Unity's Best Quarter: Vector Doubles Its Own Guide, Runtime Data Goes Live, and GAAP Profit Moves Up a Quarter

Results vs. Consensus

Unity reported before the open on August 6 and held its call at 8:30 a.m. ET the same morning, as it has for each of the past four quarters. The quarter beat consensus on every line and beat the company's own guidance by more, which is the more informative comparison: the guide was set only three months ago, after the Q1 print had already re-rated expectations upward.

Against the Company's Own Guidance

The guide-relative scorecard is where the quarter's character shows. Every line cleared the ceiling of the range, and the profitability lines cleared it by roughly four times the margin the revenue lines did.

Year-Over-Year Comparison

Quarter-Over-Quarter Comparison

Revenue

Total revenue of $546.5M grew 24% year-over-year, but the number that describes the business is strategic revenue of $486.4M, up 38%. The gap between them is the deliberately dying portfolio: non-strategic revenue of $60.1M fell 33% and consists of the ironSource Ads network, sunset effective April 30, and the Supersonic publishing business, sold to Tripledot Studios on August 4. By Q3 that line is guided to roughly $20M, essentially one month of Supersonic, and after that it goes to zero. The optical drag that made 2025 and early 2026 hard to read is nearly done, though it is widest in Q3: the guide implies total revenue growth of 19% to 21% against strategic growth of 44% to 47%, so the two rates diverge before they converge. They meet in Q4, once the non-strategic base laps out.

Within strategic revenue, the composition is lopsided and getting more so. Strategic Grow of $329.0M grew 63% and now represents 68% of strategic revenue. Strategic Create of $157.5M grew 5% as reported, or 14% excluding a $12M one-time revenue item in the prior-year quarter, and represents the remaining 32%. Management was explicit that the ironSource shutdown contributed almost nothing to the Grow number, which matters because the easy explanation for a 63% print would have been captive migration of ironSource advertisers onto Vector.

"Of note, the sunsetting of the ironSource Ad Network had a negligible positive impact on Vector growth in the quarter with only $3 million in Q2 Vector revenue growth from ironSource customers as a result of the shutdown." — Jarrod Yahes, CFO

Assessment: disclosing the $3M figure unprompted was the right call and closes off the most obvious skeptical reading of the quarter. Roughly $128M of incremental year-over-year strategic Grow revenue, of which $3M is migration, means the growth is being won in open competition for advertiser budget rather than harvested internally.

Margins

Adjusted EBITDA of $160.2M produced a 29% margin against 21% a year ago, an 800 basis point expansion, and grew 77% year-over-year, more than double the 38% growth rate of strategic revenue. The mechanism is visible line by line in the reconciliation: adjusted sales and marketing fell 2% in absolute dollars to $94.1M (17% of revenue, from 22%), adjusted general and administrative fell 9% to $37.9M (7% of revenue, from 9%), and adjusted research and development rose 17% to $161.1M but still declined as a percentage of revenue, from 31% to 30%. Adjusted gross margin held at 83%, unchanged year-over-year, so none of the leverage came from cost of revenue.

That last point is the important one. A margin story built on gross-margin expansion in a business at 83% has limited room left. A margin story built on holding two of three operating expense lines flat in dollars while revenue compounds has considerably more, and it is what management is guiding to: Q3 adjusted EBITDA of $185–190M implies a 33% margin, up another 400 basis points sequentially and 1,000 basis points year-over-year.

Assessment: the margin expansion is structural, not cyclical, and the incremental drivers for the second half are already contracted rather than forecast. The ironSource cost base leaves the system after an April 30 sunset, and the Supersonic cost base leaves after an August 4 close. Both are done deals as of this report, which makes the Q3 margin guide unusually well-underwritten for a 400 basis point sequential step.

EPS and Cash

Adjusted EPS of $0.28 grew 56% year-over-year and beat the $0.25 consensus. GAAP loss per share narrowed to $(0.05) from $(0.26), with the GAAP net loss of $22.7M down from $107.4M. The GAAP convergence is the point: intangible assets net fell to $163.1M from $650.5M at December 31, reflecting the Q1 impairment plus ongoing amortization run-off, and quarterly amortization of intangibles fell to $77.5M from $86.2M. Combined with stock compensation down 25% in dollars to 13.8% of revenue, the two largest non-cash charges standing between adjusted and GAAP profitability are both shrinking at pace.

Free cash flow of $202.0M grew 59% and exceeded adjusted EBITDA for the quarter, helped by a $63.1M build in publisher payables. Cash and equivalents including restricted cash reached $2,357.1M against $2,237.4M of convertible notes, so Unity ended the quarter with roughly $120M of net cash, its first net cash position since 2021, before the convertible notes were drawn down.

Assessment: the cash conversion is genuine but flattered this quarter by working capital. Publisher payables rising with Vector volume is a real and durable feature of a scaling ad network rather than a pull-forward, though it does mean free cash flow will run above EBITDA in periods of rapid Grow acceleration and below it if growth ever flattens. On a first-half basis, $268.4M of free cash flow against $298.5M of adjusted EBITDA is the cleaner conversion read.

Segment Performance

Unity also reports on its legacy two-segment basis, which allocates the dying non-strategic revenue into Grow. On that basis Grow Solutions revenue was $389M against $287M, up 35%, and Create Solutions revenue was $158M against $154M, up 2%.

Strategic Grow

This is the quarter Vector stopped being a recovery story and started being a growth story on its own terms. For four consecutive quarters the product had compounded at roughly 15% sequentially, a cadence steady enough that management internally planned the fifth quarter at 12–13%, as disclosed on this call. It delivered roughly 23%, and the beat was not a late-quarter surge that pulled forward Q3 demand: the third-quarter guide of 19–21% sequential implies the elevated rate substantially persists.

"Our expectations for growth in Vector in Q2 were a robust 12% to 13% growth quarter-over-quarter. Instead, our team delivered nearly double that, racking up 23% quarter-over-quarter growth and establishing an accelerated momentum that has carried over to elevated results in Q3." — Matt Bromberg, President & CEO

Management attributed the step to an unusually heavy product cadence, more than 20 major Vector updates in the quarter, of which the most consequential was the Day 28 return-on-ad-spend capability released at the end of Q1. Day 28 ROAS lets advertisers optimize against a 28-day return window rather than the standard seven days, which materially changes which users the bidding system will pay up for. Campaign spend on the new product grew nearly three times from Q1, and more than 25% of the advertiser base has adopted it.

The scale milestone landed early. Vector passed a $1 billion annual run rate two quarters ahead of the company's own schedule, having not existed six quarters ago.

"Keep in mind, 6 quarters ago, the Unity Vector product didn't exist. It is now at substantially over $1 billion in annual run rate, 2 quarters earlier than expected. And yet, we still consider ourselves to be in the very early stages of product development." — Matt Bromberg, President & CEO

Assessment: the guide-versus-actual gap is the most valuable disclosure in the quarter, because it tells you management does not yet have a reliable model of its own product's growth rate. That cuts both ways. It means upside can arrive unforecast, as it did here, and it means the 19–21% Q3 guide carries wider error bars than its precision implies. The structural read is that Vector's growth is a function of product velocity plus data quality plus model self-improvement, and Unity currently controls all three levers and is pulling them simultaneously.

Strategic Create

Create's reported 5% growth understates it and its 14% ex-item growth flatters it slightly, because the $12M one-time revenue item being excluded sat in the prior-year base. The underlying business grew mid-teens for a fifth consecutive quarter, driven by average revenue per user rather than seat count: price increases, minimum annual customer commitments, and continued strength in China.

"In Create, strategic revenue was $157 million, up 14% year-over-year when excluding the impact of a onetime revenue item in the prior year. Create momentum continues to be driven by ARPU growth, supported by price increases and minimum annual customer commitments as well as strong growth in China." — Jarrod Yahes, CFO

The forward guide of $159–163M implies 7–10% year-over-year growth, a step down from the 14% underlying rate, and only about 1% to 3.5% sequentially. Management did not reconcile the deceleration, and it is the one number in the guidance package that points the wrong way.

The strategic development in Create is not the revenue line but Unity 7, announced two weeks before the call at the Unite conference in Seoul, with a Q4 2026 beta and a Q1 2027 full release. Unity 7 rearchitects the engine so that human developers and coding agents can work in the same project simultaneously, makes the Model Context Protocol implementation free, and opens the API so agents can drive Unity from outside the application.

"We've made our MCP free and opened our API, which enables developers to use the command line and coding agents to control Unity directly from inside their own workflows." — Matt Bromberg, President & CEO

Assessment: Create is doing its job in the thesis, which is to compound quietly at a mid-teens rate while Grow carries the growth. The Q3 guide implying 7–10% deserves watching, since two consecutive single-digit quarters would break the "Create has stabilized" pillar that has held since Q3 2025. Unity 7 is strategically the more important item, and it contains an unresolved tension the call did not address: if developers increasingly reach Unity through agents and the command line rather than the seat-based editor, the seat-based pricing model for Create is exposed even as platform usage rises.

Non-Strategic Wind-Down

The portfolio surgery announced at Q1 is now complete. The ironSource Ads network closure took effect April 30, and the Supersonic publishing business closed its sale to Tripledot Studios on August 4, two days before the print. Assets held for sale of $56.0M and liabilities held for sale of $19.9M sat on the June 30 balance sheet as a result. Restructuring and reorganization costs of $31.4M ran through the quarter, against $10.9M a year ago.

The company also made a strategic investment in AppsFlyer alongside Meta, Google, and Moloco, framed as preserving independent measurement in mobile attribution.

Assessment: the wind-down is the reason the Q3 margin guide is credible, since both cost bases exit the system in the second half. It is also the reason the year-over-year comparisons become clean from Q3 onward. The AppsFlyer investment was disclosed without a dollar amount, ownership percentage, or accounting treatment, and a co-investment in the leading independent measurement vendor alongside three of the largest ad platforms is not a trivial item to leave unsized.

Key Operating Metrics

Key Topics & Management Commentary

Overall Management Tone: the most confident posture in our coverage of this name, and the first call in which management spoke prospectively rather than defensively for its full length. Where the prior three calls were organized around rebutting a bear case (the flat-Grow fear at Q4, the world-model threat at Q1), this one was organized around a forward framework, with the CEO opening on a flywheel construct and declaring the company "done looking backwards." The confidence was earned by the numbers and was notably not accompanied by any loosening of disclosure: management remained as unwilling to quantify runtime data, Vector dollars, or Unity AI adoption as it was when the stock was at $17.

1. Vector Broke Its Own Guide, and Management Does Not Fully Know Why

The central fact of the quarter is a 10 percentage point miss in the company's own favor on its most important forecast. Planning 12–13% sequential growth for a product with four consecutive quarters of roughly 15% sequential history, and then printing 23%, means the internal model does not capture the product's own dynamics. Pressed on whether step-function improvements are a repeatable pattern, the CEO described a three-part mechanism rather than a forecast.

"So you need continuous product enhancement, which we are accelerating the velocity of which is accelerating and which we feel we have lots more to do. Now every quarter won't be as busy as this quarter in terms of product enhancements. We're not going to do 20 product enhancements every quarter, but we are going to continually enhance the product as quickly as we possibly can." — Matt Bromberg, President & CEO

Assessment: the honest read is that Q2 combined an unusually heavy release cadence with a high-leverage product (Day 28 ROAS) at exactly the moment runtime signal came online, and management is telling you the cadence component will not repeat at that intensity. The 19–21% Q3 guide is therefore the better run-rate estimate than 23%, and the risk to the thesis is not that growth reverses but that the Street extrapolates the step rather than the trend.

2. Runtime Data Went Live, Four Quarters After It Was First Promised

The longest-running item on the Unity thesis is the proposition that behavioral signal from the roughly three billion monthly players on games built with the engine constitutes an advertising data asset no competitor can replicate. It has been promised since 2025 and dated to Q2 2026 since the Q4 call. It landed.

"Towards the end of the second quarter, we began for the first time in our history to unlock this advantage by incorporating signals from our runtime directly into our Vector AI models. While this effort remains very early, the results are extremely encouraging, and we gain more and more confidence each day that runtime represents a deep and sustainable competitive advantage for Unity." — Matt Bromberg, President & CEO

The strategic framing management uses to distinguish its position from the platforms is worth recording, because it is the clearest articulation the company has given of why this asset is structurally different from a retargeting pixel.

"Every major participant in this world has an approach. If you're Meta, you own your apps. If you're us, you have access to the 3 billion people playing with Unity game, and that's going to put us in a really good spot over the long term." — Matt Bromberg, President & CEO

Assessment: delivery on schedule converts the pillar from a promise to an asset, and it arrived roughly six to eight weeks before the call, meaning essentially none of Q2's 23% is attributable to it. That makes it genuine unmodeled upside for the first time. The offsetting fact is that management has now declined to size it for five consecutive quarters and has escalated from "too early to quantify" to arguing quantification lacks analytical value.

3. GAAP Profitability Pulled Forward a Quarter

At Q1 management put a hard date on GAAP net income profitability for the first time: Q4 2026. One quarter later it moved the date closer, which is a rarer event than a revenue beat.

"Lastly, we're pleased to report that we're pulling forward our expectation for achieving GAAP net income profitability from the fourth quarter of 2026 to the third quarter of 2026." — Jarrod Yahes, CFO

The bridge is mechanical and visible on the face of the statements. GAAP loss from operations narrowed to $32.2M from $118.8M. Intangible assets net fell to $163.1M from $650.5M at year-end, so the amortization charge that has suppressed GAAP results since the ironSource acquisition is running off. Stock compensation fell 25% in dollars to 13.8% of revenue.

Assessment: pulling a profitability date forward is a high-confidence signal because it is easy to verify in one quarter and embarrassing to miss. With Q3 adjusted EBITDA guided to $185–190M against roughly $85M of quarterly depreciation and amortization and roughly $76M of stock compensation, the arithmetic is tight but achievable, and the guide implies management has run it.

4. The Balance Sheet Flipped to Net Cash

Unity ended the quarter with $2,357.1M of cash including restricted cash against $2,237.4M of convertible notes, roughly $120M of net cash and the first such position in its public life. The $557.2M note due November 2026 sits in current liabilities and will be retired from the balance sheet rather than refinanced.

"This brings our cash balance to $2.36 billion with Unity flipping from a net debt position to a net cash position this quarter. Our near-term capital plans center around delevering our balance sheet, and we expect to pay off our 2026 convert in November." — Jarrod Yahes, CFO

Assessment: this fully retires the balance sheet as a thesis consideration. It was already a contained risk after the 2025 refinancing; it is now a source of optionality, and the CFO flagged as much in noting that a delevered balance sheet plus record cash generation "opens up tremendous flexibility from a capital allocation standpoint." No buyback was announced and none should be assumed, but the capacity now exists.

5. Day 28 ROAS: the Product Behind the Number

The specific mechanism behind Vector's step-change is a longer optimization window. Standard performance marketing in mobile gaming optimizes to a seven-day return signal, which systematically underweights players whose value accrues over a longer horizon. Extending to 28 days changes the bid calculus for exactly the high-lifetime-value cohorts advertisers most want.

The adoption data is the strongest new disclosure of the quarter: campaign spend on Day 28 grew nearly three times sequentially, and more than a quarter of the advertiser base has adopted it within roughly one quarter of general availability.

Assessment: a product that a quarter of the advertiser base adopted in a single quarter with spend tripling is doing something the alternatives are not, and it explains why the growth did not require migration from ironSource. The unmodeled question is how much runway remains: with 25% adopted, three-quarters of the base has not yet moved, which is either a large remaining tailwind or evidence that adoption concentrates among the sophisticated minority.

6. The Flywheel Reframe

The CEO devoted his opening remarks and three separate Q&A answers to reframing how the two segments relate, arguing that the market's habit of modeling Create and Grow as independent businesses misreads the strategy. The construct is that easier game creation produces more games, more games drive more platform usage and make discovery harder, and harder discovery drives advertising demand, all informed by runtime data.

"Folks historically are very focused on the Unity application, and it's just not how we think about it. It's not how it works. More folks create interactive entertainment, more folks on the platform, more need for advertising." — Matt Bromberg, President & CEO

Assessment: the flywheel is analytically real but currently unmeasurable, and it is being introduced at a moment when it conveniently reframes the weaker segment as an input to the stronger one rather than a business that must justify itself. We take it seriously as strategy and discount it entirely in the model until an attach-rate or cross-sell disclosure exists.

7. Unity 7 and the Agent-Native Rearchitecture

Unity 7, announced at Unite Seoul, is positioned as a generational change rather than a version increment: humans and coding agents working in the same project simultaneously, free MCP, open API, no forced upgrade, and full backward compatibility with Unity 6.

"this isn't just another engine upgrade for us. It's a complete change in how developers are going to interact with our technology and how they're going to utilize coding agents in concert with Unity and how they're going to collaborate with their teams." — Matt Bromberg, President & CEO

The release cadence is itself a data point. The CEO noted the gap between Unity 5 and Unity 6 was roughly seven years, and between Unity 6 and Unity 7 roughly 18 months.

Assessment: this is the strongest structural answer yet to the world-model disruption case, and it is a different answer from the one given at Q1. At Q1 the argument was that Unity ingests generative output and assembles it. Here the argument is that Unity becomes the substrate agents operate inside, which is a more defensible position because it does not depend on winning any generation contest. The unaddressed risk is the pricing model: a free MCP and an open API that let agents drive Unity from outside the editor are excellent for platform reach and awkward for per-seat Create revenue.

8. Create's ARPU Engine and the Q3 Deceleration

Create's mid-teens underlying growth continues to come from price and commitment structure rather than seat expansion, with China called out again as a strength, consistent with the last four quarters. Management noted five straight quarters of sequential growth in the segment.

Against that, the Q3 guide of $159–163M implies 7–10% year-over-year and only about 1% to 3.5% sequentially. No explanation was offered and no analyst asked for one.

Assessment: this is the quarter's clearest unforced disclosure gap. A guide implying roughly half the underlying growth rate, in the segment most exposed to the AI-disruption narrative, went unexplained on a call where management otherwise volunteered detail freely. The benign reading is conservatism plus the tail of the 2025 large-customer comp; the unfriendly reading is that price-led growth is running out of room. Two single-digit quarters would challenge a standing thesis pillar.

9. Operating Leverage as Declared Strategy

Asked for a long-term incremental margin framework, management declined to give a target but committed to the direction, with the CEO framing it as a resource-allocation discipline that matters more in good times than bad.

"We are not going to fall into that trap. We're going to continue to apply capital to our fastest-growing businesses and prioritize those and not get distracted by things that are not creating growth and profitability for us. That's our strategy." — Matt Bromberg, President & CEO

The CFO supplied the quantitative frame, noting 200 basis points of margin expansion per quarter since Q1 2025 and 400 basis points guided for Q3 alone.

Assessment: a company that grows revenue 38% while cutting two of three operating expense lines in absolute dollars has demonstrated the discipline rather than merely asserted it. The absence of a long-term target margin is a modest negative for modeling, but the quarterly cadence disclosure is a usable substitute.

10. Netflix, Platform Relationships, and Commerce

Unity signed a multi-year agreement to support the Netflix multi-platform games ecosystem, which management framed as emblematic of its platform relationships rather than as a discrete revenue event.

"The partnership calls for us to comprehensively support the Netflix multi-platform games ecosystem with the Unity engine in the years ahead." — Matt Bromberg, President & CEO

Separately, Unity Commerce reached general availability on June 30. The product is free to developers, targets direct-to-consumer mobile monetization that management sizes at roughly 15% of the mobile market, and carries three stated benefits: helping publishers avoid app-store fees, giving Unity visibility into purchase behavior that improves the ad models, and a small direct economic benefit that scales over time.

Assessment: the Netflix deal was announced with no economics, no term, and no minimum, so it is a credibility marker rather than a modelable item. Commerce is the more interesting of the two because the second stated benefit, purchase-behavior signal feeding the ad models, is the same architectural logic as runtime data and compounds with it.

11. Advertiser Penetration Is Intensive, Not Extensive

Asked where Unity sits on advertiser penetration, management said the growth is not coming from new logos.

"I would say that, by and large, in the game space, the vast majority of folks already know who we are, and are already spending with us to one degree or another." — Matt Bromberg, President & CEO

Assessment: this is an important qualifier on the growth algorithm and cuts against the most bullish framing. If nearly all gaming advertisers already spend with Unity, then Vector's growth is budget share within existing accounts, which is won on performance and can be lost the same way. It is a higher-quality growth source than land-grab, and a more contestable one.

Guidance & Outlook

The guide's shape is the story. Strategic revenue at the $545M midpoint sits roughly 1% above where the Street was carrying it, which by itself would have been an unremarkable in-line guide. Adjusted EBITDA at the $187.5M midpoint sits roughly 23% above consensus. The market was modeling the revenue acceleration and was not modeling the margin structure that comes with it, which is why the reaction was as large as it was on a revenue guide that barely beat.

Implied sequential ramp: the strategic revenue guide requires 11–13% sequential growth, composed of 15–17% sequential in Grow and roughly 1–3.5% in Create. Vector at 19–21% sequential does the work; Create at the low end would represent its weakest sequential quarter in over a year.

Street at: roughly $538.8M strategic revenue and roughly $152.2M adjusted EBITDA before the print. The revenue guide clears consensus by about 1% at the midpoint; the EBITDA guide clears it by about 23%.

Guidance style: demonstrably conservative on the metric that matters most. Management planned Vector at 12–13% sequential and delivered 23%. It guided strategic revenue to $455–465M and delivered $486.4M, and adjusted EBITDA to $130–135M and delivered $160.2M. Three consecutive quarters of clearing the top of the range means the Q3 guide should be read as a floor rather than a midpoint, with the caveat that the company's forecasting error has been widening in its own favor, which is its own kind of imprecision.

What is still missing: a full-year framework, for the fifth consecutive quarter. Management now guides one quarter at a time with no annual revenue range, no annual segment outlook, and no Vector dollar target. The GAAP-profitability date partially substitutes as an out-year anchor, and its pull-forward strengthens that substitute, but it is not a revenue framework.

Analyst Q&A Highlights

Eight questions were taken in a notably short and uncontentious session. The striking feature is what was not asked: no one questioned the Create guidance deceleration, the AppsFlyer investment size, or the absence of a full-year framework. The Netflix relationship drew a single request for scope and economics, which management declined to answer. The Q&A was almost entirely forward-looking and strategic, which is itself a signal about how the print was received.

Whether Runtime Data Can Be Isolated and Measured

The opening question went straight to the quarter's most-watched item, asking whether the runtime contribution was material or small given that it came online only late in the quarter. The response declined the premise of the question rather than the number, arguing that isolating any single input misunderstands how a self-learning system compounds. It is a defensible position and also a convenient one.

Q: "D28 was a big driver of strength for Vector in the quarter. Was runtime fee a material contribution or relatively small? And then based on what you've seen so far, what does it tell you about what you can expect that to contribute in the coming quarters?"
— Matthew Cost, Morgan Stanley

A: "I don't know there's any analytical value for anyone to single out the impact of runtime data. Our performance, as we've noted, is the sum of the impact of all the product enhancements, the higher quality data and the compounding model improvements. And these impacts are self-perpetuating as our system continues to learn in real time."
— Matt Bromberg, President & CEO

Assessment: this is an escalation. For four quarters the position was that it was too early to quantify the runtime contribution, which implied quantification would eventually come. The position is now that quantification has no analytical value, which implies it never will. Investors should assume the runtime contribution will never be separately disclosed and should value it as an embedded feature of Vector's growth rate rather than as a line item that will one day be revealed.

Whether Step-Function Improvements Are a Repeatable Pattern

The most useful exchange for modeling purposes probed whether the 23% quarter reflected a new baseline or an unusually productive release cycle. The answer separated the three drivers and conceded that one of them, product cadence, will not sustain at Q2's intensity.

Q: "Vector growth, 23% in the quarter, you're guiding 20% in the forward quarter. Maybe I missed this in the past, but I don't think I've heard you guys sort of call out sort of step function model improvements on a quarter-by-quarter basis. Has that been a driver sort of historically or something that we should think about as a potential future opportunity?"
— Clark Lampen, BTIG

A: "Now every quarter won't be as busy as this quarter in terms of product enhancements. We're not going to do 20 product enhancements every quarter, but we are going to continually enhance the product as quickly as we possibly can. But product enhancements are just one leg of that stool. The second leg of the stool is higher quality data, better signal."
— Matt Bromberg, President & CEO

Assessment: the most candid answer on the call, and it argues against extrapolating 23%. The useful structure is that two of the three drivers (data quality and model self-improvement) are continuous and compounding while the third (release cadence) is lumpy and just had an exceptional quarter. That supports the high-teens sequential guide as the run rate and treats quarters like this one as episodic upside.

Unity AI Unit Economics and the Credit Model

A question on how the company balances included credits against margin in the Unity AI subscription tiers received an answer about product strategy rather than unit economics. Management reframed toward developer choice and platform usage, declining to address the credit-versus-margin trade directly.

Q: "Unity AI, I would love to get a sense from you as to how you balance the need to kind of preserve margin and unit economics relative to the amount of credits you include in each of the subscription tiers. How do you think about finding that balance, delivering enough value to the customer while protecting the bottom line?"
— Alec Brondolo, Wells Fargo

A: "our strategy here is to give developers a choice. You can use our bespoke AI or use your own. We don't care. Our primary goal is to maximize the usage of the authoring platform that I was just describing to Matt. So we're not looking to restrict our customers, and which is why we opened up both our CLI and our API in Unity 7."
— Matt Bromberg, President & CEO

Assessment: the question was about gross margin on inference and the answer was about strategic positioning. That is a dodge, though possibly an honest one: if Unity genuinely does not care whether developers use its model or a third party's, then inference cost is not a margin question because Unity is not underwriting much inference. The unresolved issue is that thirteen weeks into open beta there is still no adoption, conversion, or pricing disclosure for the product designated as the answer to the world-model threat.

Unity 7 as an Architectural Break Rather Than a Version

A request to explain in plain terms how Unity 7 differs from its predecessor produced the call's most substantive product answer, centered on multi-party and agent collaboration inside a single project, the elimination of the traditional upgrade cycle, and a compressed release cadence.

Q: "can you explain to us in simple terms, please, how it's different from the previous generation and how you see it supporting growth in both of your segments?"
— Vasily Karasyov, Cannonball Research

A: "I would also note that I believe the wait between Unity 5 and Unity 6 was something like 7 years. We were closer to 18 months, and that's going to continue to accelerate. So it is a very different product than it was before."
— Matt Bromberg, President & CEO

Assessment: the release-cadence compression from seven years to 18 months is the most concrete evidence offered that AI has changed Unity's own internal productivity, which is the claim underlying the entire product-velocity leg of the Vector story. It is also the only place on the call where the AI-improves-our-execution argument was supported with a verifiable number rather than an assertion.

Long-Term Incremental Margin Framework

The question sought a structural framework for incremental margins and the trade-off against reinvestment. Management declined to name a terminal margin but supplied the quarterly cadence and an explicit commitment to not spend the upside.

Q: "how would you think about long-term incremental margins in the business from both a business mix perspective, but also the balance you would likely have to strike between maintaining that momentum with growth investments and letting some elements of just the pure incremental margins of this mix flow through to the bottom line?"
— Eric Sheridan, Goldman Sachs

A: "We've experienced 200 basis points of margin expansion each quarter since the first quarter of 2025. And based on the strategic actions we've recently executed, we're looking at 400 basis points of operating margin expansion of EBITDA margin expansion in the third quarter alone. We also believe that there is further opportunity for margin expansion in the business."
— Jarrod Yahes, CFO

Assessment: no terminal margin was given, which is a real modeling gap for a business whose margin trajectory is now the primary valuation driver. The 200 basis points per quarter cadence since Q1 2025 is a usable proxy, and the acknowledgment of "further opportunity" beyond a 33% Q3 margin implies management sees a structurally higher terminal level than the mid-30s the current trajectory reaches by year end.

Cross-Sell Between the Engine and the Ad Platform

A question on how Unity 7 unlocks cross-sell into advertising drew the call's fullest articulation of the platform thesis, culminating in a claim about market expansion that goes well beyond the current customer base.

Q: "I want to go back to Unity 7 and think about how we should be thinking about the cross-sell in terms of the advertising business and then core Create. You mentioned it earlier, but it's always been this massive opportunity for Unity. Can you really unpack that as we start to think about '27 and the next generation of Unity of bringing those 2 sides closer together?"
— Andrew Boone, Citizens

A: "I am a fond of saying creation is the new consumption. What I mean by that is we believe that tens of millions of new folks will begin creating interactive entertainment just as they today create linear video."
— Matt Bromberg, President & CEO

Assessment: the prosumer-expansion thesis is the most speculative claim management made and was offered without a timeline, a product, or a pricing model. It should carry no weight in a 12-month view. The narrower and more credible part of the same answer, that games built through agentic workflows arrive with Vector and commerce configured by default, is a genuine structural advantage and the mechanism by which Unity 7 could lift attach rates without a sales motion.

Where Advertiser Penetration Stands

The closing question asked how much of the addressable advertiser base Unity has captured and what it would take to accelerate adoption. The answer characterized the market as already penetrated and the work as optimization within existing accounts.

Q: "Could you maybe just give us an update on where you think you are in terms of advertiser penetration as sort of like a percentage of all the advertisers in the market, how that's trended over time and what you're doing to accelerate that adoption or if you need to do anything at all in that advertisers just come to you when they see returns?"
— Omar Dessouky, Bank of America

A: "it's not as if the issue for us is we've got to ramp up some big new sales force to ring doorbells to try to get folks to try Vector. It is much more of what I describe as an optimization and account management function in which we test and scale individual games and do the work around data and integration to ensure that we are able to deliver as much scale and as much value as we possibly can for each advertiser."
— Matt Bromberg, President & CEO

Assessment: a valuable disclosure delivered as a throwaway. It confirms the growth is budget-share capture inside accounts Unity already serves, which explains both the operating leverage (no sales-force build) and the risk profile (performance-contingent, and reversible if a competitor's models improve faster). It also implies the growth ceiling is set by advertiser budgets and Unity's share of them rather than by logo count.

What They're NOT Saying

1. No full-year 2026 framework, for a fifth consecutive quarter. Management guides one quarter forward with no annual revenue range, no annual segment outlook, and no margin target beyond Q3. The pulled-forward GAAP profitability date is a partial substitute, but a company guiding a 44–47% strategic acceleration and declining to frame the year is withholding the disclosure the Street has asked for since the Q4 2025 call.
2. Runtime contribution moved from unquantified to unquantifiable. The shift from "too early to size" to "no analytical value in singling out" is a meaningful change in posture on the single most-discussed element of the thesis. It should be read as a permanent disclosure decision rather than a timing one.
3. Unity AI still has zero commercial metrics. Thirteen weeks after entering open beta, there is no user count, no conversion-to-paid figure, no attach rate, and no realized pricing. At Q1 the ~70% day-five retention statistic was offered as early proof; this quarter even that was not updated. The product designated as the answer to the world-model disruption case remains commercially unsized.
4. Vector is still not disclosed in dollars. "Substantially over $1 billion in annual run rate" is the most precise figure given, and it arrived alongside a sequential growth rate but no absolute quarterly revenue. Analysts cannot independently track the Vector-versus-legacy split on which the entire Grow thesis rests.
5. The Netflix deal has no disclosed economics. A multi-year platform agreement described as important was announced without revenue contribution, term length, minimum commitment, or exclusivity. It functions as a credibility signal and nothing more.
6. The AppsFlyer investment was not sized. No dollar amount, no ownership stake, no accounting treatment, and no governance detail for a co-investment made alongside Meta, Google, and Moloco in the leading independent mobile measurement vendor. Both the strategic rationale and the balance-sheet impact are unassessable.
7. The Supersonic sale price was not disclosed. The business closed to Tripledot Studios on August 4 with $56.0M of assets held for sale on the balance sheet and no consideration disclosed, so the gain or loss and the cash proceeds are unknown ahead of the 10-Q.
8. Create's guided deceleration went unexplained. Guiding 7–10% year-over-year against a 14% underlying rate, with only about 1% to 3.5% sequential, is the one number in the package pointing the wrong way, and neither management nor any analyst addressed it.
9. Unity 7's effect on Create pricing was not addressed. Free MCP and an open API let agents drive the engine without a developer in the seat-based editor. Management's answer, that it does not care how customers access the interface because monetization happens downstream, sidesteps what happens to seat-based Create revenue if editor usage declines while platform usage rises.
10. No terminal margin target. Management confirmed "further opportunity for margin expansion" beyond Q3's 33% without naming a ceiling, leaving the primary valuation driver of the stock without a management-supplied endpoint.

Market Reaction

• Pre-print setup: Unity closed at $35.47 on August 5, down 19.7% year-to-date against an S&P 500 up 12.8%, but up 19.6% over the trailing 30 days and 11.1% over the trailing twelve months. The 52-week closing range was $17.13 to $49.47, so the stock entered the print in the upper-middle of its range after a sharp run into the event.
• Reaction session (August 6, print day): the stock opened at $36.01, up just 1.5%, traded as low as $34.80 (below the prior close) and as high as $41.66, and closed at $40.81, up 15.1% or $5.34. Volume was 41.0M shares against a 30-day average of 7.6M, or 5.4x. The S&P 500 fell 0.2% on the day.
• Follow-through (August 7): the stock opened at $42.54, ranged $41.78 to $44.36, and closed at $43.00, up a further 5.4%. Volume of 18.5M was 2.4x the 30-day average.
• Two-session cumulative: $35.47 to $43.00, up 21.2%. At $43.00 the stock is down 2.6% year-to-date, having entered the print down 19.7%, and sits 13.1% below its 52-week closing high and 151% above its February closing low.

The intraday shape on the print day is more informative than the close. Unity reported before the open, so the entire package including the Q3 guide was public before the bell and the stock opened up only 1.5%, then traded below the prior close within the first hour. The buying came through the session and the close was near the high on 5.4x volume. That pattern says the initial pre-market read was cautious and conviction built during and after the 8:30 a.m. call, which is consistent with the guide's shape: a strategic revenue number roughly 1% above consensus is easy to skim past, while the 23% adjusted EBITDA beat against consensus and the GAAP-profitability pull-forward required the call to contextualize.

The follow-through day matters as much as the reaction day. Unity has twice in the past year gapped up on a strong print and given the move back within the session, at Q2 2025 and again at Q1 2026, both times because a crowded or heavily-shorted position book used the strength to exit. This time the stock closed near the high and then added 5.4% on above-average volume the next day, with no giveback. Two-day persistence on declining but still elevated volume is the signature of position-building rather than a short squeeze exhausting itself.

What the move does not represent is a valuation reset from depressed levels. Entering the print Unity already traded at roughly 7.6x enterprise value to trailing revenue, having re-rated from 3.7x at the February low and 5.7x after the Q1 print. The two-day move took it to 9.3x. The re-rating that the Q1 recap identified as underway is now substantially complete, and the stock has recovered nearly its entire year-to-date drawdown in two sessions.

Street Perspective

Debate: Is Vector's Step-Change a New Baseline or a One-Quarter Cadence Effect?

Bull view: the bull case on the Street is that 23% sequential growth against a 12–13% internal plan, followed by a 19–21% guide for the next quarter, describes a product whose growth rate has structurally reset rather than spiked. Runtime data came online only in the final weeks of Q2 and contributed nothing meaningful to the quarter, so the compounding data-quality leg of the growth algorithm is still almost entirely ahead. On this read the $1 billion run rate arriving two quarters early is evidence the company systematically underestimates its own product, and the Q3 guide should again be treated as a floor.

Bear view: the bear camp contends that management itself explained the quarter as a function of more than 20 product enhancements and said plainly that this cadence will not repeat. Strip the unusually heavy release cycle and the underlying rate is closer to the 15% the product ran at for four straight quarters than to 23%. The Q3 guide of 19–21% already implies deceleration, comparisons harden materially through 2027 as the base compounds, and management's own admission that nearly all gaming advertisers already spend with Unity means the growth must come from budget share, which is contestable in a way that new-logo growth is not.

Our take: the bear has the better of the mechanics and the bull has the better of the direction. Extrapolating 23% is a mistake management explicitly warned against, and the honest base case is the high-teens sequential rate the company guided. But the bear's case is about the rate of deceleration from an exceptional level, not about whether growth continues, and it understates the runtime option, which is genuinely unmodeled and arrives on top of a growth rate that already works. We model high-teens sequential Vector growth through 2026 decelerating into the low teens during 2027, and treat runtime as upside we do not underwrite.

Debate: Has the Multiple Already Captured the Story?

Bull view: the bull argument is that 8.6x enterprise value to annualized forward strategic revenue, or roughly 25x annualized forward adjusted EBITDA, is not demanding for a business guiding 44–47% strategic revenue growth at a 33% margin with a net cash balance sheet and GAAP profitability arriving next quarter. The rule-of-40 arithmetic on the Q3 guide is roughly 78, a figure very few software businesses at this scale reach, and comparable AI-driven advertising platforms have historically carried materially higher multiples at lower growth rates.

Bear view: the bear camp contends the easy money has been made. The multiple has gone from 3.7x trailing revenue in February to 9.3x today, the stock is up 21% in two sessions and 151% off its low, and the entire year-to-date drawdown has been erased. From here the return requires either sustained upside surprises against a guide the Street has now marked up, or further multiple expansion beyond a level that already prices in successful execution. Meanwhile the disclosure gaps that justified a visibility discount are all still in place.

Our take: both are right about different horizons, and the resolution is that the valuation cushion is gone while the earnings trajectory is not. At 25x forward adjusted EBITDA growing that line roughly 70%, the multiple is not the constraint it would be if growth were decelerating toward the market's rate. What has changed is the margin of safety: in May we were underwriting a cheap business with a resolved bear case, and today we are underwriting a fairly-priced business with a proven one. That is still an Outperform, at lower implied return and with less room for a stumble.

Debate: Does Unity 7 Settle the World-Model Question or Reopen the Pricing Question?

Bull view: the bull view is that Unity 7 is a stronger answer to the AI-disruption case than Unity AI was, because it does not require Unity to win any generation contest. By rearchitecting so that coding agents operate inside Unity projects, making the protocol layer free and opening the API, Unity positions itself as the substrate that agentic development runs on rather than a tool competing with agents. Games built that way arrive with Vector, commerce, and live services configured by default, which lifts attach rates with no sales motion. A cadence compression from seven years to 18 months between major releases is offered as evidence the company can now out-ship the threat.

Bear view: the bear camp contends that opening the platform solves the strategic problem by creating a commercial one. Create is priced per seat in the editor. If agents drive Unity from the command line and the API, editor seats are decoupled from platform value precisely as usage grows, and management's response, that it does not care how customers access the interface, concedes the point rather than answering it. The guided deceleration to 7–10% in Create is read by this camp as the first evidence of that pressure rather than as conservatism.

Our take: the bull has the stronger strategic argument and the bear has identified the right unanswered question. Substrate-for-agents is a more durable position than assembly-layer-for-generated-content, and it is the best articulation Unity has offered of where its value sits in an AI-native stack. But the pricing model has not been updated to match the architecture, and the Q3 Create guide is the first datapoint that could be read either way. We hold the disruption risk contained on the strength of the architecture and add the Create monetization model as the specific thing to watch, since it is where a strategically correct decision could still cost revenue.

Model Update & Valuation

Every operating line in our framework moves favorably and the valuation line moves against us, which is the mirror image of the Q1 update where both moved in our favor.

Valuation framework: at the August 7 close of $43.00, with 440.0M shares outstanding and roughly $120M of net cash, Unity carries an enterprise value near $18.8B. Against trailing twelve-month revenue of $2,028M that is 9.3x, up from 7.6x entering the print, 5.7x after the Q1 print, and 3.7x at the February closing low. On forward bases the picture is less stretched: annualizing the Q3 strategic revenue guide midpoint gives $2.18B and a multiple of 8.6x, and annualizing the Q3 adjusted EBITDA guide midpoint gives $750M and a multiple of 25.1x.

The forward EBITDA multiple is the one we weight most, because the margin line is now growing faster than revenue and is the driver of the equity story. Roughly 25x a figure guided to grow 69–74% year-over-year is not an aggressive price, and the rule-of-40 score implied by the Q3 guide is approximately 78. Set against that, the trailing revenue multiple has nearly tripled from the February low and the stock has erased a 20-point year-to-date deficit in two sessions, so anyone underwriting further multiple expansion is underwriting sentiment rather than arithmetic.

Valuation impact: our framework moves up on earnings, not on multiple. We now carry strategic revenue compounding in the mid-40s through the second half of 2026 and decelerating toward the high-20s during 2027 as the Vector base compounds and comparisons harden, with adjusted EBITDA margins reaching the mid-30s by the end of 2026 and drifting higher thereafter. Applying a forward EBITDA multiple in the low-to-mid 20s to that trajectory supports a value modestly above the current price, which is a materially narrower gap than the one that existed at the Q1 upgrade. The rating is maintained on the strength of the earnings trajectory rather than on the discount.

Thesis Scorecard Post-Earnings

Overall: the thesis is strengthened materially on operations and weakened on price. Seven of nine pillars moved favorably or held, including delivery on both of the structural promises made at Q1 (runtime go-live and the GAAP-profit bridge), one of them a quarter early. The two that moved against us are the two that always constrain a stock after a print like this: valuation, whose trailing-revenue multiple has gone from 5.7x to 9.3x since the Q1 close, and disclosure, which has not improved in five quarters and is now the entirety of the residual bear case.

Action: hold existing positions and add only on weakness. This is no longer the asymmetric setup it was in May, when a proven business traded at a de-rated multiple with its chief overhang freshly resolved. The business has since gotten demonstrably better and the price has gotten commensurately higher, which leaves the risk/reward favorable but ordinary. The specific things that would take us to higher conviction are a full-year framework, any commercial metric on Unity AI, and evidence that Create's guided deceleration is conservatism rather than the start of a trend. The specific thing that would take us to Hold is two consecutive single-digit Create quarters alongside Vector settling below the mid-teens sequentially.