Tinder DAUs Grew for the First Time in 3½ Years on $10 Million of AI Spend — Match Group's 14% EBITDA Jump Is Mostly a Base Effect

Match Group reported third-quarter revenue of $891.2 million, a hair below the roughly $892 million analysts expected and inside the $885-895 million it had guided. Adjusted EBITDA of $344 million came in above the $330-335 million guided, and GAAP diluted EPS of $0.73 beat the $0.71 consensus. The beat was in margin, not in sales — revenue for the quarter was $23.0 million lower than a year earlier while adjusted EBITDA was $43 million higher.

The engagement headline came out of the same release: Tinder's daily active users returned to year-over-year growth in September for the first time in three and a half years, and U.S. daily users grew in October. That is the first hard evidence that Tinder's three-year user decline has stopped. The cash-return machinery also ran: trailing-twelve-month free cash flow of $1.14 billion funded $614.5 million of buybacks and $182.2 million of dividends, and the board declared a $0.20 quarterly dividend.

Decompose GAAP operating income and the AI story separates cleanly from the margin story. Operating income rose $47.2 million to $268.6 million. Cost of revenue fell $40.5 million even as revenue slipped, lifting gross margin about 380 basis points to 76.8%. General and administrative expense fell $55.8 million. Research and development rose $14.8 million and sales and marketing rose $15.9 million — the lines where product engineering and user acquisition live. Match Group does not disclose how its roughly $10 million AI budget is split across expense lines, so the honest reading is that the AI money shows up as research and development while the margin shows up in general and administrative.

There is a second-order point in the gap between the two profit measures. Adjusted EBITDA rose 14% while GAAP operating income rose 21%. The difference is the add-backs — stock compensation, depreciation and other items — which shrank from about $80.5 million to about $75.4 million of the gap between the two. In other words, the same 14% headline is built on a cost base that carried heavier charges a year ago, which is why the margin is better without the business being better.

Match Group plans to spend about $10 million on AI tools in 2026, roughly double the $5 million first budgeted, and is slowing hiring to absorb the cost without laying anyone off. The payoff the company describes is engineering velocity, not headcount: the Missed Connections feature moved from idea to shipped product in weeks rather than the months or quarters that used to be normal, and real-time adaptive recommendations begin testing in the fourth quarter. That is engineering cycles measured in weeks instead of quarters on a product whose users had been leaving.

The conversion gap is the entire investment case. Part of the revenue shortfall was self-inflicted: Match Group said Tinder's user-experience testing cost about $8 million of revenue in the second quarter and was guided to about $10 million in the third, with the full-year drag lowered to $30 million to $40 million. Strip that out and Tinder's direct revenue of $480 million was close to flat year over year rather than down 2%. The payers, though, still fell: 8.9 million for Tinder, 13.8 million for the group.

Tinder's economics make the conversion worth more than the top line suggests. The segment runs a 50% adjusted EBITDA margin on $480 million of quarterly direct revenue, so each 1% of Tinder revenue is worth about $2.4 million of quarterly segment profit, or roughly $10 million a year. That is not a needle-mover against roughly $1.35 billion of adjusted EBITDA implied for 2026. The engagement inflection is an argument about the multiple, not about 2026 earnings.

Hinge is the part of Match Group a growth investor should look at first. Its $221 million of quarterly direct revenue annualizes to about $884 million and its $88 million of adjusted EBITDA annualizes to about $352 million — Hinge is now 36% of Tinder's profit pool on 46% of Tinder's revenue, with adjusted EBITDA growth roughly twice as fast. Hinge's monthly users rose 12%, and its European expansion markets were about 15% of direct revenue and about 25% of global monthly users, with Brazil and Mexico up at least 50% since late-2025 launches and about 1.5 million monthly users in India. Everyone Everywhere is the mirror image: revenue down 20%, adjusted EBITDA down 28%, a shrinking annuity that still contributes $45 million a quarter of profit.

Management guided the fourth quarter to $865 million to $875 million of revenue, down 1% to flat, and adjusted EBITDA of $327.5 million to $332.5 million — about 11% below the year-ago quarter at the midpoint, on a 38% margin against 38.6% just reported. That is roughly 400 basis points of margin compression implied in a single quarter and is the strongest single argument against the bull case. The published guidance does not itemise what drives the step-down; the company says only that it is slowing hiring to absorb higher AI costs.

The arithmetic underneath the year is the real story. Nine months of 2026 adjusted EBITDA of $1,018 million plus the midpoint of the fourth-quarter guide implies roughly $1.35 billion for 2026, against about $1.24 billion in 2025 — growth of roughly 9% on revenue that guidance leaves flat at $3.47 billion to $3.48 billion. For 2026 the entire company story is margin, which is precisely why the 2027 question — whether engagement converts into payers — is the only thing that matters from here.

The numbers only make sense once you know what each buyer is paying for. Bumble's buyer paid $6.05 a share in cash — a 138% premium to the unaffected Oct. 9, 2026 close — for a business whose revenue fell 15% in the second quarter, and the roughly $1.3 billion transaction value works out to about 1.5 times trailing sales. On trailing EBITDA the mark looks far higher, at about 17.6 times, but that is flattered: Bumble's reported trailing EBITDA of $92 million carries a $173 million charge in second-quarter other expenses, and adding that back puts the mark nearer six times. Grindr, growing revenue 33% at a 42% adjusted EBITDA margin, trades at 5.1 times sales and 18.6 times trailing EBITDA.

Here is the synthesis. Match Group is a cash-flow machine with a re-accelerating engagement base and a still-shrinking payer base. Trailing free cash flow of $1.14 billion against a market capitalization of about $9.6 billion is an 11.8% yield, and $614.5 million of buybacks plus $182.2 million of dividends in the last twelve months is what turned a 2.5% revenue decline into 18% growth in GAAP diluted EPS. The market is paying for the buyback to outrun the revenue decline, and that is a workable trade for as long as the cash flows hold. What is new this quarter is that the product experiment finally moved users — and what is not yet new is that it has not moved payers.

So the honest position is neither the margin story nor the user story alone. On the margin story, the third-quarter beat was a base effect and the fourth-quarter guide gives it back. On the user story, the September and October inflections are the first credible evidence in three and a half years, and they cost roughly $10 million to produce — about 0.3% of revenue. The asymmetry sits in the payer count: if Tinder's 8.9 million payers turn, a stock on roughly nine times its own guided 2026 adjusted EBITDA re-rates toward the category's growth mark; if they do not, the buyback keeps per-share numbers rising on a shrinking base, which is a slow way to go nowhere.