Unity AI Answers the Genie Bear: A Beat Above a Raised Guide, Vector's Fourth Straight Sequential Quarter, and a GAAP-Profit Date — Upgrading Unity to Outperform

Results vs. Consensus

A quarter ago we downgraded Unity to Hold. The downgrade was explicitly not a verdict on the business, which had just delivered the best operational quarter in our coverage, but on a risk profile that had widened underneath the numbers: a Google Genie world-model narrative that re-rated the multiple before the print, a flat Q1 guide that broke the beat-and-raise pattern, a withheld full-year framework, and a cluster of governance signals. We named the path back to Outperform in three gates. This quarter is the first read on whether management walked through them, and the short answer is that it walked through two of the three at speed, cleared a raised bar on the financials, and put a date on GAAP profitability that did not exist at the Q4 call. Resolving the Hold is the substance of this note.

Q1 2026 Scorecard

Year-Over-Year Comparison

The year-over-year picture is the cleanest in the coverage and a clear acceleration even off the Q4 print. Strategic revenue at +35% is the headline the company now wants the Street to anchor on, and it is the right one: it strips out the deliberately-shrinking IronSource Ad Network and shows the part of the business Unity is actually investing behind growing at more than twice the reported total rate. Strategic Grow at +49% is a faster year-over-year rate than the +11% total-Grow print last quarter precisely because the legacy ad network has been pulled out of the comparison rather than dragging it down. Strategic Create at +15% is the fourth consecutive quarter of mid-teens year-over-year growth, a consistency the CFO called "phenomenal" and which sits awkwardly against the bear thesis that the engine-subscription business is the one being disrupted. The 800-basis-point margin expansion is the operating-leverage story compounding: revenue accelerating at high flow-through margins while sales-and-marketing and G&A both fall in absolute dollars and as a percentage of revenue.

Quarter-Over-Quarter Comparison

The sequential move looks modest on the reported total ($508.2M vs. $503.1M, +1.0%) but the headline understates it for the same reason the Q4 guide looked flat: the IronSource Ad Network is still rolling off inside the total, masking the strategic acceleration underneath. On the line that matters, Vector held its ~15% sequential cadence for a fourth straight quarter, and the margin expanded another 200 basis points to 27% even as the company kept investing into AI-focused R&D (up 17% year-over-year inclusive of cloud inference and AI hiring). The sequential story is the optical flat-to-up total that the Q4 call warned about, sitting on top of a strategic business that is plainly compounding. What changed since February is that the market now has the strategic-revenue cut to see through the optics, and a Q2 guide that translates the underlying momentum into a reported +29–32% strategic growth rate.

Quality of Beat

Revenue: The $508.2M is operationally driven and, on the strategic cut, the highest-quality revenue print in the coverage. There is no FX windfall (Unity remains predominantly USD-billed), no acquisition contribution (IronSource is being wound down, not lapped), and no one-time license element. The upside-to-guide was once again Vector, which management characterizes as its highest-return revenue, and the disclosure that Vector is now 80% larger year-over-year and represents 80% of strategic Grow is the cleanest single statement of where the franchise's growth now lives. The reported total carries a shrinking ~$58M of nonstrategic Grow (the IronSource residual) that will be gone within a quarter; on the part of the business Unity is keeping, the growth rate is 35% and the quality is not in dispute.

Margins: The 27% adjusted EBITDA margin is the operating-leverage thesis printing at its highest level in over two years, and the mechanics are clean: revenue accelerating at high contribution margins, with adjusted sales-and-marketing and G&A both down year-over-year in dollars and as a share of revenue, and the savings deliberately redeployed into R&D (adjusted R&D +9%, AI-focused R&D +17%). The more important margin story is below the EBITDA line. Stock-based compensation fell 20% in dollars to 15% of revenue, roughly half its year-ago share, and M&A amortization is set to fall from $117M this quarter to sub-$25M for all of 2027. Those two declines are what convert a widening adjusted-EBITDA margin into a credible GAAP-profitability date, which management now places in Q4 2026.

EPS: Adjusted EPS of $0.23 is the one line that did not clean-beat: it sat $0.01 below the Street's higher $0.24 estimate while clearing the lower $0.20 estimate by three cents. We read the modest shortfall against the high estimate as immaterial next to the EBITDA beat and the cash story, and not as an operational tell; the gap is well within the noise of share-count and below-the-line modeling on a company whose adjusted profitability is rising fast. The cash line is the one that underpins the equity: free cash flow of $66M took the trailing-twelve-month figure to $463M, up 50% year-over-year, against a $2.15B cash balance that comfortably funds the $558M November 2026 convert from the balance sheet with no raise. The GAAP loss this quarter is real but is dominated by the $279M non-cash impairment tied to the IronSource Ad Network sunset and the Supersonic divestiture, a strategic clean-up cost rather than an operating deterioration.

Segment Performance

Unity reports two segments, and this quarter reframed how it reports them. Grow is the advertising and monetization business (the Unity Ad Network powered by Vector, plus the legacy IronSource-era ad products now being sunset), and Create is the engine-subscription business (Unity Pro / Enterprise plus the non-gaming "industry" vertical). The change is that, beginning Q2, all guidance is framed on strategic revenue, with the IronSource Ad Network and the Supersonic publishing business carved out as nonstrategic and run to zero. The effect is to surface a business growing 35% that the reported total (+17%) had been hiding behind a deliberately-shrinking legacy line.

Grow: Vector Now 80% of the Segment, IronSource Being Removed Entirely

Strategic Grow revenue of $278.7M was up 49% year-over-year, with the upside again driven by Vector, which management said grew ~15% sequentially for the fourth quarter in a row and is now 80% larger than a year ago. The composition shift the Q4 call flagged is now essentially complete: Vector is 80% of strategic Grow, and the IronSource Ad Network has gone from a slowly-declining 11% of Grow to a line the company is actively eliminating, with the network closed on April 30 and the Supersonic publishing business in a divestiture process. Management framed the decision as a pure quality upgrade, concentrating resources on the differentiated Vector platform and shedding commoditized, lower-margin ad-network revenue.

"Strategic Grow revenue in the first quarter was $279 million, representing 49% year-over-year growth. Revenue upside compared to our guidance and our expectations was once again driven by the exceptional performance of Vector. While our growth was impressive, it does not yet reflect any impact from our runtime data, which we believe will provide us with a sustainable competitive advantage for many years to come." — Jarrod Yahes, CFO

The contrast with last quarter could not be sharper. Three months ago the Q1 Grow guide was flat sequentially, and that flatness, a mechanical artifact of a shrinking IronSource line and a holiday-rich base, did real damage to a beat-and-raise-conditioned stock. This quarter the company removed the source of the optics altogether: by carving IronSource and Supersonic into a defined nonstrategic bucket (~$50M in Q2, ~$45M in Q3, then gone), the reported Grow line will increasingly equal the Vector-driven strategic line. The Q2 guide makes the point numerically, with strategic Grow guided to +50–52% year-over-year.

Assessment: The flat-Q1 fear that anchored the Q4 sell-off is resolved, and resolved emphatically. Vector did not merely hold its cadence; it overwhelmed the legacy drag to the point that the company chose to delete the drag entirely rather than wait it out. The "return to sequential growth in Q2" that management promised in February is now visible in a +50–52% strategic Grow guide, and the decision to define and shed nonstrategic revenue is the clean structural answer to the disclosure complaint we lodged last quarter. The single most important near-term tell from the Q4 commitments, whether the flat Q1 proved to be the optical trough, grades cleanly delivered.

Create: Fourth Straight Mid-Teens Quarter, the Segment the Bear Case Targets Keeps Accelerating

Strategic Create revenue of $153.7M was up 15% year-over-year, the fourth consecutive quarter of mid-teens growth and, in the CFO's words, a "phenomenal" run of consistency. The drivers are the same three that have compounded all year: the impact of annual price increases, continued strength in China, and a non-gaming "industry" business where Unity is "really a leader in auto HMI." Management also tied the segment's stability to a multi-year quality push, citing a 22% decline in user-reported issues since the Unity 6 launch and a maintained ~70% share of mobile game creation that allows it to pass through moderate price increases while reinvesting in product.

"In Create, strategic revenue was $154 million, up 15% year-over-year. The consistency of this business over the last year has been phenomenal with 4 straight quarters of mid-teens year-over-year growth… This has allowed us to maintain a robust 70% market share in mobile game creation while passing along moderate price increases and allowing us to invest aggressively in our products." — Jarrod Yahes, CFO

The forward framing has a wrinkle worth flagging: the Q2 strategic Create guide of +11–14% year-over-year explicitly excludes the impact of a large customer win the company is comping from 2025, so the underlying rate is steadier than the optically-decelerating guide implies. Beneath the headline, management was explicit that the Create business model is itself evolving, with the new Unity AI pricing scaling on first- and third-party agent connections in addition to seats, so that revenue can grow with usage as agentic and human consumption rises rather than being capped by a per-seat model.

Assessment: Create remains the underappreciated leg, and once again it strengthened on the very dimension the world-model narrative attacks. A subscription franchise growing 15% for a fourth straight quarter, holding 70% of mobile creation, with a fresh pricing model designed to monetize AI usage rather than be displaced by it, does not look like a business being disrupted in real time. The 70% share figure is the new disclosure that matters most here: it quantifies the installed-base moat that the Genie thesis has to overcome, and it is the empirical counterweight to the abstraction that world models make engines obsolete. We continue to underwrite Create at a low-double-digit-plus strategic rate and note, again, the irony that the segment carrying the most narrative risk has the steadiest printed trajectory in the company.

Industry Vertical and China: The Non-Gaming Engine Inside Create

Management spent more time than usual on the non-gaming side of Create, which it now calls out as a distinct growth driver. The "industry" business (auto, manufacturing, and other enterprises using Unity for interactive 3D content across platforms) was described as growing "very strongly," with Unity positioned as a leader in automotive human-machine interface (HMI) work, the in-dash and cockpit display software that ships in modern vehicles. Consistent with prior quarters, management declined to break out the dollar contribution, but the qualitative emphasis was heavier than in Q4.

"Our industry business outside of gaming continues to grow very strongly. We're really a leader in auto HMI. There's a range of use cases where people are looking to have interactive content for extremely sophisticated models available across a range of platforms and operating systems, which really plays to Unity's strength." — Jarrod Yahes, CFO

Assessment: The industry vertical is a genuine and under-modeled diversifier inside Create, and one with a structurally different demand driver (enterprise digital-twin and HMI adoption) than the gaming cycle. It is not yet a sized revenue thesis, and management is right not to pitch it as one, but the recurring callout and the auto-HMI leadership claim support the view that Create's growth is broader-based than a pure gaming-engine read would suggest. We treat it as a quality marker for the segment's durability rather than a modeled line.

Key Operating Metrics

Key Topics & Management Commentary

Overall Management Tone: The posture reversed from last quarter. Where Q4 was defensive, opened by a CEO framing the call around tuning out a "noisy" environment, Q1 was forward-leaning and product-led, opened with "Unity is on an incredible trajectory" and built around shipping rather than rebutting. Management was most confident on Vector, the margin trajectory, and the GAAP-profit date, where the numbers carry the argument, and notably chose to engage the world-model question on the offensive, with a launched product rather than a logical paragraph. The one place the tone stayed disciplined-to-evasive was on quantification: management again declined to size runtime-data contribution or break out Vector dollars, the industry vertical, or the commerce take rate, preferring framework-level answers. The net is a team that walked into this call with the operational and product wind at its back and used it.

1. Unity AI: Turning World-Model Output Into Engine Input

The most consequential disclosure of the quarter is a product, not a number. Unity AI, the first Vector-driven enhancement for the Unity engine itself, went into public beta the week of the call. It is an integrated agent tuned specifically for Unity, with full project context, that writes code directly into a Unity project across scene hierarchies, packages, assets and performance controls. The strategically critical capability is what it does with generative world-model output: it ingests image pixels, outputs primitives, performs mesh up-scaling and textures, and constructs the content pipeline almost instantly, converting the raw material a Genie-class model produces into a production-ready Unity scene.

"We've remarked on many occasions that generic world models would be a great source of prototyping material and that game engines like Unity could ingest those pixels, transform them into Unity scene and enable developers to build deep interactive content and systems around them. Well, that's just what Unity AI now does. It ingests image pixels, outputs primitives and does mesh up scaling and textures, constructing the entire content pipeline almost instantly." — Matthew Bromberg, CEO

This reframes the entire Genie debate that drove the Q4 downgrade. For two quarters management argued world models were "complementary, not duplicative" as a thesis; this quarter it shipped the interface that makes the argument literal, positioning Unity as the consumption-and-assembly layer that sits downstream of any world model rather than the engine those models replace. The CEO was explicit that the "contest" framing misses what is happening: the future combines what generative systems do best (speed, personalization, variation at scale) with what engines do best (determinism, persistence, consistency), and that the bottleneck shifts from generation, "a commodity if it hasn't already," to direction and realization, which is a tools problem Unity has spent 20 years solving.

Assessment: This is the single most important development for the thesis and the reason the rating moves. At the Q4 downgrade we wrote that management's rebuttal was "coherent and more right than wrong" but "unfalsifiable on any near-term horizon," and that an unsizable tail on Create was incompatible with an Outperform. Unity AI does not make the tail disappear, but it changes its character from passive (defend the engine and hope) to active (architect the engine as the layer every world model feeds into). A ~70% day-five retention rate on the public beta, one week in, is early but real evidence of product-market fit. The Genie bear is no longer a narrative Unity is absorbing; it is a narrative Unity is now monetizing the answer to. We are upgrading the bear pillar from EMERGING to CONTAINED on the strength of this, with the caveat that the proof remains forward (adoption, attach, and the rest of the 2026 product pipeline must land).

2. Vector: The Flywheel, the Fourth Straight Quarter, and the 28-Day ROAS Unlock

Vector posted its fourth consecutive quarter of ~15% sequential growth and is now 80% larger than a year ago, exceeding the company's own ambitious expectations. The CEO walked through the mechanics with unusual specificity, describing a flywheel of product improvement and signal improvement: better product (a new day-28 return-on-ad-spend planning tool versus the prior day-7 horizon) drives more advertiser spend, while better signal lets Unity retrain its self-learning models, and the loop compounds. The day-28 ROAS product was singled out as a recent unlock, driving an 80% incremental campaign lift and a 37% ROAS improvement versus the old day-7 benchmark.

"We rolled out last quarter a day 28 ROAS product, which enables our customers to plan and predict their return on a 28-day basis instead of just a 7-day basis. And it's driven an incremental lift of 80% for campaigns and showing a 37% ROAS improvement versus the old day 7 benchmarks… better signal, better models, better ROAS, better product, more ad spend and you begin again. That's the flywheel of our business." — Matthew Bromberg, CEO

Management framed the growth as broad-based across geographies, campaign types, platforms and genres, with no geographic divergence to call out, and reiterated that none of it yet reflects runtime data. On the question of where the ceiling is, the CEO's answer was that industry conversion rates sit in the single digits, so the headroom for improvement "is almost infinite," which is the same no-natural-ceiling framing from last quarter but now backed by a fourth consecutive quarter of delivery.

Assessment: Vector's durability, the Q4 commitment we most wanted confirmed, graded cleanly delivered and then some. A fourth straight ~15% sequential quarter with +80% year-over-year growth, a concrete product unlock (day-28 ROAS) quantified at an 80% campaign lift, and the entire trajectory still predating runtime data is a stronger durability argument than the +72% January figure that anchored the Q4 call. The "single-digit industry conversion rate" framing is the most useful new way to size the runway: if Vector's edge is conversion prediction and the industry baseline is low-single-digit, the improvement headroom is structural rather than cyclical. This is the load-bearing bull pillar, and it is intact and strengthening.

3. The GAAP-Profitability Date: Q4 2026, on Three Concrete Levers

Management did what the Street had been asking for and put a date on GAAP profitability: Q4 2026. Critically, it was framed not as an aspiration but as the arithmetic of three converging levers. First, adjusted EBITDA margins up 800 basis points. Second, stock-based compensation down 20% in dollars and to 15% of revenue, "literally cut in half as a percentage of revenue year-over-year." Third, M&A amortization running off almost entirely by end-2026, from $117M this quarter to $80M in each of Q2 and Q3, $55M in Q4, and sub-$25M for all of 2027.

"As a result, we now forecast Unity to be GAAP net income profitable by the fourth quarter of 2026… You should expect M&A amortization to fall off quite precipitously this year. From the first quarter where we experienced $117 million of amortization, we would expect $80 million of amortization in each of Q2 and Q3, dropping down to $55 million in the fourth quarter. And in calendar year 2027, that amortization for the full year should be sub-$25 million." — Jarrod Yahes, CFO

This is the disclosure that most directly addresses the "no anchor" complaint from Q4. Management still declined to give a full-year revenue range, but it gave the out-year a different and arguably more durable anchor: a hard profitability milestone with the bridge laid out line by line. For a name whose adjusted-to-GAAP gap has long been the bear's favorite stick, watching SBC halve and amortization run off on a disclosed schedule is the cleanest path-to-profitability framing Unity has provided.

Assessment: The GAAP-profit date is a high-quality substitute for the full-year revenue framework the Street wanted and did not get. A revenue range would have been an estimate; the profitability bridge is mostly mechanical (amortization and SBC schedules are far more predictable than top-line), which makes the Q4 2026 milestone more credible than a revenue guide would have been. It does not fully close the forward-visibility gap (Bear-5), but it meaningfully narrows it, and it gives the equity a concrete, datable catalyst on the path to a clean GAAP profit. We weight it as the second-most-important disclosure of the quarter after Unity AI.

4. Portfolio Surgery: IronSource Ad Network Sunset, Supersonic Divestiture, and the $279M Charge

Unity took the knife to its non-core portfolio. It sunset the IronSource Ad Network (closed April 30) and launched a strategic process to divest the Supersonic publishing business, the two actions that drove the $279M non-cash impairment in the GAAP results. Management framed both as deliberate quality and focus decisions: shedding commoditized ad-network revenue and a lower-margin publishing operation to concentrate on the strategic platform. The CFO quantified a tangible margin benefit, at least 200 basis points of operating-profit improvement upon the Supersonic divestiture, on top of the cost reductions that flow through as the IronSource cost base leaves the system in the back half of the year.

"We are currently in the midst of the strategic process for Supersonic. The profitability of Supersonic is such that as we divest the business, that will naturally cause our margin profile to improve. We expect at least 200 basis points of operating profit improvement upon the divestiture of that business." — Jarrod Yahes, CFO

The flip side, which the bears will note, is that the Q2 numbers carry the revenue loss of these actions before the cost savings land. The CFO was explicit that Q2 absorbs one fewer-than-normal month of IronSource Ad Network revenue with the costs still in the system, creating a quarter of operating deleverage, before the costs come out in the back half and the margin benefit shows up.

Assessment: The portfolio surgery is the right call and a credible margin story, and it is the structural fix to the disclosure complaint from Q4: by defining and shedding nonstrategic revenue, management has removed the very ambiguity that made the Q4 Grow optics look flat. The $279M impairment is an accounting acknowledgment of past M&A, not an operating problem, and the disclosed ~200bps Supersonic uplift plus the IronSource cost run-off give the back-half-2026 margin expansion a concrete basis. The near-term cost is the optics: Q2 takes the revenue hit before the cost benefit, which the guide already reflects. We treat this as net positive, with the timing mismatch a one-quarter cost worth paying.

5. Runtime Data: Going Live in Vector This Quarter, Still Not Quantified

The runtime-data catalyst, feeding proprietary sequential behavioral signal from Made-with-Unity games into Vector, advanced from "live in Q2" framing to an explicit go-live during the current quarter, with the CEO reiterating positive offline test results and the >90% Developer Data Framework opt-in. The strategic argument got sharper: runtime data is real-time and sequential, "not click data, not conversion data, not post-back data," a different category of signal that captures the order in which players do things, which management believes is a powerful predictor of conversion. Consistent with every prior quarter, the CEO declined to promise a step-change, framing the benefit as compounding rather than a "lightning strike."

"What we believe is really quite valuable about runtime data is that it's real time, so it's not delayed. It is sequential… It's not click data, it's not conversion data, it's not post-back data. It's a different category of signal… we expect to graduate our testing to live production models during the course of Q2." — Matthew Bromberg, CEO

Assessment: This grades the Q4 commitment on runtime-data timing as on track: the go-live holds for this quarter, the opt-in remains above 90%, and the sequencing argument is the clearest articulation yet of why the signal is differentiated. Management again refused to quantify the magnitude, which remains appropriately disciplined for an unproven feature but is the one place the disclosure is still thin. The key analytical point is unchanged and bullish: Vector is already growing 80% year-over-year with none of this in the model, so runtime data is upside to a thesis that already works, and it is precisely the proprietary behavioral moat that no world model can replicate. We weight the firm go-live as a modest positive and the refusal to quantify as prudence rather than evasion.

6. The Business Model Evolves: Consumption Pricing for an Agentic World

Management spent meaningful airtime on how Unity's pricing must change as AI moves to the center of the platform. The principle articulated was that pricing should scale with usage and value created rather than penalizing customers for being more productive, because "customers value the outputs, not the inputs." The new Unity AI pricing already reflects this, scaling on first- and third-party agent connections in addition to seats, and the existing enterprise minimum-commit model is described as well-suited to a consumption overlay. The CFO framed this as opening "new areas of growth like consumption into our models," with both human and agent users interacting with the software concurrently.

"The new pricing models we've introduced with Unity AI, which account for the number of first- and third-party agent connections in addition to seats, ensures our pricing scales fairly with usage rather than penalizing creators who use AI to be more productive. Over time, we expect this will allow us to scale revenues from both agentic and human consumption." — Jarrod Yahes, CFO

Assessment: This is the commercial complement to Unity AI's technical story, and it is strategically important: it is how a per-seat subscription business avoids being deflated by AI productivity gains, by attaching price to agent connections and consumption rather than headcount. It is early and unquantified, and the transition from seat-based to consumption-based pricing carries execution and customer-acceptance risk. But the direction is right, and it directly addresses the most sophisticated version of the AI-disruption bear, that AI collapses the value of seats. We treat it as credible strategic direction and a positive-skew option on Create monetization, with the burden of proof ahead.

7. Unity Commerce: Launching This Quarter With Committed Partners

The cross-platform commerce product (previously framed as Unity IAP) is on track to launch this quarter, now with named committed partners including Voodoo and SciPlay. The pitch is a single native dashboard to manage catalogs, pricing and payouts across mobile, web and PC, removing the overhead of juggling multiple SDKs and payment systems. Management again framed the value to Unity as multi-dimensional, data, economic revenue share, and merchandising/web-shop optimization, and again declined to size it, but the move from "GA in Q2" to a launch with marquee partners is incremental progress.

"The Unity commerce platform is also on track to launch this quarter, and we already have a set of committed partners like Voodoo games and SciPlay working with us to ensure we do it right. Developers shouldn't have to be fintech experts to run a global business." — Jarrod Yahes, CFO

Assessment: Commerce remains genuine optionality with a clear secular tailwind (app-store disintermediation) and a structural advantage (engine-native integration), and the named-partner progress is a modest de-risking of the launch. The explicit link back to enriching Vector's behavioral picture is the strategically interesting part, transaction data feeds the same personalization engine. It is still not a revenue thesis and management is right not to pitch it as one. We carry it as a free option on the commerce layer and on incremental Vector signal, not a modeled line.

8. Capital Allocation: $2.15B Cash, the November Convert, and a High M&A Bar

The balance sheet continues to strengthen. Cash rose to $2.15B on the back of $66M of quarterly free cash flow ($463M trailing twelve months, +50% year-over-year), and management's stated intention is to pay off the $558M convertible note due November 2026 from cash on the balance sheet, reducing leverage rather than refinancing. On uses of cash, the CFO set a deliberately high bar for M&A, emphasizing organic investment in the product roadmap and a reluctance to distract the product teams, which reads as a direct response to the IronSource/Supersonic experience the company is now unwinding.

"Right now, we are planning on paying off the '26 convert that's coming due in November… we're really focused on our product road map, organic investment in our business… there is a high threshold as we evaluate M&A opportunities. So I think you'll see us be prudent with our cash, not distract our product teams." — Jarrod Yahes, CFO

Assessment: The convert overhang we tracked through the prior three quarters is now fully a non-issue: paying off the $558M November note from a $2.15B cash balance, with $463M of annual free cash flow behind it, removes the last of the refinancing-and-dilution concern. The high M&A bar is the right posture and a quiet acknowledgment that the IronSource-era acquisitions added complexity the company is now shedding; capital discipline after a write-down is the credible signal. The balance sheet is not where the risk lives, and capital allocation is now a source of support rather than a swing factor.

Guidance & Outlook

The guide is the cleanest contrast with the Q4 setup, and it cuts the opposite way. A quarter ago the Q1 guide was flat sequentially and below consensus, breaking the beat-and-raise pattern; this quarter the Q2 guide pairs an optically-modest total ($510M midpoint, held down by the deliberate IronSource runoff) with a strategic-revenue guide of $455–465M (+29–32% year-over-year) and an adjusted EBITDA guide of $130–135M (+44–49% year-over-year) that sits above the ~$131M consensus. The composition is a +50–52% strategic Grow line driven by Vector, a +11–14% strategic Create line (steadier than it looks once a 2025 large-customer comp is excluded), and a defined nonstrategic bucket running to zero. Management framed the second half as the period when the strategic-action cost savings land and margins reach "record levels."

The qualitative framing was confident on strategic growth, the margin trajectory, and the GAAP-profit milestone, and still silent on full-year revenue. The CFO described two known back-half margin catalysts beyond normal operating leverage: the IronSource cost base leaving the system, and the ~200bps Supersonic divestiture uplift. The one acknowledged near-term wrinkle is the Q2 margin optics: with one fewer month of IronSource Ad Network revenue and its costs still in the system, Q2 carries some operating deleverage before the back-half recovery, which is why the $132.5M EBITDA midpoint implies a slightly lower margin than Q1's 27% before re-expanding.

Implied sequential setup: The $510M total midpoint is roughly flat off Q1's $508.2M optically, but that is the IronSource runoff at work; on strategic revenue the +29–32% guide implies a clean sequential step. Given Vector's demonstrated pattern of exceeding internal expectations for four straight quarters, the realistic path likely lands at or above the strategic midpoint.

Street at: Consensus for Q2 sat near the $510M total and ~$131M EBITDA; the guide is in line on revenue and modestly above on EBITDA, a constructive setup rather than the below-consensus guide that defined Q4.

Guidance style: Confident and specific on strategic revenue and EBITDA, deliberately conservative on the reported total (which absorbs the IronSource runoff), disciplined on the GAAP-profit bridge, and still absent on full-year revenue. The pivot to strategic-revenue guidance is itself the structural disclosure improvement we asked for last quarter; the missing full-year top-line range is the one remaining gap.

Analyst Q&A Highlights

Where Unity AI Sits Against the Frontier Models Making Games

The opening exchange went straight to the competitive heart of the quarter: with frontier-model labs demoing game creation as a use case, where does Unity AI fit in a world proliferating with general-purpose tools that can make some kind of game? Management's answer leaned entirely on context and specificity, arguing that general-purpose coding agents lack Unity-engine context and project context, the gap that matters when shipping a real game across platforms, and disclosed the early attachment data as evidence the differentiation is landing.

Q: "When you think about how that compares to what you can do with Unity AI and your ability to get in front of the next generation of game creators… given that there's going to be this proliferation of other tools that can make some sort of games. I guess, where do you feel you sit in the future of these new AI game creation tools?"
— Matthew Cost, Morgan Stanley

A: "General purpose coding agents are really powerful, but what they lack is Unity engine-specific context as well as the context of the project you're building itself. And that gap matters a lot when shipping a real game across multiple platforms… we're seeing really strong attachment rates in that product. So 70% of the users who adopt it are continuing to work with the product 5 days in… This kind of attachment appears to be a function of better performance of our AI product than just generic models alone."
— Matthew Bromberg, CEO

Assessment: This is the exchange that most directly retires the Genie bear, and it does so with data rather than rhetoric. The ~70% day-five retention figure is the first quantified evidence that Unity's context-engineering advantage (knowing both the engine and the project) produces a stickier product than generic models, and the CEO's framing of Vector as the personalization layer now powering both Grow and Create ties the whole AI story together. It does not prove durability, one week is one week, but it converts an abstract "we'll be fine" into a measurable early signal, which is exactly what the rating upgrade needed.

Why Runtime-Data Sequencing Is a Different Category of Signal

A multi-part question pressed for granularity on why runtime data is powerful, surfacing the "sequencing" idea that understanding the order of a user's in-app actions is a strong predictor of conversion. Management embraced the framing and extended it, distinguishing runtime data from the click, conversion and post-back data the ad industry already uses, and tying its value to its real-time, sequential, in-engine nature.

Q: "I think people are trying to unpack like why the runtime data is powerful… this idea of sequencing keeps coming up… why do you think sequencing is going to be such a powerful avenue of conversion prediction in the mobile game and app space?"
— Alec Brondolo, Wells Fargo

A: "It is sequential, which means it helps us to understand not just what consumers are doing, but… the order in which you do things is critical to understanding what it is that you're doing… It's not click data, it's not conversion data, it's not post-back data. It's a different category of signal. And we're excited about the impact we think it will have."
— Matthew Bromberg, CEO

Assessment: The exchange is the clearest articulation to date of the runtime-data moat, and it matters because it grounds an abstract "proprietary signal" claim in a concrete mechanism: sequence-aware, in-engine behavioral data that the ad-tech incumbents structurally cannot see because they do not run the game. Management still would not quantify the contribution, but the "different category of signal" framing is a credible explanation of why this is differentiation rather than incremental data volume. For modeling, it reinforces treating runtime data as upside optionality with a Q2 go-live, not a sized line.

Composition of Strategic Grow Growth and the Margin Levers Beyond Vector

A two-part question sought to decompose the >50% strategic Grow growth between model enhancements and self-learning, and separately to identify margin levers beyond the natural leverage from Vector's growth. On composition, management declined to delineate the individual contributions but confirmed all three axes (model, product, signal) are contributing with runtime data still to come. On margins, the CFO named two specific back-half catalysts on top of ongoing automation-driven leverage.

Q: "Based on the guide for strategic grow revenue… Is there a way that you can help sort of frame up the composition of growth between model enhancements and self-learning?… [and] additional margin levers that you have at your disposal… are there any other places where you see room to continue slimming down the cost structure?"
— James Heaney, Jefferies

A: "80% of that strategic grow revenue line is now the Unity Ad Network powered by Unity Vector… We don't break out or delineate the individual contributions… [On margins:] with respect to some of the strategic actions… the costs haven't come out of the business as of yet. That will take place over the back half of the year… We expect at least 200 basis points of operating profit improvement upon the divestiture of [Supersonic]."
— Matthew Bromberg, CEO / Jarrod Yahes, CFO

Assessment: The useful disclosure here is the 80%-of-strategic-Grow-is-Vector figure, which quantifies how thoroughly the segment has become a Vector story, and the two named margin catalysts (IronSource cost run-off and the ~200bps Supersonic uplift) that give the back-half margin expansion a concrete basis rather than a hand-wave. The refusal to decompose model-versus-signal growth is consistent and defensible; the components are not separately observable and breaking them out would invite quarter-to-quarter noise. The exchange supports both the revenue durability and the margin-expansion legs of the thesis.

The Create Commercial Opportunity From AI-Lowered Creation Barriers

A question connected the unusually positive release-volume and sign-up callouts to the Create opportunity, asking how AI-lowered creation barriers change the commercial opportunity for both professionals and a new prosumer class. The CEO's answer was the offensive case for Create under AI: tools that automate the common, repetitive systems work let professionals reach the creative differentiation faster, while a new and ultimately much larger prosumer class is newly enabled to build interactive entertainment.

Q: "It feels like browser AI tools, a lot of the work on the product side that you're doing with Create is closing some of the gaps… how do you guys see this sort of collectively changing the commercial opportunity for the Create business as it relates to… both professionals and hobbyists?"
— Clark Lampen, BTIG

A: "Tools that enable more efficiency, allow you to get to that creative head end much faster and will ultimately result in better games… But we're equally excited about the creation of a new class of creators that ultimately is going to be much, much larger than the professional class… all those creators will be adding interactive elements… because that is, by far, the most direct way to increase engagement."
— Matthew Bromberg, CEO

Assessment: This is the strategic vision that underpins the AI-as-tailwind thesis: AI does not shrink Create's market, it expands it, both by accelerating professionals and by enlarging the creator base toward "the many." It is a vision, not a number, and the prosumer-TAM expansion is unquantified and unproven. But it is internally coherent with the printed evidence the CEO cited (new sign-ups +20% quarter-over-quarter, the fastest since 2020; Made-with-Unity games +12% sequentially; new mobile apps +60% year-over-year), which is the first quarter in a while those leading-indicator metrics have been positive enough to volunteer. We weight it as credible direction with the prosumer monetization still to be proven.

Cloud-Cost Trajectory and Whether a Minimum Contribution Margin Is Defended

A question probed the tension between investing in AI/runtime experimentation and protecting contribution margin, asking whether Unity defends a minimum gross-margin floor. The CEO declined to commit to a specific floor, framing the choice explicitly as a preference for accelerating revenue growth over optimizing cloud costs in the current investment phase, while pointing to a structural long-term opportunity to bring cloud costs down toward larger competitors' levels.

Q: "I wanted to know how you think about the evolution of cloud costs and your contribution margin in the context of that and whether you want to maintain some minimum kind of contribution margin above which investments would not exceed."
— Omar Dessouky, Bank of America

A: "We've experienced 82%, 83% adjusted gross margins, which are inclusive of our cloud costs… If given a preference of optimizing cloud costs versus accelerating revenue growth in the future, we would tilt towards accelerating revenue growth… with the knowledge that we will optimize cloud costs over time as the business scales and grows."
— Matthew Bromberg, CEO

Assessment: The honest read is that management is explicitly choosing growth over near-term gross-margin optimization, which is the right call at Vector's current return profile but is also the answer that lets margin "bounce around" quarter to quarter on investment timing. The 82–83% adjusted gross margin and the claim that cloud costs are structurally higher than larger competitors' (and therefore a future leverage source) are both useful disclosures. The lack of a defended floor is a modest watch item, the same OpEx/cloud-investment discipline question that recurs for any scaling AI platform, but it is bounded by the 27% EBITDA print and the GAAP-profit commitment, which together cap how loose the investment can get.

The Q2 Margin Step-Down and the IronSource Cost Timing

A pointed follow-up flagged that the Q2 margin guide steps down slightly and asked whether the cause was operating expense, contribution margin, or the runtime-data experimentation load. Management was direct that the larger driver is the strategic-action timing, two fewer months of IronSource Ad Network revenue with the costs still in the system, with continued cloud investment a secondary factor.

Q: "If I'm looking at the second quarter, it looks to me like your margins are guided down a little bit. Does that come from operating expenses or from this kind of contribution gross margin leverage…? Is that the driver of the slightly lower margin?"
— Omar Dessouky, Bank of America

A: "The larger contribution is the fact that we have taken some strategic actions, which are bringing down total revenue in the second quarter. There's 2 fewer months of the ironSource Ad Network, the cost of which are still in the system in the second quarter, and we would expect to leave the system in the back half of the year. So it's that operating deleverage that hits us in the second quarter, but we know we're going to get those costs out in the back half."
— Matthew Bromberg, CEO

Assessment: This is the most useful modeling exchange on the margin line and the antidote to any read of the Q2 step-down as a deterioration. The slightly lower Q2 margin is a timing artifact of revenue leaving before its associated costs do, not contribution-margin erosion, and management was explicit the costs come out in the back half, which is when it expects record margins. For modeling, it argues Q2 is the optical margin trough of the IronSource wind-down (mirroring how Q1 2026 was the optical revenue trough), with H2 stepping up as the cost base normalizes. It is the margin-side analogue of the Grow-optics story and reinforces underwriting the back half above the Q2 print.

What They're NOT Saying

1. Still no full-year 2026 revenue or segment framework. Management put a hard date on GAAP profitability (Q4 2026) and a detailed amortization schedule on the table, but provided no full-year strategic-revenue range, no Vector dollar target beyond the qualitative trajectory, and no annual segment outlook. The GAAP-profit anchor is a real and arguably more durable substitute, but the top-line range the Street most wanted is still withheld.
2. Runtime-data contribution remains unquantified. The go-live moved to "during Q2" with >90% opt-in, but management again declined to size the revenue impact or even bracket it, repeating the "compounding, not a lightning strike" framing for a fourth straight quarter. Disciplined for an unproven feature, but it means the single most-hyped catalyst still cannot be modeled.
3. Vector is still not disclosed in dollars. Management quantified Vector's year-over-year (+80%), its share of strategic Grow (80%), and its sequential cadence (+15%, fourth straight), but the absolute dollar level remains undisclosed, so the Street still cannot independently track the Vector-versus-everything-else split on which the thesis turns.
4. Unity AI adoption is one data point deep. The ~70% day-five retention figure is encouraging but is a single, one-week-old metric on a public beta. There is no disclosure of beta user count, conversion to paid, or the pricing realized on agent connections, so the commercial trajectory of the most strategically important new product is still essentially unsized.
5. The "pipeline of new products before end-2026" is unspecified. The CEO referenced an "AI native" product pipeline scheduled before the end of 2026 with the potential to "redefine our industry," but named none of it beyond Unity AI and commerce. Intriguing as a forward signal, but impossible to value without specifics.
6. Create's large-customer comp was flagged but not sized. The Q2 strategic Create guide excludes "a large customer win we're comping from 2025," which implies the reported rate is being depressed by a known 2025 contribution, but management did not quantify the comp, leaving the underlying Create growth rate slightly ambiguous.
7. Commerce economics still unquantified, a quarter from launch. Unity Commerce launches this quarter with named partners, but with no take-rate figure, no addressable-volume target, and no contribution timeline, consistent with prior quarters but now at the doorstep of general availability.

Market Reaction

• Pre-print setup (May 6, 2026 close): $27.28, entering the print down 38.2% year-to-date (from $44.17 at 2025 year-end) but up 32.0% over the trailing twelve months (from $20.66) and up 23.9% over the trailing 30 days (from $22.01). The 52-week closing range was $17.13–$49.47. The critical context is that the stock entered this print already recovering off its lows (up nearly 24% in a month) but still deeply down on the year, a setup that combines washed-out, heavily-shorted positioning with a fragile, sentiment-driven base.
• Reaction session (May 7, 2026, reported before the open): The stock gapped up to open at $29.86 (+9.5%) and traded an intraday range of $26.33 to $30.65 (−3.5% to +12.4%), before fading through the session to close at $26.73, down 2.0% (−$0.55) on the day. The print produced a sharp early pop on the headline beat-above-guide and the GAAP-profit date; the pop did not hold.
• Volume: 32.8M shares traded versus a 13.8M 30-day average, a 2.4x surge, consistent with active two-way repositioning into the gap rather than a one-directional stampede.
• Benchmark context: The S&P 500 closed down 0.4% on the reaction day and was up 7.6% year-to-date, so Unity's intraday fade was idiosyncratic, not a market event, though the broadly flat tape removed any index tailwind from the session.

The price action is a classic gap-and-fade and, in our read, more a function of positioning than of the print. A name down 38% year-to-date, heavily shorted, and already up ~24% in the trailing month gapped 9.5% on a clean beat, and the move attracted profit-taking and short-covering-then-re-shorting that bled the gain to a 2% loss by the close. The fundamentals that drove the early pop, revenue above a raised guide, a 27% margin, strategic revenue +35%, Unity AI, and a hard GAAP-profit date, are unambiguously the strongest set in the coverage; the tape simply did not reward them on day one. This is the inverse problem from Q4, where a great print met a bad setup and fell 26%; here a great print met a recovering-but-fragile setup and gave back an early spike. We weight the print and the resolved overhang over a single session's positioning-driven fade, and note that a 2% down day on this print, against a Q4 that saw a 26% down day on a comparable beat, is itself evidence the narrative has shifted.

Street Perspective

Debate: Does Unity AI Actually Neutralize the World-Model Threat?

Bull view: The bull case on the Street is that Unity AI fundamentally reframes the Genie debate by making world-model output an input to the engine rather than a substitute for it: the product ingests generative pixels and converts them into editable Unity scenes, positioning Unity as the indispensable assembly-and-distribution layer downstream of any generative model. The ~70% day-five retention on the public beta is offered as early proof the context-engineering advantage produces a genuinely better tool than general-purpose agents, and the accelerating Create print (+15% for a fourth straight quarter) as evidence the disruption was always theoretical.

Bear view: The bear camp contends that a one-week-old beta with a single retention data point proves very little about a multi-year displacement risk, that incumbents always build the "AI-native" product that supposedly co-opts the disruptor, and that if generation truly becomes a commodity, the value migrates to whoever owns the generative model and the distribution surface, not necessarily to the engine in the middle. On this read, Unity AI is a necessary defensive move that does not by itself prove the engine layer captures the economics in an AI-native creation stack.

Our take: The bull has materially the better of this debate now than at Q4, which is exactly why the rating moves. Unity AI converts the "complementary, not duplicative" thesis from an assertion into a shipped product with an early, measurable attach signal, and the architecture (world model as upstream content source, engine as downstream systems-and-distribution layer) is the most credible articulation yet of where Unity's durable value sits. The bear's caution about a single data point is fair, and we hold the pillar at CONTAINED rather than RESOLVED for that reason. But an unsizable existential tail has become a sizable, addressable product opportunity with early traction, and that is a different risk profile than the one that forced the downgrade.

Debate: Is the Strategic-Revenue Pivot Clarity or Goalpost-Moving?

Bull view: The bull view is that the pivot to strategic revenue is exactly the disclosure improvement the Street demanded: by defining the IronSource Ad Network and Supersonic as nonstrategic and running them to zero, management surfaces a 35%-growth business that the reported +17% total had been masking, and removes the optical drag that made the Q4 Grow guide look flat. The accompanying ~200bps Supersonic margin uplift and the cost run-off give the framing a concrete financial benefit, not just a presentational one.

Bear view: The bear camp counters that re-centering the narrative on a hand-picked "strategic" subset is the oldest trick in the guidance playbook, that the reported total still grew only 17% and will be optically flat sequentially in Q2, and that excluding both a legacy ad network and a "large customer comp" from the growth framing risks flattering the rate the Street anchors on. Combined with the still-withheld full-year revenue range, the bears argue the disclosure is being curated rather than expanded.

Our take: We side with the bull on substance while taking the bear's discipline point. Carving out a deliberately-shed, commoditized, lower-margin ad network is a legitimate clarification, not a cosmetic one, because that revenue is genuinely going away and is genuinely lower-quality; this is not the same as excluding a struggling core line. The strategic cut is the more honest representation of the business Unity is actually building. The bear's fair point is the still-absent full-year top-line range, which keeps a sliver of the Q4 forward-visibility complaint alive, and we keep the governance/disclosure pillar at EMERGING (improving) rather than declaring it resolved.

Debate: At a Recovering ~5x Sales, Is the Re-Rate Beginning or Capped?

Bull view: The bull view is that the washed-out ~4x trailing-sales multiple of February is beginning to normalize as the overhangs lift: with the Genie bear answered by a shipped product, Grow's return-to-growth proven, a 27% margin, a GAAP-profit date, and the convert paid from cash, the case that took the stock to its lows has been systematically dismantled, and a de-rated, heavily-shorted name re-rates violently when the bear thesis breaks. The early 9.5% gap is offered as a preview of the repricing once positioning resets.

Bear view: The bear camp notes that the gap faded to a loss the same session, that a 2% down day on a clean beat is not the tape of a stock about to re-rate, and that the multiple compressed on a real, if reframed, long-duration risk that one quarter of product progress does not fully retire. On this read, "cheap with improving fundamentals" can persist for several quarters before the market pays up, and the absence of a full-year framework caps how much conviction the out-years can carry.

Our take: This is the crux of the rating, and it is why the call is Outperform rather than a cautious Hold-with-upside. The de-rating thesis has been dismantled gate by gate: the one live valuation risk we carried into Q4 (Bear-2) has eased from MATERIALIZING toward a business demonstrably re-accelerating into the multiple, and the new disruption tail (Bear-4) has a shipped-product answer. At a recovering mid-single-digit sales multiple for a business growing strategic revenue 35% with a 27% margin, a GAAP-profit date, and the franchise's chief overhang reframed, the risk/reward has tilted from balanced to favorably skewed. The bear's single-session-fade point is real but is a positioning observation, not a fundamental one; we are underwriting the twelve-month re-rate, not the day-one tape.

Model Update & Valuation

The operational and risk-weighting lines in our framework both move favorably this quarter, the mirror image of Q4, where operations rose but the risk band widened. The table frames the changes against where we carried the model after the Q4 downgrade.

Valuation framework: At the reaction-day close of $26.73, Unity trades at roughly 5x trailing sales, recovering off the ~4x washout of February but still well below the ~9.6x forward-sales multiple it carried before the Genie re-rating. The case for the multiple to keep recovering is that the specific risks that took it to the lows have been addressed in sequence: the flat-Grow fear is gone (strategic Grow +49%, Q2 +50–52%), the world-model tail has a shipped-product answer (Unity AI), the path to GAAP profitability has a date and a bridge (Q4 2026), and the convert is paid from cash. What has not changed is the absence of a full-year revenue framework, which keeps a residual visibility discount on the name. Our framework now carries strategic revenue compounding in the low-30s% near-term decelerating toward the 20s as the base grows, a margin structure expanding to record levels in H2 2026 and beyond, GAAP profitability from Q4 2026, and free cash flow scaling faster than revenue. On those inputs, a mid-single-digit sales multiple for a 30%+ strategic grower with a 27% margin and a resolved central overhang is too low, and the direction of travel is up.

Risk/reward: The up-case is that the multiple re-rates as the dismantled bear thesis sinks in: Vector compounds, runtime data lands as upside in H2, Unity AI converts early retention into paid agentic consumption, margins hit record levels, and a heavily-shorted, washed-out name re-rates toward a growth-appropriate multiple. The down-case is that world models improve faster than Unity AI can co-opt them and the disruption narrative re-hardens, or that the still-absent full-year framework was masking a deceleration that shows up once the IronSource runoff stops flattering the strategic-growth optics, or that the consumption-pricing transition proves harder to execute than the framing implies. The difference from a quarter ago is decisive: the one live risk we named at the downgrade (valuation) is easing because the business is re-accelerating into the multiple, and the new tail we flagged (world-model disruption) now has a product answer with early traction. Cheap-and-recovering operations, a resolved central overhang, and a hard profitability date is the basis for the upgrade to Outperform.

Thesis Scorecard Post-Earnings

This scorecard grades the same pillars carried in the standing thesis against what Q1 revealed, and grades the eight specific commitments management made on the Q4 call. The pillar status tags move where the quarter justifies it, most importantly the world-model bear (Bear-4) and the valuation bear (Bear-2), both of which improved materially.

Overall: The thesis is operationally strengthened and risk-profile-narrowed at the same time, the exact inverse of the uncomfortable Q4 combination. All four bull pillars are confirmed (two strengthening, runtime data de-risked), the original single-product and balance-sheet bears are further toward resolution (Bear-3 now a non-issue), the valuation bear (Bear-2) is easing as the business re-accelerates into the multiple rather than away from it, and, most importantly, the world-model bear (Bear-4) moves from an unquantifiable EMERGING tail to a CONTAINED risk with a shipped-product answer. The one pillar still flagged is governance/forward-visibility (Bear-5), which is improving (strategic-revenue pivot, GAAP-profit date) but not resolved while a full-year revenue range remains withheld. The net of a better business and a narrower risk band is a thesis that again supports the conviction an Outperform requires.

Action: Upgrade to Outperform from Hold. We set three gates at the downgrade, and two tripped decisively this quarter: Grow's clean return-to-growth (strategic Grow +49%, Vector accelerating) and AI reframed from threat to tailwind (Unity AI ingesting world-model output, ~70% day-5 retention). The print cleared a raised bar with a 27% margin and added a hard GAAP-profit date the out-year previously lacked. The third gate, a full-year revenue framework, is still open, but it matters less now that the IronSource noise is being removed, the disclosure has pivoted to strategic revenue, and the profitability milestone anchors the out-year. At a recovering ~5x sales for a 30%+ strategic grower with an answered central overhang and a datable path to GAAP profit, cheap-and-recovering operations against a contained tail is an Outperform. The path to a higher-conviction Outperform runs through three observable gates for Q2: runtime data landing with an early read on contribution, Unity AI converting beta retention into paid agentic consumption, and either a full-year framework or the back-half margin step-up that confirms record levels.