Las Vegas Sands Q3 2026 Earnings: The $3.2B Quarter the House Didn't Win

Las Vegas Sands earned $304 million on $3.22 billion of revenue in the third quarter — and the entire shortfall fits inside one line item at one property. Marina Bay Sands' mass-table win percentage collapsed from 24.5% to 18.7% even as gamblers put 12.6% more money on the tables, wiping roughly $150 million of high-margin casino win off the quarter. Volumes grew, hotel rooms ran 95.8% full at a record $1,038 nightly rate, and management still bought back $950 million of stock and raised the dividend 20%. This is the first casino ledger in the Open Ledger library, and it opens with the lesson every casino quarter teaches: hold rate moves the top line more than occupancy ever will.

The Headline Numbers​

Las Vegas Sands' fiscal year is the calendar year, so every figure below comes from the Form 10-Q for the quarter ended September 30, 2026 and the earnings release attached as Exhibit 99.1 to the Form 8-K — every income-statement line agrees between the two.

Net income here and everywhere below is attributable to Las Vegas Sands — the $352 million of consolidated Q3 net income less the $48 million belonging to the minority shareholders of Sands China, deducted the way the filing deducts it. Read the table top to bottom and the damage concentrates downward: revenue off 3.5%, operating income off 14%, net income off 27%. That steepening is the tax line — the effective rate jumped from 15.6% to 22.5%, which alone accounts for roughly a third of the pre-tax-to-net gap widening year over year.

The GAAP-to-non-GAAP bridge is unusually clean this quarter — no debt-retirement losses, no impairment drama:

Development expense — the pursuit of new markets and the New York downstate license effort — is the only recurring add-back of size, and even it halved year over year from $72 million. There is no adjusted-EPS magic here: the miss is in reported casino win, full stop.

Revenue Deep Dive​

The portfolio split tells the quarter. Macao grew; Singapore shrank; Singapore's decline was bigger.

Macao is a volume story, and management said so directly. "Our ongoing investments in enhanced service and hospitality offerings [are] contributing to growth in volumes across all gaming segments as compared to the prior year," CEO Patrick Dumont said in the release. The exhibits back him: Londoner rolling-chip volume surged 39% to $3.22 billion, Venetian rolling volume rose 44% to $917 million, and non-rolling drop grew at the Londoner ($2.80 billion, up 24%) and company-wide. The Plaza and Sands Macao — the premium-mass and grind properties — both posted double-digit revenue growth. This is genuine demand strength, and it is the quarter's most bullish fact.

Marina Bay Sands is the mirror image: revenue down $142 million, of which casino accounts for $161 million — every non-gaming line grew. Rooms added $8 million, food and beverage $9 million, the mall and convention lines held. A property whose hotel runs 95.8% occupied at $1,038 a night is not suffering a demand problem. It suffered a cards problem: rolling-chip volume fell 7% to $8.43 billion while rolling win slipped to 4.53% (against a 4.6% theoretical), and mass win collapsed 5.8 points to 18.7% on 12.6% higher drop. Dumont's other demand line — "positioning our portfolio of properties to capitalize on the growing demand for premium travel and world-class entertainment experiences in the region" — reads as forward positioning, not a description of the quarter just printed. On the signal scan, robust demand is the only one of the seven themes with a verbatim hit; pricing power, supply tightness, upcycle language and demand-outpacing-supply are all absent. Notably, management never once uttered the words "win percentage" or "hold" in the prepared remarks — the release lets the exhibits confess what the commentary omits.

The Margin Story​

Every margin compressed, in both geographies, and the mechanics differ by segment. At Marina Bay Sands the story is pure operating leverage in reverse: casino revenue fell $161 million while the cost of running the building barely moved, so $153 million of the property's $153 million EBITDA decline is gross-profit air. A 45.6% EBITDA margin would still be the envy of nearly any business on earth — it just is not 51.7%.

Macao's compression is more interesting because revenue went the right way. Macao revenue rose $34 million while Macao EBITDA fell $72 million — costs outran sales. Consolidated resort operations expense rose 5.3% to $2.10 billion against a 3.5% revenue decline, and the property tables show the pain distributed: Venetian margin -3.4 points, Londoner -5.9 points, Parisian -4.4 points. Only the two smallest properties held or expanded margin. When volumes grow "across all gaming segments" and EBITDA falls 12%, the business is buying its top line with comps, commissions, labor and reinvestment — the classic mass-market treadmill, running faster to stand still.

Then the tax line takes its cut. The 22.5% effective rate, "primarily driven by a 17% statutory rate on our Singapore operations," is the structural price of MBS being the profit engine: Singapore taxes what Macao's gaming concession largely does not. As long as MBS out-earns Macao, the blended rate stays in the twenties, and every point of it falls straight through to the net margin. Pricing power is the unclaimed hero of the quarter — MBS average daily rate up 5.7% to $1,038, RevPAR up 6.1% — real pricing momentum that management's commentary never mentions, perhaps because claiming it would invite the question of why none of it reached EBITDA.

The One Big Question: Was Marina Bay Sands Weak, or Just Unlucky?​

Strip out luck and re-ask the quarter. The release's hold-normalization exhibit does exactly this for rolling-chip play:

Last year's quarter was the lucky one: MBS held 4.84% against a 4.20% theoretical, flattering casino revenue by $57 million. This year's 4.53% against 4.60% is essentially fair — the +$6 million adjustment is noise. So roughly $60 million of the $161 million casino decline, more than a third, is just lapping good cards. Any analysis that annualizes Q3 2025's hold as normal is modeling luck as earnings power.

But hold normalization cuts only one way here, and the remaining $100 million of decline is real softness. Rolling volume fell 7% — fewer VIP chips in play, the one unambiguous demand signal in the MBS data and the number worth watching next quarter. And the mass hold collapse (18.7% vs. 24.5% on higher drop) sits outside the rolling-only hold adjustment entirely: mass win is structurally supposed to be the stable, high-volume annuity, and a 5.8-point swing on $2.9 billion of drop is worth roughly $90 million of theoretical win in a single quarter. Mass hold mean-reverts — the cards do not remember last quarter — but anyone modeling MBS on last year's 24.5% was modeling a hot streak, not a house edge.

The honest decomposition: about a third luck rolling off, about a third VIP volume softness, about a third mass-hold noise. None of the three is structural. All three can recur.

Tracking a $19B Company in Plain Text​

A casino company is a near-perfect plain-text accounting subject: two geographies, six named properties, and a balance sheet where three decades of capital allocation are fossilized in treasury stock and land concessions. Modeling Las Vegas Sands in Beancount forces every one of those fossils to reconcile — this is the company. Our conventions for how we model every company deduct the Sands China minority's earnings share exactly as the income statement does, so the ledger's bottom line is the $304 million attributable to LVS shareholders, never the $352 million consolidated total.

Here is the quarter, as the ledger records it (income negative, expenses positive — Beancount sign conventions):

; FY2026Q3 Income Statement — three months ended September 30, 2026
; Check: −3,216 + 1,408 + 94 + 151 + 25 + 71 + 21 + 333 + 88 + 5 + 37 + 340 + 21 + 4 − 23 + 186 + 1 + 102 + 48 + 304 = 0 ✓
 
2026-09-30 * "Las Vegas Sands Corp." "FY2026Q3 Income Statement"
  Income:Revenue                         -3216 MUSD  ; net revenues
  Expenses:CostOfRevenue                  1408 MUSD  ; casino
  Expenses:CostOfRevenue                    94 MUSD  ; rooms
  Expenses:CostOfRevenue                   151 MUSD  ; food and beverage
  Expenses:CostOfRevenue                    25 MUSD  ; mall
  Expenses:CostOfRevenue                    71 MUSD  ; convention, retail and other
  Expenses:OtherNet                         21 MUSD  ; provision for credit losses
  Expenses:SellingGeneralAdministrative     333 MUSD  ; general and administrative
  Expenses:SellingGeneralAdministrative      88 MUSD  ; corporate
  Expenses:SellingGeneralAdministrative       5 MUSD  ; pre-opening
  Expenses:SellingGeneralAdministrative      37 MUSD  ; development
  Expenses:DepreciationAndAmortization     340 MUSD  ; depreciation and amortization
  Expenses:DepreciationAndAmortization      21 MUSD  ; amortization of leasehold interests in land (total expenses 2,598; operating income 618)
  Expenses:OtherNet                          4 MUSD  ; loss on disposal or impairment of assets
  Income:OtherNet                          -23 MUSD  ; interest income
  Expenses:Interest                        186 MUSD  ; interest expense, net of amounts capitalized
  Expenses:OtherNet                          1 MUSD  ; other expense (income before income taxes 454)
  Expenses:IncomeTax                       102 MUSD  ; income tax expense (effective rate 22.5%)
  Expenses:OtherNet                         48 MUSD  ; net income attributable to noncontrolling interests, deducted as the filing does
  Equity:Adjustments                       304 MUSD  ; net income attributable to Las Vegas Sands Corp. offset (RE set by balance assertion)

The one balance-sheet number that tells the narrative is treasury stock: $11,530 million — against $6,115 million of paid-in capital. Three years of buybacks ($6.98 billion since the program resumed in late 2023, 19% of shares outstanding, 21 million more shares at $45.31 this quarter) have now pushed parent stockholders' equity outright negative, to a $344 million deficit. The company funded it honestly — the $1.26 billion Venetian seller loan came home in full this year, and $4.27 billion of revolver plus a $4.57 billion expansion facility stand behind the $15.3 billion of debt — but the direction is unmistakable: this balance sheet is being levered, deliberately, to retire equity. The ledger's other curiosity is the $2,995 million of leasehold interests in land — prepaid Macao concessions and Singapore land premiums, including the $848 million MBS expansion parcel bought in April 2025 — an asset no domestic casino operator carries and the reason this ledger needed its own account.

The Multi-Year Arc​

Three years of compounding, one quarter of noise. Revenue grew 25% from FY2023 to FY2025, EBITDA 28%, attributable net income 33% — and the share count shrank 9% underneath it all, so per-share compounding runs hotter than every one of those figures. The casino share creeping from 72.5% to 75.2% is the mix story to watch: the higher-margin gaming line is doing more of the work each year, which flatters margins in good-hold years and concentrates the exact risk Q3 just demonstrated. Nine-month 2026 revenue is up 6.2% on 9M 2025 even after the soft quarter, and nine-month operating income of $2.14 billion is already ahead of last year's pace — the year is fine; the quarter was cards.

The Verdict: Bull vs. Bear​

Bull Case

• Volumes are growing "across all gaming segments" in Macao while MBS mass drop rose 12.6% — demand is healthy and hold mean-reverts, so the $150 million of missing win is timing, not erosion.
• MBS pricing power is intact and unclaimed: 95.8% occupancy at a $1,038 average rate (+5.7%), with every non-gaming line growing — the building is full at record prices.
• Capital returns are aggressive and funded: $950 million of buybacks this quarter, a 20% dividend hike to $1.44 annualized for 2027, $5.06 billion of authorization remaining, and the $1.26 billion seller loan repaid in full.
• The Londoner is ramping into a premium-mass engine — rolling volume up 39% year over year — while the MBS expansion land ($848 million paid, arena to come) locks in the next decade of Singapore growth.
• Nine-month operating income of $2.14 billion already exceeds last year's nine-month pace, and the buyback has cut diluted shares 6% year over year — per-share earnings compound even when the quarter does not.

Bear Case

• Costs are outrunning sales in Macao: revenue +1.8%, EBITDA -12.0%, with margins compressing at every major property — the mass-market treadmill is speeding up, and "robust demand" has not reached the bottom line yet.
• MBS mass win at 18.7% may not be pure noise — if premium-mass mix is shifting toward lower-hold play, the 51.7%-margin era is over and models anchored to it are wrong.
• The 22.5% effective tax rate is structural, not transient — Singapore's 17% statutory rate taxes the profit engine Macao's concession does not, permanently capping net-margin expansion.
• Parent equity is now negative $344 million against $15.3 billion of debt with $1.57 billion of maturities due within a year — the buyback is levering the balance sheet, and debt markets, not management, set the limit.
• Sands China's net income fell 29.8% on 1.4% revenue growth — the Macao profit engine that funds the dividend is deteriorating faster than the consolidated numbers suggest.

Our Take: Buy the operations, fade the financial engineering. The Q3 miss is luck rolling off a lucky comparable — volumes, occupancy, room rates and nine-month operating income all say the underlying business is as healthy as it has ever been, and nothing in the quarter changes the MBS expansion thesis. But a management team that answers a soft quarter with $950 million of buybacks into a negative-equity balance sheet and a 20% dividend hike is optimizing for shareholders who sell, not shareholders who hold. The operations deserve a premium multiple; the leverage deserves a discount rate. Net those honestly and Las Vegas Sands is fairly priced for perfection on luck — hold the stock, and watch rolling-chip volume and resort-ops cost growth, the two numbers that will decide whether Q4 repeats Q3.