull case

The local franchise is real, and the pricing runway is longer than the bears think. Life Time raises price 10%+ a year while visits per member rise — 122 million visits in 2025, up 18.4% in two years against 7.8% membership growth. Roughly two-thirds of members still pay below rack rate, and the CFO’s disclosure that the $17–20M monthly gap “never closes” means the price lever is a perpetual annuity rather than a one-time catch-up. Dynamic Personal Training is growing 18% with trainers “booked solid” — a capacity problem, which is the best kind. MIORA is a genuine option on the longevity and GLP-1 spending wave, and the clinical evidence says GLP-1s drive gym usage rather than replace it.

The capital cycle is working in Life Time’s favour, not against it. The mid-tier is being liquidated, boutique is in outright bust (Xponential guiding revenue down 16%, Orangetheory closing 26 US units), at-home digital is in structural decline, and Equinox needed a $1.8bn rescue raise. Life Time is the best-capitalised operator in the only segment gaining both revenue and price, self-funding through a channel its competitors cannot access.

And the inflection is closer than the numbers suggest. Twenty-nine clubs are under three years old and 17 more are under construction — capital already spent, earning nothing yet. More than half of 2026 growth capex buys 2027-and-beyond openings. When that cohort matures, free cash flow inflects mechanically, which is precisely management’s “>$400M by 2030” plan. Measured from the pre-COVID FY2019 base rather than the depressed 2022 trough, incremental ROIC is roughly 14–20% — comfortably above the cost of capital. The 5.2% consolidated figure is a portfolio average dragged down by LBO goodwill and idle construction, not a statement about the marginal club.

11.3 The strongest bear case

The company does not generate cash and calls its asset sales free cash flow. Positive CFO-less-capex in one of five years. Cumulative organic free cash flow of −$945M since the IPO against $983M of sale-leaseback proceeds. In its best year ever, on record revenue and margins, the business consumed roughly $115M of cash on a recurring basis. The reported $206.5M of “free cash flow” is the proceeds of selling seven buildings.

The returns do not clear the cost of capital and never have — 2.8%, 4.1%, 5.2%; roughly 7.65% even on the most generous lease-adjusted construction, against a WACC of 8.1–8.4%. The advertised unit return was walked down from “over 40%” to “over 30%” and quietly re-labelled from return-on-capital to cash-on-cash, and even that does not reconcile to the portfolio’s actual 19–20%.

The growth algorithm is decelerating and its composition is deteriorating. Six consecutive quarters of decline from 13.5% to 8.6%, guided to 6.3–7.3%. Pure price is 3.0 points — inflation. Volume is negative. The mix lever that supplied 3.5 points is nearly spent. And the longest view is the most damning: center memberships are still 3.7% below their 2019 level with 29.5% more clubs, and real revenue per square foot has been flat for six years.

The financing arbitrage runs backwards and is getting worse. Life Time sells buildings at an implied ~7.9% cap rate while it can borrow secured at 5.409%, forfeiting the freehold residual on 25-year leases. Price-to-book on disposals has fallen from 1.31x to 1.04x, the lease discount rate has risen 131bp to 9.28%, and gross capex per opening has risen 60% to $65.6M. Meanwhile management is paid 100% on Adjusted EBITDA, with the leverage metric deleted the moment sale-leasebacks achieved it.

And the one stress test available says the moat is thinner than claimed: in 2020 Life Time lost 41.3% of its memberships while Planet Fitness lost about 10%.

11.4 The assumptions that actually matter

1. Does comparable-centre growth stabilise at 6–8%, or continue through it? Falsifies the bull if FY2026 comes in below the 6.3% floor. Falsifies the bear if FY2027 comps hold at 7%+ after the qualified-medical drag laps.
2. Does organic free cash flow — excluding sale-leaseback proceeds and one-timers — turn positive and stay positive? Falsifies the bear if FY2027 CFO-less-capex is positive with sale-leasebacks below $200M. Falsifies the bull if FY2027 requires another target increase.
3. Does consolidated ROIC cross the cost of capital? Falsifies the bear at a sustained 9%+ lease-adjusted. Falsifies the bull if it stalls in the 6–7% band as the ramping cohort matures.
4. Does the sale-leaseback channel hold its terms? Falsifies the bull if implied cap rates move above ~8.5% or price-to-book falls below 1.0x. Falsifies the bear if 2026’s ~$400M clears at cap rates at or inside 2025’s.
5. Is retention genuinely at a record? Unfalsifiable today, because the company will not disclose it. Disclosure of a numerical attrition rate at or below ~25% would be the single most bullish datapoint available; continued refusal is itself informative.

11.5 The factor-positioning read

The tape and the factor model add a genuinely non-obvious piece of evidence.

Life Time’s factor loadings show no Momentum loading whatsoever — it is ElasticNet-zeroed in all four nested model specifications, as are Value, LowVolatility and Beta. A stock up 55% in a quarter with zero momentum loading is not a crowded momentum trade, which removes the most common mean-reversion tell. Roughly 85% of variance is idiosyncratic (specific volatility 35.2% against ~38% total, model R² 25–28%): this is an operational re-rating on the company’s own prints, not a factor ride.

The material loadings are SmallSize +0.41, Real Estate +0.415, and InterestRate −0.413 — and in every specification the Real Estate loading exceeds the Consumer Discretionary loading (+0.098). The market is already treating Life Time as a rate-sensitive, long-duration property business rather than a consumer operator, which independently corroborates the sale-leaseback framing in Sections 6 and 7 and quantifies the cap-rate exposure that drove the 2023 drawdown.

Where consensus may be offsides is on classification rather than direction. The factor-similar peer set contains no fitness operator at all — LabCorp, UGI, TransUnion, FirstService, BXMT, Colliers, Quest. If the market is pricing Life Time as a real-estate-services compounder, the relevant question becomes what return on capital such businesses earn, and 5.2% is not it.

Two further positioning facts. Life Time rallied into headwinds on two of its three real loadings — SmallSize is −5.7% over 126 days and the InterestRate factor is at a +1.91 z-score — while its nominal sector, Consumer Discretionary, sits at a −1.80 z-score. The single fitting tailwind is REITs, at an extreme +3.67 (126-day) and +3.82 (252-day) z-score. Extremes revert more often than they extend, though that is interpretation and regime-dependent, not a forecast.

Net: this is not a momentum crowd to fade and not a falling knife to catch. It is a high-beta, small-mid-cap, overwhelmingly idiosyncratic re-rating that the tape has fully validated — and precisely because the tape has agreed with the operational story, the remaining risk sits in the multiple rather than in the narrative.

12. Fact vs. Interpretation

13. Open Questions

1. What is the current attrition rate? The single most important undisclosed metric. The only figure ever published (16.1%, 1H2019) predates COVID and the entire premium repositioning. Without it the captivity claim cannot be verified and the mix strategy cannot be graded. Management claims record retention on four consecutive calls while declining to quantify it.
2. What definition reconciles “>30% cash-on-cash” to the portfolio’s disclosed 19–20%? Is it pre- or post-corporate overhead, pre- or post-maintenance capex, and does it charge the lease obligation the sale-leaseback proceeds create? The ten-point gap has never been addressed publicly.
3. What cap rate is Life Time actually achieving on sale-leasebacks? The company discloses proceeds but never the yield. Our ~7.9% estimate is derived, and it materially contradicts management’s stated 6.5–7.0%. This is the single most important undisclosed input to the entire growth model.
4. How many of the 189 clubs are the only premium full-service facility in their trade area? This is the missing datapoint for the local-scale thesis, and it is not disclosed anywhere.
5. Where does the qualified-medical run-off end, and what does underlying volume look like afterwards? When the drag laps, does volume prove to be the +4–5% management implies, or flat?
6. Is the $3.5bn owned-real-estate valuation supportable? Book value of the owned estate and the basis for the market-value claim are not in the filings.
7. What is the refinancing plan for the April 2028 swap expiry? The largest unmodelled exposure in the capital structure.
8. Is there a succession plan? Akradi has run this business for 34 years, designed the format, drives site selection and personally describes the pricing engine. Nothing is disclosed.
9. Will the board add a return-on-capital or per-share metric to the incentive plan? The 2026 relative-TSR addition suggests the question is live.
10. What did Q2 2026 (reporting 30 July 2026, four days after this report) show on comparable-centre growth and membership units? This report is written pre-print and the answer will immediately update assumptions 1 and 5 in Section 11.4.

14. What Must Be True

14.1 For the bull case

14.2 For the bear case

The single cleanest test either way is item 3 in both tables: FY2027 cash from operations less capital expenditure, with sale-leaseback proceeds disclosed separately. If that number is positive on a sub-$200M sale-leaseback programme, the bear case is substantially dead. If it is negative for a seventh year in eight while the target climbs again, the bull case rests entirely on a 2030 promise.

15. Source Appendix

See Appendix B below for the full source list. Principal sources: Life Time Group Holdings FY2021–FY2025 Forms 10-K and the Q1 2026 Form 10-Q (SEC EDGAR, CIK 0001869198); DEF 14A proxy statements 2022–2026; the October 2021 Form S-1/424B4; 51 Forms 8-K and 256 Section 16 filings covering July 2021 to July 2026; six earnings-call transcripts (Q4 2024 through Q1 2026) via the ROIC.ai data service; ROIC.ai computed fundamentals and ratios; AZI daily price history and own-history valuation percentiles; the FactorsToday factor model; Health & Fitness Association industry data; and named trade and financial press cited inline.

Sections 1–15 contain no investment recommendation and no price target; the Claude's Take block at the head of this article is a clearly-labelled exception representing the author’s own subjective view. This article is general information and not investment advice. The author may hold positions in securities mentioned.

APPENDIX A — Standard Diligence Questionnaire

Life Time Group Holdings, Inc. (NYSE: LTH) — 26 July 2026

A standard diligence questionnaire applied to this company. Fact / Interpretation / Assumption labels are applied where the distinction matters.

General

What thoughtful questions have other investors asked about this company?

Four recur, and all four are the right questions.

“Is the reported free cash flow real?” — the sharpest and most productive. Life Time’s non-GAAP free-cash-flow definition adds back proceeds from sale-leaseback transactions and land sales. FACT: the FY2025 reconciliation is CFO $870.5M + sale-leaseback proceeds $227.4M − capex $891.5M = $206.5M. Strip the asset sales and the figure is −$21.0M; strip the $94.2M of one-time COVID credits and Zurich proceeds inside CFO and it is roughly −$115M.

“What is the true leverage?” — management markets 1.6x net debt to adjusted EBITDA. FACT: that excludes $2,686.1M of lease liabilities carried at a 9.28% discount rate over 16.5 years. Rent-adjusted, leverage is ~3.44x on EBITDAR, or 4.79–4.86x lease-inclusive on adjusted EBITDA.

“Does the “>30% cash-on-cash” claim reconcile?” — INTERPRETATION: no. Portfolio club-level EBITDA of $5.30–5.44M on $25–30M of net invested capital is 18–22%, before any corporate overhead. The claim has also been walked down from “over 40%” in the FY2022 10-K and re-labelled from return on invested capital to cash on cash.

“Is the price-led growth model sustainable?” — the most important forward question. Management’s own Q1 2026 decomposition puts pure price at 3.0 of 8.6 points, mix at 3.5, in-centre at 2.3, and volume at −0.2.

A fifth question deserves to be asked more than it is: why has the company never disclosed an attrition rate since going public?

Cyclicality & Earnings Nature

Are earnings at a cyclical high or low? INTERPRETATION: a high, and a partly artificial one. FY2025 was the best year in company history on revenue, margin, EBITDA and net income — and roughly 24% of GAAP EPS was non-recurring (a $54.6M CARES Act credit, a $39.6M Zurich settlement, a $12.6M discrete option-exercise tax benefit). Adjusted EBITDA margin of 27.5% is a record against 23.0% pre-COVID. Management’s own FY2026 guidance concedes the point: GAAP net income of $340–345M is below FY2025’s $373.7M on revenue up 11%.

Driven by the external environment or internal actions? Predominantly internal. The margin expansion is a deliberate repricing and mix-optimisation programme — raising dues 10%+ annually, running off low-dues qualified-medical memberships, and de-densifying clubs. The external contribution is a genuinely favourable affluent-consumer backdrop (households above $250k drive 49.7% of US consumer spending) and a goods-to-experiences rotation.

How stable are revenues? Stable in normal conditions, fragile under access interruption. Roughly 73% of centre revenue is contractual monthly dues, billed by EFT. But 2020 is the available stress test: revenue fell 50.1%, centre memberships 41.3%, and adjusted EBITDA swung by $500.9M — a ~52.6% decremental margin — while rent rose 12.2%.

Outlook for products/services? Core club demand is healthy: 122M visits in 2025 (+18.4% over two years) and 12.5 monthly visits per membership (+4.8%). Dynamic Personal Training is the strongest line (+18% sessions for two consecutive years, capacity-constrained). MIORA (7–8 longevity clinics) is a genuine option with real regulatory complexity. Life Time Digital’s ~3.3M subscribers are free, non-paying accounts with monetisation “not in the near term” — value at zero.

How big will this market be? FACT: 81.0M Americans held gym memberships in 2025 (+5.2%, an all-time high; 26.1% penetration), ~7 billion annual visits, ~$47bn of US industry revenue, 55,000+ facilities. LTH holds ~1.1% of members but ~7% of revenue — the arithmetic of the premium niche. Growing, not shrinking. Almost entirely domestic: 31 US states plus one Canadian province, with no stated international ambition.

Business Quality & Competitive Moat

Is the industry getting more or less competitive? More, in three of four tiers, and the pattern is tier-specific. HVLP is late-boom (330–350 new boxes in 2026; Planet Fitness cut same-club-sales guidance from 4–5% to ~1% and fell 31%). Boutique is in outright bust (Xponential guiding revenue 16% lower, a $17M FTC settlement; Orangetheory closed 26 US units). Mid-tier is exiting — good for survivors. Premium is mid-boom and capital is arriving: Equinox raised ~$1.8bn for a 25–40-club pipeline, Bay Club is acquiring, Club Studio is going 15→50. But premium adds only ~40–60 large-format boxes a year against 330–350 HVLP and 500+ boutique studios, because entitled 100,000 sq ft sites and $60–90M of capital are constraints franchise debt cannot relax.

How profitable is the business (ROIC, ROE)? FACT: ROIC 2.84% (FY23), 4.06% (FY24), 5.22% (FY25); ~6.6% ex-goodwill and ex-construction-in-progress; ~7.65% on the most generous lease-adjusted construction. ROE 13.0% reported, 10.0% normalised. INTERPRETATION: against a WACC of 8.1–8.4%, Life Time has not earned its cost of capital in its public life. Planet Fitness earns 11.66%.

How profitable is the industry — how many competitors, what barriers to entry? Median operator EBITDA margin ~23.6%; LTH’s 27.5% is only ~4pp above it, and is struck before the capital intensity that produces the 5.2% ROIC. Nationally there are no barriers — 55,000+ facilities, no operator above ~10% share, and a 2020 graveyard (24 Hour Fitness, Gold’s Gym, Town Sports, YouFit). Locally, barriers are real: a ~97,000 sq ft box costs $60–90M gross, takes 2–4 years to entitle and build and 3–4 more to ramp, and a trade area supports one, rarely two.

Can it be easily understood? Yes — this is among the most legible business models in any sector. Build a big club in a wealthy suburb, charge $230 a month, sell training and spa services inside. The complexity is entirely in the financing.

Can it be undermined by foreign low-cost labour? No. The service is physically delivered on a specific site. Equipment is sourced from Italy and Sweden, so tariff exposure is de minimis. The relevant labour risk is domestic wage inflation (2.5–3% centre labour; 22 states raised minimum wages in 2026) — and it is asymmetrically harder on the $15–30/month tier, where labour is a far larger share of a much smaller revenue per member.

Do brands matter? Partly. The Life Time brand supports premium positioning and mix migration. But run Greenwald’s test — does the brand permit price increases without volume loss? — and it is a partial pass at best: pure price is 3.0 of 8.6 comp points (roughly inflation), volume is −0.2, and centre memberships remain 3.7% below 2019 despite 29.5% more clubs.

What is the nature of competition? Local and format-based, not national. Management named neither Equinox, Planet Fitness, Crunch nor Orangetheory on any of six earnings calls. Akradi locates the advantage in gestation and complexity: “the incredible moat that is around this company … because it takes such a long time to develop these things.” He claims boutique flow is one-way: “Nobody is leaving a Life Time to go to studios. But on the reverse, we do see the reverse.” Corroborated by in-centre revenue +15.1% and DPT +18%.

Customers’ switching costs? Situational — high for families, near zero for individuals. Real friction: swim progressions, summer camps, childcare, racquet leagues, ~60% couple/family memberships, and an on-hold tier at $15/month that preserves re-entry without a new enrolment fee (50,556 memberships parked there). Real counter-evidence: month-to-month contracts, statutory 3–10 day cancellation windows, no data or equipment lock-in, an implied ~30%+ gross annual attrition rate from the only figure ever disclosed — and 2020, in which LTH lost 41.3% of memberships while Planet Fitness lost ~10%.

Financial Condition & Balance Sheet

Assets not fully recognised on the balance sheet? Yes, materially. Management asserts ~$3.5bn of owned real-estate market value against $1.5bn of funded debt; the 55 owned centres are carried at depreciated historical cost. UNVERIFIED ASSERTION — repeated on calls but not reconcilable from the filings. Also unrecognised: the brand, the 34-year development capability, the site pipeline (85–100 deals), and the ~1.6M-member relationship base.

Off-balance-sheet liabilities? Post-ASC 842 the leases are on balance sheet, but two items sit outside the recognised liability: $545.8M of signed-but-not-commenced leases, and the gap between the $2,686.1M discounted liability and $5,249.7M of undiscounted future payments — $5,795.5M in total contracted rent including signed-not-commenced. Also off-balance-sheet in economic substance: the master-lease structure putting 55 properties under 15 master leases with only 14 lessors, which removes the ability to hand back a single underperforming club.

How conservative is the accounting? The GAAP accounting is clean; the non-GAAP presentation is aggressive. No restatements, no revenue-recognition games, no capitalisation abuse; sale-leaseback gains are properly excluded from adjusted EBITDA and the resulting rent is properly deducted (each transaction mechanically lowers adjusted EBITDA). Two aggressive choices: (a) the “free cash flow” definition adds back asset-sale proceeds; (b) the marketed leverage ratio excludes operating leases. One presentation blemish: the FY2025 MD&A cites a “$2.6M net sale-leaseback gain” for FY2024 while Note 10 shows $19.8M — the bridge is a $17.2M related-party loss in Note 14, and the MD&A wording obscures it.

How CapEx-hungry is the business? Extremely, and increasingly so. FY2025 capex $891.5M on $2,995.3M of revenue (29.8%) at 3.01x depreciation. FY2026 guidance is $1,145–1,205M against adjusted EBITDA of $925–940M — capex exceeds EBITDA by ~$240M. Even excluding growth, maintenance plus modernisation runs ~$235–290M a year. Gross growth capex per opening has risen from $40.9M (2022) to $65.6M (2025).

Capital Allocation & Management

How much FCF does the business generate, and what is the philosophy? On the honest definition, none. Positive CFO-less-capex in one of the last five years (FY2024, +$50.6M, at the trough of the growth-capex cycle). Cumulative FY2021–25 organic FCF −$944.5M against $982.5M of sale-leaseback proceeds. The philosophy is explicit and management is candid about it: sale-leaseback proceeds support “our ongoing focus on generating annual positive free cash flow,” and the long-range plan targets >$400M of free cash flow by roughly 2030, at which point Akradi says “I don’t want to sell any of my real estate.”

Significant acquisitions recently? None. Growth is entirely organic — a genuine positive. The only material disposal was the triathlon events business (a $4.9M gain in FY2023). The recurring “transactions” are sale-leasebacks: 27 properties for ~$962M since 2021.

Buying back shares? Yes, newly and well. A $500M authorisation was approved 19 February 2026 — the first ever. Executed to date: ~$73M (15%), comprising ~0.4M shares at ~$26.75 in Q1 2026 and 2,192,500 shares at $28.60 for $62.7M in May 2026, bought privately from Leonard Green, TPG and Partners Group. Stock is $42.33 ten weeks later. Caveats: a related-party purchase from affiliates with board designees, and only 15% deployed before a 48% run.

Issuing large amounts of new shares to insiders? No. SBC was $51.8M in FY2025 — 1.73% of revenue and 0.86% of market capitalisation, falling as a share of revenue from 2.26%. Share count is up 15.4% since end-2021 (~3.3% a year), of which the only primary raise was 6.0M shares at $21.75 in August 2024. (FY2021’s $334.3M was one-time IPO vesting.)

Compensation policy of directors/management? This is the weakest link in the file. The 2025 short-term plan was 100% Adjusted EBITDA (tranches $740M/$760M/$780M; actual $825.2M; every NEO at maximum). Long-term PSUs also vest on Adjusted EBITDA (2025 tranche 200%, 2024-granted tranche 250%). The 2026 short-term plan is “again based entirely on our Adjusted EBITDA.” There is no ROIC, no free-cash-flow and no per-share metric anywhere. The leverage metric used in 2024 was deleted for 2025 because it “improved dramatically” — and as defined (excluding operating leases) it could be improved simply by selling buildings and leasing them back. CEO total compensation was $15,199,475 in FY2025 against a median employee of $9,816 (a 1,548:1 ratio); perquisites include aircraft, a car with tax gross-up, and $42,000 of personal administrative support with a $35,205 gross-up. The one improvement: 2026 PSUs are split evenly between Adjusted EBITDA and a new relative TSR vs the Russell 2000 metric — the first external measure ever used.

Motivations of management? Founder-driven, growth-oriented, and honestly disclosed even when unflattering. Akradi has built this business since 1992, holds ~16.3M shares (7.3%, ~$690M) with zero pledging company-wide, and bought 273,411 shares of his own money at an average $14.26 during the 2022–24 drawdown. He is also candid against interest — conceding Life Time Living earns “extremely below the IRR of our club operations” and that the supplements digital channel is “mediocre. It’s so-so.” Against that: his only sale (5.0M shares at $30.13, February 2025, $150.65M) was discretionary rather than 10b5-1, though mechanically driven by a 9,388,000-share option struck at $10.00 expiring that October; no insider has bought a share since 31 May 2024; and the company pays ~$11.5M a year of rent to entities affiliated with the CEO and/or a director, including a 2024 sale-leaseback to a CEO/director-affiliated entity on which LTH booked a $17.2M loss at an 8.75% implied rent yield. The CEO’s daughter is VP of Real Estate and Location Intelligence ($0.506M in 2025).

Valuation & Market Data

Is the stock an ADR, MLP, or K-1 issuer? No. Life Time Group Holdings, Inc. is a Delaware corporation filing 10-K/10-Q, listed on the NYSE, issuing a Form 1099 — no K-1, no ADR, no partnership units, single share class.

Dividend policy? None, ever. Distributions are restricted by the credit agreement and the 6.000% notes indenture; substantially all subsidiary net assets are restricted. Capital return began in 2026 through buybacks only.

How profitable is the business? Layered: excellent at the club level ($5.30M of four-wall EBITDA per club, a 34.4% four-wall margin), good at the consolidated P&L level (27.5% adjusted EBITDA margin, 16.1% operating margin), and inadequate at the capital level (5.22% ROIC vs 8.1–8.4% WACC). The gap is corporate overhead (~$260M/yr), 2015-LBO goodwill ($1,235M), 46 immature or under-construction clubs, and ~$235M/yr of maintenance capex the unit metric excludes.

Is net income diverging from cash from operations? Yes, and in the flattering direction — but the flattery is in both lines. FY2025 CFO of $870.5M was 2.33x net income of $373.7M, consistent with $296.3M of D&A, $51.8M of SBC and an $87.5M deferred-tax benefit. That is a normal, benign gap for a capital-intensive operator. The problem is not the CFO-to-NI relationship but the CFO-to-capex relationship: $870.5M of CFO against $891.5M of capex. And both lines contain the same $94.2M of one-time credits.

Where does the stock trade? At $42.33 (24 July 2026): market cap $9,416M, EV $10,799M (funded debt) or $13,451M (lease-inclusive). 11.6x FY2026E adjusted EBITDA; 28.2x FY2026E GAAP EPS; 33.3x FY2025 normalised EPS of $1.27; 2.99x book — the 99.3rd percentile of its own history — and 3.11x sales at the 95.9th. Short interest 6.2% of float, up 19% month-on-month.

Risks & Downside

What factors would cause the stock to decline? In descending order of likelihood: (1) comparable-centre growth breaking below the 6.3% guidance floor as the mix lever exhausts — six consecutive quarters of deceleration make this the base-rate expectation, not a tail; (2) a sale-leaseback cap-rate backup raising net invested capital per club — the 2023 drawdown (−43%) is the live demonstration; (3) an affluent-household wealth shock, given a ~52.6% decremental margin and a member base with $160,000 median household income; (4) multiple compression from a 99th-percentile price-to-book absent any operating disappointment at all; (5) construction-cost inflation outrunning what net-lease buyers will pay; (6) premium-segment overbuild as Equinox, Bay Club and Club Studio expand into the same affluent catchments.

Risk of a catastrophic loss? Low. Funded debt of $1,525.4M against a claimed $3.5bn of owned real estate; net leverage 1.6x (3.44x rent-adjusted); liquidity $823.0M; a $650M undrawn revolver with no maintenance covenant below 30% drawn; effectively 100% fixed-rate debt with no meaningful maturity before 2031 ($1,445.1M thereafter). The genuine tail scenario is a cap-rate backup coinciding with an affluent-consumer shock, which would simultaneously raise the cost of funding growth and cut the cash flow to fund it against $339.2M of contractual rent. Even then this is a multiple event, not a solvency event — as 2022 showed, when the stock fell 58% and the business survived intact. Deferred item worth naming: the interest-rate swaps expire in April 2028.

Chance of a total loss? Very low. 189 hard, saleable assets in a deep net-lease bid; a member base of ~1.6M people generating $2.1bn of annual contractual dues; positive and growing operating cash flow; and a founder holding ~$690M of unpledged stock. The company survived a 50.1% revenue collapse in 2020 without restructuring. The risk in Life Time is to the multiple and to the return on incremental capital — not to the existence of the equity.

Recent News & Events

Has the business environment changed recently? Yes, in four ways, and they do not point the same direction. (1) Favourably: the mid-tier continues to exit, at-home digital is in structural decline (Peloton subscriptions −8%), and the GLP-1 debate has resolved in the industry’s favour — a William Blair survey of 300 users found gym membership rising 3pp to 35% post-initiation with 72% working out more. (2) Unfavourably: the first crack appeared in the value tier when Planet Fitness cut same-club-sales guidance from 4–5% to ~1% and fell 31%. (3) Structurally: single-tenant net-lease cap rates compressed to 6.80% in Q1 2026 but ticked wider in Q2 2026 — directly relevant to a company funding ~45% of its growth programme through sale-leasebacks. (4) Regulatorily: the Eighth Circuit vacated the FTC’s “click-to-cancel” rule in July 2025 and the FTC issued a new ANPRM on 11 March 2026; ROSCA enforcement never stopped, and the FTC has a live complaint against LA Fitness. This is asymmetrically an HVLP issue — LTH members visit ~39 times a quarter and need no cancellation friction to be retained.

Significant acquisitions? None by Life Time. Industry-wide: Leonard Green acquired a majority of Crunch from TPG Growth (April 2025); Bay Club acquired 425 Fitness (June 2025); Equinox raised ~$1.8bn (March 2024).

Change in accounting policies? None. No restatements or policy changes across the five-year corpus. The only accounting item worth knowing is longstanding: related-party sale-leasebacks are recorded “at their contractually stated terms” and are not adjusted for off-market terms — which is why the 2024 related-party transaction produced a $17.2M loss.

Recent changes — new markets, facilities, management? Facilities: 190 centres at 31 March 2026, accelerating to 12–14 openings a year with a 2026 class of ~1.2M sq ft — nearly double 2024 and 2025 — averaging ~94,000–95,000 sq ft (vs ~66,000 in 2025), 11 of 14 ground-up, and planned at 3,500–4,000 memberships versus 4,400–4,600 today. Markets: a deliberate shift toward denser urban and vertical formats; MIORA scaled from 2 to 7–8 clinics. Capital structure: leverage 2.28x → 1.6x; a BB rating in June 2025, a year early; swaps fixing the term loan at 5.409% to April 2028; a first-ever $500M buyback authorised February 2026. Governance: the sponsor overhang cleared in May 2026; directors Santo Domingo (March 2026) and Small (May 2026) resigned, Wagner appointed; the board is now 12 members, 10 independent, having exited “controlled company” status in June 2025.

Timing note. Life Time reports Q2 2026 on 30 July 2026, four days after this report date. The comparable-centre growth and membership-unit lines in that print bear directly on the central question in this file.

APPENDIX B — Source Appendix

Life Time Group Holdings, Inc. (NYSE: LTH) — 26 July 2026

All sources accessed 26 July 2026 unless otherwise noted. Primary sources are listed first. Price reference throughout: $42.33, the close of 24 July 2026.

1. Primary — SEC filings (CIK 0001869198)

The trailing five-year corpus (1 July 2021 – 26 July 2026) was enumerated and mirrored locally: 472 filings — 220 Forms 4, 69 Forms 144, 51 Forms 8-K, 31 Forms 3, 38 Schedules 13D/G, 14 Forms 10-Q, 5 Forms 10-K, 5 DEF 14A, 4 Forms 424B7, 2 Forms 424B5.

Local mirror: output/LTH/sources/ (by form type, with MANIFEST.csv and filing_index_LTH.txt).

2. Primary — company communications

Transcripts sourced via the ROIC.ai data service; local copies in output/LTH/transcripts/.

3. Quantitative data services

4. Peer and comparable-company sources

• Planet Fitness, Inc. (NYSE: PLNT) — SEC filings and ROIC.ai fundamentals; FY2025 revenue, unit count, margins, ROIC (11.66%); Q1 2026 same-club-sales guidance cut from 4–5% to ~1% and the 31% share-price decline.
• Xponential Fitness (NYSE: XPOF) — 2025 unit closures (140 against 341 gross openings), 2026 revenue guidance ~16% below 2025, the $17M FTC settlement and ~$22.75M franchisee class-action settlement.
• Equinox Group (private) — the March 2024 ~$1.8bn capital raise led by Sixth Street and Silver Lake with Ares, HPS, L Catterton and the Related principals; club count and pipeline, per company announcements and trade press.
• Peloton Interactive (NASDAQ: PTON) — subscription decline of 8% year-on-year to 2.66M; fifth consecutive year of revenue decline.
• Orangetheory / Bay Club / Crunch / EoS / Chuze / VASA / solidcore / Club Studio — unit plans, closures and ownership changes per trade press cited below.

5. Industry data

6. Regulatory sources

• US Court of Appeals for the Eighth Circuit — vacatur of the FTC Negative Option (“click-to-cancel”) Rule, July 2025.
• Federal Trade Commission — Advance Notice of Proposed Rulemaking on the Negative Option Rule, 11 March 2026 (comments closed 13 April 2026); the enforcement complaint against Fitness International, LLC (LA Fitness); the March 2026 Xponential Fitness settlement.
• Gibson Dunn and Goodwin Procter client alerts on the vacatur, continuing ROSCA enforcement, and the ANPRM.
• State health-club statutes as described in LTH’s own 10-K Item 1, with specific reference to the Illinois Physical Fitness Services Act, Washington trust/bond requirements, Virginia escrow rules and the New Jersey three-year contract cap.
• Minnesota Court of Appeals (reversal, 11 August 2025) and Minnesota Supreme Court (review denied, 29 October 2025) in the Zurich COVID business-interruption litigation.

7. Trade and financial press

Athletech News (GLP-1 survey coverage; HVLP and mid-tier analysis; Club Studio expansion); SGI Europe and StockTitan (Planet Fitness Q1 2026 guidance cut); PR Newswire and news.lifetime.life (Life Time earnings releases); Orange County Business Journal, 20 July 2026 (Brea Mall club build cost); MMCG (US reformer-Pilates market outlook; post-pandemic industry analysis); CNBC and Motley Fool (contemporaneous coverage of the October 2023 print); Houlihan Lokey Fitness Index.

8. Analytical frameworks

• Bruce Greenwald & Judd Kahn, Competition Demystified — the three genuine competitive-advantage types; the local-market instruction; the market-share-stability and ROIC tests (15–25% indicates advantages present, 6–8% indicates absence); “market growth is the enemy of economies-of-scale advantages.”
• Edward Chancellor (ed.), Capital Returns (Marathon Asset Management) — supply-side capital-cycle analysis; the asset-growth anomaly; the late-cycle warning-sign checklist; “competition neglect.”

9. Data limitations and flagged gaps

1. No current attrition or retention rate exists in any public source. The only figure ever disclosed is 16.1% for the first half of 2019, in the IPO S-1. Every retention claim in this report is labelled as management commentary.
2. The sale-leaseback cap rate is not disclosed. The ~7.9% figure used in Sections 7 and 10 is derived from the operating-lease right-of-use addition, and is labelled an assumption throughout. It materially contradicts management’s stated 6.5–7.0%.
3. Management’s ~$3.5bn owned-real-estate valuation is unverified and not reconcilable from the filings.
4. No public premium-fitness comparable exists. Equinox, Bay Club, Invited, Crunch, EoS, Chuze and VASA are private; Planet Fitness is a franchisor with a different capital model. The comp set is directional only.
5. The own-history valuation percentiles span only 4.8 years (from the October 2021 IPO), and the P/E percentile is unusable because of the 2020–23 loss and near-zero-EPS years.
6. FY2019 and FY2020 line-item detail comes from ROIC.ai and pre-dates the mirrored SEC corpus; it is un-reconciled to a filing.
7. ROIC.ai list_earnings_calls ignores the identifier argument and returns a cross-ticker list; transcripts were retrieved by explicit year and quarter instead.
8. Aggregator enterprise values for LTH are stale, marking to the 31 March 2026 period end (~$27) rather than the live price. All enterprise values in this report were rebuilt from the balance sheet at $42.33.
9. Q2 2026 reports on 30 July 2026, four days after this report date. This report is written pre-print.