Unity Software Inc. (NYSE: U) — A Real Engine Stapled to a Catch-Up Ad Machine

Date: June 19, 2026 Format: Independent equity research — initiation Price reference: ~$27.48 (close 2026-06-18) · Market cap ~$11.8B · EV ~$12.0B

⚡ Claude’s Take

This block is the author’s own independent opinion and general information only — not investment advice. The analysis that follows is deliberately position-free and carries no price target; this block is the single, fenced-off exception.

Verdict: HOLD at ~$27 — accumulate-on-weakness below ~$22, do not chase above ~$33. Not a short. Conviction: low-to-medium.

Unity is two businesses wearing one ticker, and the market is right to be confused. The Create engine is a genuine, durable moat — roughly 70% of mobile-game creation, 71% of the top-1,000 mobile titles, real developer switching costs, and demonstrated pricing power (Unity is pushing through price increases while support tickets fall). That franchise, ~34% of revenue, is worth a clean software multiple. The other ~66% — Grow, the ironSource-derived ad network — is a contested #2 in mobile performance advertising that spent 2022–2024 losing share to AppLovin’s AXON and is now, finally, fighting back with its “Unity Vector” ML ad model (Grow +23.5% in Q1’26, Vector +80% YoY). The whole thesis reduces to one question: is Vector durable share-gain, or a transient post-launch bounce off a depressed base? Nobody — including management — yet knows.

At ~6.5x sales (39th percentile of its own ~5.5-year history), the stock prices neither the bull nor the bear — it underwrites a modest, de-risked base case. A sum-of-the-parts (Create at a software multiple + Grow at a #2-ad-network multiple) lands at ~$9–12B EV, i.e. the current price is fair, not cheap, with little embedded turnaround premium. That is why I won’t chase it here: the franchise is real but the FCF is an SBC-funded mirage (FCF-less-stock-comp was only ~+$16M in FY25, negative every prior year), GAAP losses persist, $3.17B of goodwill has never been impaired despite an abandoned Weta strategy, and the stock is a maximum-beta (~2.1), −93%-max-drawdown speculative-AI name that trades as a C3.ai/ARKK cohort member, not a compounder. Framing: a quality-asset-at-a-fair-price turnaround with a falling-knife-that-bounced tape — cheap optionality only on weakness, not at the current quote. Bullish flip: two-plus quarters of Grow >20% with rising take-rate and GAAP operating income turning positive with SBC/revenue under 15%. Bearish flip: Grow decelerates below 10% as AppLovin/Meta take the incremental dollar, or a goodwill writedown lands. Tag: “A real engine; an unproven ad machine.”

📈 Stock Price Action — Five-Year Event Map

Unity’s five-year chart is a near-complete round trip from speculation to capitulation and a tentative partial recovery: a Sep-2020 IPO that first closed at ~$68 ran to an all-time high of ~$201 (Nov-18-2021) at the peak of the unprofitable-tech bubble, then collapsed ~93% to an all-time low of ~$13.93 (Aug-7-2024) through the rate shock, the self-inflicted “Runtime Fee” fiasco, and a CEO ouster. The stock today trades at ~$27.48 (Jun-18-2026) — inside a 52-week range of roughly ~$16.75–$49.47 and still ~86% below its all-time high. The most recent leg is a violent counter-trend bounce off a ~$17 February-2026 low layered on top of a still-negative six-month and year-to-date trend. (Price levels: FACT, 5-yr daily price data. Drivers: INTERPRETATION.)

Cycle narrative. (1) Unity IPO’d into peak liquidity and tripled as the market paid any multiple for top-line growth. (2) The 2022 rate shock hit the longest-duration, cash-burning equities hardest; Unity’s persistent GAAP losses and the ironSource-merger overhang made it one of the group’s worst performers. (3) Early-2023’s AI-led tech rally and a cost-cut/profitability framing produced a ~60% recovery to the mid-$40s. (4) The Sep-12-2023 Runtime Fee announcement triggered a developer revolt and a self-inflicted credibility crisis; the partial walk-back didn’t stop the slide, and CEO Riccitiello’s October departure crystallized the governance hit. (5) Through 2024, new CEO Bromberg’s “reset” — exiting non-core lines and resetting revenue lower — ground the stock to its $13.93 all-time low. (6) From that base, the Vector turnaround narrative drove a ~3.5x recovery into late 2025. (7) That trade reversed hard into early 2026 as the high-beta AI cohort de-risked, halving the stock. (8) A sharp counter-trend bounce off ~$17 brought it back to ~$27.5, still deep below its highs.

1. Executive Summary

Unity Software is the larger of the two companies that dominate the real-time 3D (RT3D) game-engine duopoly (Unity vs. Epic’s Unreal) and, since the November 2022 all-stock acquisition of ironSource, also a top-three player in the mobile app-monetization (advertising) market. As of FY2025 it reports a single segment disaggregated into Create Solutions (the engine — subscriptions and enterprise; $621M, ~34% of revenue) and Grow Solutions (the ad network; $1,228M, ~66%). The investment debate is unusually clean: a genuinely good business (the engine) is stapled to a structurally harder one (the ad network), and the equity trades as a single, high-beta, GAAP-unprofitable, speculative-AI ticker.

The five-year financial record is poor on the surface and improving underneath. Revenue peaked at $2,187M in FY2023 (the first full year of ironSource), then fell 17% to $1,813M in FY2024 — not cyclical weakness but a deliberate “portfolio reset” under CEO Matt Bromberg (appointed May 2024) that wound down non-strategic, low-margin product lines. FY2025 revenue grew 2.0% to $1,850M, and Q1’26 reaccelerated to +16.8% YoY (Grow +23.5%) on the new Unity Vector ML advertising model. Operating losses narrowed sharply (−$882M FY22 → −$479M FY25); GAAP EBITDA reached roughly breakeven in FY25 (−$18M vs −$346M FY24). Reported free cash flow turned positive in FY23 and reached $401M in FY25.

That FCF, however, is the central quality-of-earnings issue. It is bridged from a −$402M net loss almost entirely by non-cash add-backs: $385M of stock-based compensation (21% of revenue) and $461M of depreciation & amortization (most of it acquired-intangible amortization from ironSource and Weta). FCF net of SBC was only ~+$16M in FY2025 — and deeply negative every prior year. Share count rose ~58% in five years. Unity’s reported “profitability” is, so far, substantially shareholder-funded.

The capital-allocation record is the indictment. Under prior management Unity overpaid at the 2021 cycle top — Weta Digital (~$1.53B, strategy abandoned in 2024) and ironSource (~$2.92B all-stock, ~30% of the company issued at a de-rated price) — leaving $3.17B of goodwill that has never been impaired despite the Weta reversal. A $2.5B buyback authorized in mid-2022 was largely spent near the highs and expired with ~$750M unused. No insider has bought a single share in the open market since the IPO; ~$1.22B was sold. The mitigants are recent and real: SBC is falling (from $649M FY23 to $385M FY25), employee stock options and the ESPP were eliminated effective March 2026, and the new compensation plan is genuinely well-designed (a $400M Adjusted-EBITDA bonus gate was missed in FY24 and zero bonus paid; CEO performance options vest only at $35/$50/$60/$75 share-price hurdles).

At ~6.5x FY2025 sales (39th percentile of its own history), the stock embeds a base-case turnaround — neither the bull (Vector converges toward AppLovin-style economics) nor the bear (structural-#2 ad melt + latent goodwill writedown). The entire asymmetry rests on the durability of Vector-driven Grow growth. This report takes no position and sets no price target (see Claude’s Take for the single fenced exception).

2. Business Overview

What Unity does. Unity operates a platform for creating and operating real-time interactive 2D and 3D content — primarily video games, and increasingly non-gaming “Industries” applications (automotive and aerospace design, film/animation, manufacturing digital twins, architecture, defense). The company was founded in 2004 (Copenhagen, now headquartered in San Francisco), IPO’d on the NYSE in September 2020, and as of FY2025 generated $1,849.6M of revenue at a ~74% GAAP gross margin (FACT: FY2025 10-K, filed 2026-02-11).

Unity historically reported two operating segments; as of the FY2025 10-K it reports a single reportable segment, presenting revenue disaggregated into two “solution” families. This is an important nuance — Create and Grow are managed as one business with a shared go-to-market, not as independent divisions.

Create Solutions (~$621M FY25, ~34% of revenue) — the engine. This is the Unity Editor and runtime: the software developers use to build, ship, and operate games and RT3D experiences across ~20 platforms (iOS, Android, consoles, PC, web, AR/VR). Revenue is predominantly subscription (per-seat Pro and Enterprise plans), supplemented by enterprise support, professional services, and cloud/consumption services (multiplayer hosting, LiveOps, asset management). This is the recurring, sticky, high-margin half of Unity and the source of the genuine moat. Customers range from individual indie developers (often on free or low-cost tiers) to the largest studios and, increasingly, Fortune 500 industrial firms. Create revenue had been inflated pre-reset by enterprise professional-services and consumption lines that Bromberg has since exited — which is why Create fell from $859M (FY23) to ~$620M (FY24–25) even as the core subscription business strengthened.

Grow Solutions (~$1,228M FY25, ~66% of revenue) — the ad network. This is the monetization and user-acquisition stack: the Unity Ad Network (now powered by Unity Vector, the ML ad model rolled out in 2025), the ironSource LevelPlay mediation platform, the legacy ironSource Ad Network (being sunset), offerwall, the Supersonic game-publishing business (being divested), and Aura on-device (carrier) advertising. The revenue model is a net take / agent model: Unity recognizes “the amount we retain from the transaction we facilitate through our auction and mediation platform” (FACT: FY2025 10-K). This is consumption/transactional, performance-contingent revenue — it rises and falls with advertiser spend and with the quality of Unity’s auction relative to AppLovin’s, Google’s, and Meta’s.

Revenue character. The two halves could hardly be more different. Create is recurring software with high gross margins and high switching costs but a TAM capped by the size of the professional game-development and RT3D-industrial market. Grow is a volume-and-take-rate ad business — much larger in dollars, lower-margin, far more competitive, and far more sensitive to a single technical variable (auction/ML performance). Unity is, by revenue, majority an ad-tech company that happens to own the leading game engine — a framing the market often gets backwards.

Verdict: A two-headed business — a durable, undermonetized software moat (Create) bolted to a contested, scale-dependent ad network (Grow) that now drives two-thirds of revenue and essentially all of the growth-vs-decline debate.

3. Industry Dynamics

Unity sits at the intersection of two very different industries, and they deserve separate structural verdicts.

The real-time 3D engine market — structurally GOOD, but a capped duopoly. Game engines are the foundational tooling on which interactive content is built. The market is effectively a duopoly: Unity dominates mobile and indie (Unity claims ~70% share of mobile-game creation; third-party data shows 71% of the top-1,000 mobile games and ~51% of 2024 Steam releases use Unity), while Epic’s Unreal Engine dominates AAA, console, and high-fidelity PC (and monetizes via a 5% royalty above a revenue threshold). Barriers to entry are high in the Greenwald sense — building a competitive multi-platform engine takes a decade and an ecosystem (Asset Store, trained developers, middleware integrations) that cannot be bought. But two structural features cap the prize: (1) a free-tier floor — Unreal’s royalty model and, more pointedly, the open-source Godot engine (gaining momentum among indies after Unity’s 2023 pricing fiasco) limit Unity’s pricing power at the low end; and (2) a TAM ceiling — the professional game-development market is finite, and Unity’s “Industries” expansion (automotive/film/digital twins) has been slower and lumpier than hoped. A genuine longer-term risk, named explicitly in Unity’s own 10-K risk factors, is generative-AI content creation disintermediating parts of the engine value proposition.

The mobile ad-tech / app-monetization market — structurally HARD for a #2. This is where two-thirds of Unity’s revenue lives, and it is a brutal industry for anyone who is not the leader. The market is dominated by AppLovin (the AXON ML engine + MAX mediation) and the walled gardens (Google AdMob, Meta Audience Network), with a long tail (Mintegral, Liftoff, Moloco, Unity/ironSource). The economics are extreme and winner-take-most: AppLovin’s advertising segment generated ~$3.2B revenue (+75% YoY) at ~76% EBITDA margins and holds ~39–40% of iOS ad-revenue share. Through a Marathon (capital-cycle) lens, AXON’s supernormal returns are precisely the kind of profit pool that attracts capital — and Unity, with ironSource and now Vector, is itself the capital rushing in. The hard truth is that the economics of this industry have accrued to the leader (AppLovin), not to the #2/#3 — and they did so during the period Unity owned both an engine and ironSource.

Verdict: One good-but-capped industry (engines) and one structurally difficult industry (mobile performance ads) in which Unity is a subordinate competitor. The blended industry quality is mediocre, dragged down by the larger, harder half.

4. Competitive Position

Unity has one real moat and one contested one, and naming them precisely is the key to the whole analysis.

Create (engine) — a durable moat. Type: customer captivity (switching costs) + intangibles/learning + installed-base scale. The mechanism is concrete: a studio that builds on Unity accumulates a C#/Unity codebase, a team of Unity-trained developers, multi-platform port investments, and a library of Asset Store purchases and middleware integrations — all of which must be rebuilt to switch engines. This captivity is reinforced by scale (the more developers on Unity, the deeper the talent pool, the more third-party tools target it, the more reason the next developer chooses Unity). The financial proof of the moat is the most important evidence in this report: management states it is passing through “moderate price increases” while support tickets fall post-Unity 6, and holds ~70% mobile share. Pricing power amid stable dominant share is exactly what a real moat looks like. The 2023 Runtime Fee debacle was a self-inflicted test of this captivity — developers were furious but, crucially, most did not actually leave, which both demonstrated the switching costs and revealed how recklessly prior management had exploited them.

Grow (ad network) — a contested, compressible advantage Unity lost and is fighting to rebuild. Type: data/scale flywheel in the ML auction. In principle, Unity’s engine should feed its ad model a structural data advantage (the “Create → Grow flywheel” is the bull thesis). In practice, AppLovin — which owns no engine — out-executed Unity-with-an-engine for three straight years (2022–2024), as AXON’s model quality pulled ahead of Unity’s legacy Audience Pinpointer. This is decisive evidence that the engine-data advantage is not automatic and that the relevant moat here is model quality and scale, which is the weaker, more compressible kind of advantage. Unity Vector (2025) is the first credible attempt to actually monetize the engine-data edge: four consecutive quarters of ~15% sequential growth, Vector +80% YoY in Q1’26, Vector now ~56% of Grow (Q4’25), and a March-2026 launch of day-28 IAP-ROAS campaigns that closes a specific ~2-year technical gap to AXON. Early reported uplift is ~15–20% on installs/IAPs. But this is catch-up, not leadership — off a depressed base, against a rival with ~4x the ad revenue and far higher margins, and with Q1’26 guidance already flagging a tough comp from a “large customer win” in 2025.

Direct comparison. Versus Epic/Unreal, Unity wins on breadth and mobile dominance and loses on high-fidelity AAA — a stable, durable split. Versus AppLovin, Unity is the clear #2/#3 in performance advertising with materially worse unit economics, fighting to narrow rather than close the gap. Versus pure-play game publishers (Take-Two, EA), Unity is infrastructure rather than content — it monetizes the medium, not the hits.

Verdict: A durable advantage in the engine, a contested-and-improving advantage in ads. Unity is a dominant-engine company stapled to a subordinate ad network in credible catch-up mode. The franchise quality is real on the ~34% of revenue that is Create and conditional on Vector execution on the ~66% that is Grow.

5. Growth History and Forward Opportunities

History (a peak, a deliberate reset, a reacceleration). Revenue compounded hard into the peak — $772M (FY20) → $1,111M (FY21) → $1,391M (FY22) → $2,187M (FY23, first full year of ironSource) — then fell 17% to $1,813M (FY24). The decline was not demand loss; it was the portfolio reset: management deliberately wound down non-strategic, low-margin Grow lines and exited bloated Create enterprise/professional-services and consumption revenue (Create fell from $859M to $614M). FY2025 grew 2.0% to $1,850M — the first post-reset growth — and Q1’26 reaccelerated to +16.8% YoY total, Grow +23.5%, driven by Vector. This is the inflection the bull case rests on.

Quality of the growth. Mixed. The Create subscription engine grows in the mid-to-high single digits with pricing power — modest but high-quality. The Grow reacceleration is faster but lower-quality in two senses: it is largely a recovery of share previously lost to AppLovin (so partly non-repeatable once the easy comps lap), and it is in the structurally inferior, lower-margin ad business. A transparency concern compounds this: Unity has stopped disclosing dollar-based net expansion rate and the count of >$100k customers — both former headline KPIs and the cleanest gauges of organic health. Removing the scoreboard just as the turnaround narrative takes hold is a yellow flag.

Forward opportunities. (1) Vector — the single biggest swing factor; continued ROAS improvements, IAP-campaign expansion, and self-serve advertiser onboarding could sustain 15–20%+ Grow growth if durable. (2) Unity AI (beta, launched Q1’26) — generative tools inside the editor, both an offensive monetization lever and a defensive answer to the AI-disintermediation risk. (3) Industries / non-gaming — automotive, defense, and digital-twin RT3D applications, a real but slow and lumpy adjacency. (4) Unity 6 adoption and the subscription/runtime monetization optionality on the installed base. (5) Operating leverage — the most certain “growth,” in earnings rather than revenue, as the cost base stays disciplined.

Verdict: Medium-quality growth. The high-quality piece (Create) is slow; the fast piece (Grow/Vector) is real but lower-quality and partly a recovery rather than fresh share-gain. The durability of the Vector reacceleration past Q2’26 — once ironSource/Supersonic noise is gone and the 2025 comps are lapped — is the open question that defines the thesis.

6. Financial Quality

This is where the report earns its keep, because Unity’s headline numbers and its economic reality diverge sharply.

Multi-year financials (reconciled to filings; $M except per-share):

Revenue quality. Discussed in the Growth section: a deliberate reset, then a Vector-led reacceleration, partly offset by the withdrawal of net-expansion and large-customer KPI disclosure.

Margins and operating leverage — genuinely improving. GAAP gross margin recovered to 74% (adjusted gross margin ~83% ex-amortization/SBC). The real story is opex discipline: following layoffs of ~25% of headcount in 2024 (and more in 2025), operating expenses fell ~$435M off the peak — S&M −13% YoY, G&A −35% YoY, with R&D held roughly flat at ~$930M. GAAP EBITDA swung from −$346M (FY24) to −$18M (FY25), near breakeven, and Q1’26 adjusted EBITDA grew +65% YoY on +17% revenue — real incremental-margin leverage as Vector scales. GAAP operating profitability is still roughly two years out, dependent on continued Grow growth, opex restraint, and the scheduled run-off of acquired-intangible amortization.

The central quality-of-earnings issue — “FCF” is shareholder-funded. FY2025 OCF of +$423M / FCF of +$401M (capex is trivial at ~$22M) is bridged from a −$401M net loss almost entirely by SBC ($385M, 20.8% of revenue) plus D&A ($461M, ~$419M of it acquired-intangible amortization). Strip out SBC — a real, recurring, cash-equivalent cost paid in dilution — and FCF was only ~+$16M in FY2025, and was deeply negative ($−323M to −$667M) every prior year. FY2025 is the first year reported FCF survives an SBC deduction, and only barely. The dilution is structural: shares outstanding rose ~58% in five years (~273.5M → ~432.9M, ~9.6%/yr), driven by the all-stock ironSource and Weta deals plus ongoing RSU settlement. The most important positive trend in the entire model is that SBC is falling ($649M → $596M → $385M) and Unity eliminated employee options and the ESPP effective March 2026 — a credible commitment to bend the dilution curve.

Adjusted metrics — standard but aggressive. Management’s Adjusted EBITDA ($409M FY25) and Adjusted EPS ($0.86) add back all SBC and all acquired-intangible amortization. The amortization add-back will mechanically shrink — remaining finite-lived intangibles are ~$650M, amortizing ~$435M in 2026, ~$123M in 2027, ~$93M in 2028, then ~zero — so GAAP optics improve toward breakeven by ~2028 simply as purchase accounting runs off, independent of operations. The SBC add-back, by contrast, represents a permanent ~$385M/year real cost.

Fresh impairment flag. The FY2025 10-K states no material intangible/goodwill impairment in FY23–25 — then Q1’26 took a $278.7M intangible impairment ($226.5M in COGS, $47M in S&M), collapsing reported Q1’26 GAAP gross margin to ~31%, plus smaller investment and PP&E impairments, all tied to winding down non-strategic products. More striking: goodwill of $3.166B (unchanged since FY22) has never been impaired, despite an ~85% peak-to-trough stock decline and the abandonment of the Weta strategy the goodwill partly represents. This is an aggressive accounting posture and a latent risk (see the Risk Analysis section).

Balance sheet and convertible-debt schedule (12/31/25). Cash + equivalents $2,055.8M; convertible debt $2,235.4M carrying; net debt only ~$179.5M. Three series, all out-of-the-money at ~$27:

Near-term solvency is comfortable: the $558M 2026 maturity (intended to be cash-settled) is covered ~3.7x by cash, and the blended coupon (~1%, a ZIRP-era legacy) means interest cost is tiny (~$24M/yr). Unity has opportunistically repurchased 2026 notes at discounts (booking ~$61M/$43M gains in FY24/FY25). The risk is refinancing: all three series are sub-conversion, so the ~$1.69B due 2027–2030 will likely be cash-settled, and refinancing at today’s rates would be materially costlier if the stock stays below conversion.

ROIC/ROE. Negative throughout — Unity has never earned its cost of capital on a GAAP basis since IPO. FY25 operating loss of −$479M sits on ~$6.5B of invested capital; tangible common equity is negative (goodwill + intangibles of ~$3.82B exceed total equity; accumulated deficit −$4,139M).

Verdict: Economics are improving from a very low base — operating leverage is real, SBC is falling, EBITDA is near breakeven — but the business does not yet durably earn its cost of capital, reported FCF is ~entirely SBC-funded until a marginal FY25 crossover, and management’s preferred metrics exclude its two largest real costs. Improving, not yet good.

7. Capital Allocation

Verdict up front: capital allocation has been poor-to-mixed — predominantly value-destructive on M&A, with credible but recent damage control under new management. Unity is a textbook Marathon capital-cycle casualty: it over-earned and over-valued itself at the 2021 peak, used that inflated equity to buy a portfolio it then dismantled, and diluted shareholders ~58% over five years.

M&A — the indictment.

Goodwill sits at $3,166M, unimpaired, despite the Weta reversal — economic impairment was instead routed quietly through intangible amortization (net intangibles fell $1,922M → $650M FY22–FY25, including +$77M of accelerated FY25 amortization from revising useful lives down — an implicit admission). Both major deals were struck at or near the 2021 cycle top, the classic capital-cycle error: high returns attracted capital, then mean-reverted.

Buybacks and dilution. The $2.5B buyback authorized July 2022 was ill-timed — $1.5B/42.7M shares repurchased in FY2022 near the highs, $250M in FY2023, then $0; the program expired November 2024 with ~$750M unused. Shares outstanding rose every year (294M → 433M FY21–FY25) as SBC overwhelmed repurchases. The genuine positives: Unity eliminated all employee stock options and suspended the ESPP (effective March 2026), and managed its converts opportunistically (repurchasing ~$480M of 2026 notes for ~$415M in March 2024). No dividend, appropriately.

Compensation and incentives — a bright spot. The current plan is well-designed and explicitly not EBITDA-only — a positive that pushes against the broader negative verdict. CEO Matt Bromberg’s inducement package includes a 1.0M performance stock option that vests only at $35/$50/$60/$75 share-price hurdles (i.e., he is paid for re-rating the stock, not for adjusted-metric engineering). The FY2024 bonus carried a hard Adjusted-EBITDA $400M gate — it was missed ($390M) and zero bonus was paid, evidence the plan has teeth. PSUs (March 2025) weight 75% revenue / 25% “Adjusted EBITDA Less SBC Margin” — deliberately penalizing dilution, exactly the right metric for this company.

Insider behavior — a clear negative. Across all 403 Form 4s reviewed (2021–2026), there are zero discretionary open-market common-stock purchases (code P) by any insider. Against that, ~17.4M shares were sold for ~$1.22B (Helgason ~$323M, Bar-Zeev ~$160M, Riccitiello ~$212M, Silver Lake ~$360M in 2021, Bromberg ~$13M). The directors-and-officers group stake collapsed from 6.3% (2024) to <1% (2026), and co-founder David Helgason left the board in February 2026. The complete absence of conviction buying after an ~85% drawdown is itself a signal.

Verdict: A poor M&A and buyback record (the substance of past capital allocation) partially redeemed by a genuinely good incentive design and a credible recent commitment to dilution discipline (the new regime). Management has not historically allocated capital intelligently; the jury on the current team’s allocation is still out, but the comp structure and dilution moves are encouraging.

8. Changes and Headwinds — Last Two Years

The last ~30 months are the most eventful in Unity’s public life and frame the entire current setup.

The Runtime Fee fiasco (September 2023). Unity announced a per-install fee on developers — a retroactive, unilateral change to the economics of games already shipped. The developer backlash was immediate and severe; studios threatened to leave the engine, and the episode became a case study in squandering customer trust. Unity partially walked it back within days, and fully cancelled it in September 2024, reverting to the subscription model. The lasting damage was to trust (and to Godot’s adoption), but — critically — most developers stayed, demonstrating the engine’s switching costs even under maximum provocation.

Leadership overhaul. CEO John Riccitiello “retired” October 9, 2023 in the wake of the fiasco; board member Jim Whitehurst served as interim CEO; Matt Bromberg (ex-Zynga COO) became permanent CEO May 1, 2024. Co-founder David Helgason left the board in February 2026. This is essentially a new management team executing a turnaround on assets the prior team assembled.

The “reset” (January 2024 →). Bromberg’s strategy: shrink to grow. ~25% of the workforce was cut (~$214M in separation costs), non-strategic Grow lines and the abandoned Weta film-VFX business were wound down, and revenue was deliberately reset lower. In 2025–26 the reset extended to sunsetting the legacy ironSource Ad Network and selling the Supersonic publishing business, with guidance reframed around “strategic revenue.”

Unity Vector and Unity 6. The two product pillars of the turnaround: Vector (the ML ad model now driving Grow reacceleration) and Unity 6 (the engine release driving improved retention and the pricing-power story). Unity AI (beta) launched alongside in Q1’26.

Headwinds. Intensifying competition from AppLovin (ads) and Godot (engine low-end); the AI-disintermediation risk to content creation; the withdrawal of key KPI disclosures; a high-beta tape that punishes any guidance disappointment violently; and the latent goodwill/refinancing risks discussed elsewhere.

Verdict: Net thesis-strengthening on balance — the reset and leadership change addressed real problems and produced a credible Vector-led reacceleration — but the strengthening is recent, fragile, and not yet proven durable. The same two years also contain the fiasco and the disclosure withdrawals that justify a skeptical stance.

9. Risk Analysis

Catastrophic-loss risk is low in the near term — Unity is not over-levered (net debt ~$180M against $2.06B cash), the engine franchise is durable, and there is no single cliff event. The realistic bad outcome is not bankruptcy but multiple compression if Vector disappoints and the stock re-rates toward a structural-#2 ad multiple.

10. Valuation Discussion (Embedded Expectations)

No price target, no recommendation — embedded-expectations and scenario framing only.

Setup. At ~$27.48, ~433M shares → market cap ~$11.8B; net debt ~$180M → EV ~$12.0B. Against FY25 revenue of $1,849.6M, that is ~6.5x sales (~6.2x on TTM through Q1’26). On its own ~5.5-year history Unity sits at the 39th percentile composite (P/B 4.0x = 44th; P/S 6.1x = 35th) — the cheaper third of its own range, but not remotely a trough multiple.

Comparables (note the revenue-recognition trap). Unity, Roblox, and Take-Two recognize gross revenue; AppLovin recognizes net (agent model), so APP’s ~22–24x EV/sales is not comparable to Unity’s ~6.5x.

The clean read is Unity vs. Roblox (~4.9x) and Take-Two (~5.6x): Unity trades at a premium to both pure-play game platforms despite slower growth and GAAP losses — justified only if Grow/Vector re-rates toward AppLovin-style economics. Versus AppLovin, Unity is “cheap because unproven” — but AppLovin is expensive for a reason (it earns ~76% EBITDA margins and leads the auction).

Sum-of-the-parts sanity check. Create (~$650M revenue × 8–10x software multiple) = ~$5.2–6.5B; Grow (~$1.3B × 3–4x #2-ad-network multiple) = ~$3.9–5.2B → ~$9–12B EV. The current ~$12B EV is therefore at the top of a fair SOTP range — the price already credits Create with a software multiple and Grow with a healthy ad-network multiple, leaving little embedded turnaround premium. The stock is “show-me,” not euphoric, but also not a bargain.

Embedded-expectations scenarios (illustrative; no price target):

What the market is underwriting. The current price embeds roughly the base case — a modest, de-risked turnaround. It is not pricing the bull (Vector → AppLovin convergence) and not pricing the bear (structural-#2 melt + impairment). The entire asymmetry reduces to a single swing variable: whether Q1’26’s Grow +23.5% is durable share-gain or a transient post-launch bounce. Get that right and the rest follows.

11. Variant Perception

Consensus. The sell-side and the recent tape price a Vector-led turnaround working — the ad business is “fixed enough,” the engine is a free option, and the stock’s sharp 2025 recovery and 2026 bounce reflect restored confidence. Consensus ≈ the base case, tilting bullish on momentum.

Strongest bull case. (1) The Create engine moat — sticky, high-switching-cost, ~70% mobile share with demonstrated pricing power — is underappreciated and worth a full software multiple on its own. (2) Vector structurally closes the AppLovin ROAS gap, and the engine-data flywheel finally monetizes, driving durable Grow reacceleration. (3) Operating leverage on a now-disciplined cost base flips GAAP to profit and the 2030 converts ($36) move in the money. (4) The stock is cheap versus its own history and versus the economics it would have if it converged toward AppLovin.

Strongest bear case. (1) Unity is a structural #2 in performance advertising; AXON’s data/scale moat is self-reinforcing and likely too wide to close — Unity competes for residual non-walled-garden spend. (2) The engine TAM is game-dev-capped and faces a real (if slow) gen-AI disintermediation threat. (3) Reported FCF is an SBC-funded illusion (FCF-ex-SBC ~$16M; shares +58% in five years). (4) $3.17B of goodwill has never been impaired despite the abandoned Weta strategy — a latent writedown. (5) At ~6.5x gross sales, Unity is more expensive than Roblox/Take-Two for slower, unprofitable growth.

The 3–5 assumptions that matter most, with falsification tests:

1. Grow/Vector reaccel is durable share-gain. Falsifies bull: Grow decelerates below ~10% over 2–3 quarters, or AppLovin/Meta visibly take the incremental dollar. Falsifies bear: Grow sustains >20% for 3+ quarters with rising take-rate/retention.
2. GAAP operating leverage materializes (15–25% mature margin). Falsifies bull: op margin stalls negative through FY27. Falsifies bear: 2+ consecutive quarters of positive GAAP operating income with SBC/revenue <15%.
3. Create holds a software multiple / isn’t AI-disrupted. Falsifies bull: Create growth turns negative or a credible AI-native engine substitute emerges. Falsifies bear: Create reaccelerates on runtime/subscription monetization.
4. Goodwill ($3.17B) is recoverable. Falsifies bear: no impairment; segment cash flows support carrying value. Falsifies bull: an impairment charge.
5. Real (ex-SBC) FCF inflects positive and sustainable. Falsifies bull: FCF-ex-SBC stays ~breakeven amid ongoing dilution. Falsifies bear: FCF-ex-SBC scales >$200M with a shrinking share count.

Factor-positioning input (where consensus may be offsides). Quantitatively Unity is a maximum-beta (~2.1), deep-drawdown (−93% max), negative-alpha (−0.52) speculative-tech/AI name whose nearest factor twin is C3.ai (0.99 similarity), surrounded by ARKK/ARKW/ROBT and the CWB convertibles ETF (consistent with $2.2B of converts driving part of its trading). The eye-catching m3 +398% annualized (≈+39% raw quarter) is a counter-trend bounce inside a still-negative 6-month/YTD trend (rs_6m −38%) and a stock ~86% below its high. The factor read frames this as a “falling-knife-that-bounced,” not a confirmed momentum re-rating — which corroborates the variant view that consensus has swung to “turnaround working” and would be badly offsides if the Vector reacceleration proves transient. With beta 2.1 and a −93% drawdown history, the multiple compression on any disappointment would be violent.

12. Fact vs. Interpretation Table

13. Open Questions

1. Is Vector durable share-gain or a recovery bounce? The single most important question; resolves over the next 2–4 quarters once ironSource/Supersonic noise clears and 2025 comps lap.
2. Why did Unity stop disclosing dollar-based net expansion and >$100k-customer counts? Removing the cleanest organic-health KPIs as the narrative turns warrants explanation.
3. When (if) does goodwill get impaired? $3.17B unimpaired through the Weta abandonment and an 85% drawdown is anomalous.
4. What is the steady-state mature operating margin of a blended engine-plus-#2-ad-network business, and when is it reached?
5. Does the engine-data flywheel actually confer an ad advantage, or did AppLovin already prove it doesn’t matter?
6. How are the 2027/2030 converts refinanced if the stock stays sub-conversion?
7. Can Industries (non-gaming RT3D) ever become material, or is it a perennial “next year” story?

14. What Must Be True

Bull case — what must be true:

• Grow sustains >20% growth for 3+ quarters with rising take-rate, demonstrating Vector is durable share-gain (not a comp-driven bounce).
• GAAP operating income turns positive within ~2 years with SBC/revenue falling below 15%.
• Create holds its ~70% mobile share and pricing power, and is not disrupted by gen-AI.
• The convertibles are refinanced or converted without a dilutive/expensive event.
• Falsification test: a single quarter of Grow below ~10%, or a guide that resets the Vector trajectory lower, breaks the bull thesis.

Bear case — what must be true:

• Grow decelerates toward high-single-digits as AppLovin/Meta capture the incremental performance-ad dollar, confirming structural-#2 economics.
• Real (ex-SBC) FCF stays near breakeven amid continued dilution; GAAP losses persist.
• A goodwill or further intangible impairment crystallizes the past M&A overpayment.
• The multiple compresses toward a #2-ad-network level (~3.5–4x sales).
• Falsification test: two consecutive quarters of >20% Grow growth and positive GAAP operating income with SBC/revenue <15% breaks the bear thesis.

15. Source Appendix

See the Source Appendix (Appendix B) for the full citation list. Primary sources: Unity Software FY2021–FY2025 10-Ks (EDGAR CIK 0001810806), FY2025 10-K filed 2026-02-11; Q1’26 results and earnings-call transcript (2026-05-07); DEF 14A proxies (2024–2026); the Form 4 insider corpus (403 filings); 8-K material-event filings (runtime fee, CEO transition, reset). Quantitative data: ROIC.ai (statements, ratios, enterprise value), public price history and valuation-index own-history percentiles, and FactorsToday (factor loadings, leaderboard, related stocks). Peer context drawn from public filings of AppLovin (APP), Roblox (RBLX), Take-Two (TTWO), and Electronic Arts (EA). Third-party engine-share and competitive data cited inline with URLs and access dates.

No buy/sell recommendation and no price target appears in the body sections; the single fenced exception is the Claude’s Take block. Management commentary is treated as hypothesis and validated against filings and external evidence throughout.

APPENDIX A — Standard Diligence Questionnaire

Unity Software Inc. (NYSE: U) — supplemental to the research memo. Fact / Interpretation / Assumption labels applied where it matters.

General

What thoughtful questions have other investors asked about this company? The recurring institutional questions cluster on five points: (1) Is “Unity Vector” durable share-gain against AppLovin, or a transient post-launch bounce off a depressed base? (2) Is the reported positive free cash flow real, given ~$385M of annual stock-based compensation? (3) Why has $3.17B of goodwill never been impaired despite the abandoned Weta strategy and an ~85% stock drawdown? (4) Why did management stop disclosing dollar-based net expansion and the >$100k-customer count? (5) Can the durable engine moat (Create) ever offset the structural-#2 economics of the ad network (Grow)? These map directly onto the the Open Questions section.

Cyclicality & Earnings Nature

Are earnings at a cyclical high or low? Neither in the normal sense — Unity is GAAP loss-making and has been every year since IPO. On a margin basis, earnings are at a cyclical/structural trough that is inflecting upward: operating losses narrowed from −$882M (FY22) to −$479M (FY25), and GAAP EBITDA reached near-breakeven. (INTERPRETATION: the inflection is real but early.)

Driven by the external environment or internal actions? Predominantly internal — the FY24 revenue decline and margin improvement both stem from the deliberate “portfolio reset” (exiting low-margin lines, ~25% headcount cut), not from market demand. The Grow reacceleration is part internal (Vector) and part external (mobile-ad-spend environment).

How stable are revenues? Mixed. Create (subscription, ~34%) is recurring and stable. Grow (ad take-rate, ~66%) is consumption/performance-based and volatile — sensitive to advertiser budgets and to Unity’s auction quality versus competitors.

Outlook for products/services? Engine: slow, steady growth with pricing power. Ads: the swing factor — 15–20%+ growth if Vector is durable, high-single-digits if it is a recovery bounce.

How big is this market — growing, shrinking, domestic or international? The RT3D engine market is large and growing but capped; the mobile ad-tech market is very large (~tens of billions) and growing but winner-take-most. Unity is global (San Francisco HQ; significant operations in Israel via ironSource, plus Europe and Asia).

Business Quality & Competitive Moat

Is the industry getting more or less competitive? More. Ads: AppLovin’s lead widened 2022–2024; Google/Meta walled gardens persist. Engine: Godot (open-source) gained momentum after the Runtime Fee fiasco, and gen-AI content tools loom.

How profitable is the business (ROIC, ROE)? Negative on a GAAP basis throughout — Unity has never earned its cost of capital since IPO. Tangible common equity is negative. (FACT)

How profitable is the industry — competitors, barriers? Bifurcated. The engine duopoly has high barriers (decade-plus build, ecosystem, switching costs). Mobile ad-tech has extreme winner-take-most economics — AppLovin earns ~76% EBITDA margins; the #2/#3 earn far less.

Can the business be easily understood? Mostly — but the single-segment presentation, the net-vs-gross revenue distinction in ads, and the SBC-funded FCF require careful reading.

Can it be undermined by foreign low-cost labor? Not directly; this is IP/software/network-driven, not labor-cost-driven. The real substitution threat is technological (gen-AI, open-source engines).

Do brands matter? Moderately — “Unity” and “Unreal” are strong developer brands, but the moat is switching costs and ecosystem, not consumer brand.

Nature of competition? Engine: feature/performance/ecosystem and price. Ads: ML auction/ROAS quality and scale of data.

Customers’ switching costs? High in the engine (codebase, trained talent, ports, Asset Store) — demonstrated by developers staying through the Runtime Fee crisis. Low-to-moderate in ads (advertisers and publishers multi-home across networks and chase ROAS).

Financial Condition & Balance Sheet

Assets not fully recognized on the balance sheet? The Create engine’s installed base, brand, and developer ecosystem are largely internally generated and under-recognized — the genuine economic asset. Conversely, $3.17B of acquired goodwill may be over-recognized given the Weta reversal.

Off-balance-sheet liabilities? None material flagged beyond ordinary operating leases; convertible debt is on balance sheet.

How conservative is the accounting? Aggressive in two respects: goodwill never impaired despite a strategy abandonment, and Adjusted EBITDA/EPS add back all SBC and all acquired-intangible amortization. Revenue recognition (net take in ads) is standard.

How CapEx-hungry? Very capital-light — capex ~$22M FY25 (~1.2% of revenue). The “investment” is in people (R&D/SBC), not physical assets.

Capital Allocation & Management

How much FCF, and how is it used? Reported FCF $401M FY25, but only ~+$16M net of SBC. No dividend; the buyback lapsed; cash is used to manage converts and fund operations. (FACT)

Significant acquisitions recently? The era of large M&A appears over — the current regime is divesting (Supersonic) and sunsetting (ironSource Ad Network), not buying. The legacy deals (Weta ~$1.53B 2021, ironSource ~$2.92B 2022) define the poor historical record.

Buying back shares? Not currently — the $2.5B authorization expired Nov-2024 with ~$750M unused; net share count keeps rising from SBC.

Issuing large amounts of new shares to insiders? Historically yes (SBC peaked at $649M FY23), but falling ($385M FY25) and options/ESPP eliminated effective March 2026 — a credible reversal.

Compensation policy? Well-designed and a relative bright spot: CEO performance options vest only at $35/$50/$60/$75 share-price hurdles; a $400M Adjusted-EBITDA bonus gate was missed (FY24) and zero bonus paid; PSUs use “Adjusted EBITDA Less SBC Margin” to penalize dilution.

Motivations of management? New team (Bromberg since May-2024) incentivized to re-rate the stock and reduce dilution. Caveat: directors-and-officers stake collapsed from 6.3% to <1%, and zero open-market buying — limited “skin in the game.”

Valuation & Market Data

ADR, MLP, or K-1 issuer? No — ordinary NYSE-listed U.S. common stock; standard 1099 treatment.

Dividend policy? None, and none expected — appropriate for a sub-scale, GAAP-unprofitable growth company.

How profitable is the business? GAAP-unprofitable; ~$409M management Adjusted EBITDA (which excludes ~$385M of real SBC).

Is net income diverging from cash from operations? Yes, and materially — FY25 net loss −$402M vs. OCF +$423M, an ~$825M gap bridged by SBC ($385M) and D&A ($461M). The divergence is the core quality-of-earnings issue: cash generation is real operationally but a large share is funded by dilution.

Risks & Downside

What would cause the stock to decline? A Grow/Vector deceleration; a goodwill impairment; continued GAAP losses; a high-beta/macro risk-off; a guidance reset. (See the Risk Analysis section.)

Risk of catastrophic loss? Low near-term — net debt is only ~$180M against $2.06B cash, and the engine franchise is durable. The realistic bad case is severe multiple compression, not insolvency.

Chance of total loss? Very low. Unity has a defensible core asset, a manageable balance sheet, and no near-term solvency cliff.

Recent News & Events

Has the business environment changed recently? Yes — the Vector-led Grow reacceleration (Q1’26 +23.5%) is the material change, alongside the ongoing reset (sunsetting ironSource Ad Network, selling Supersonic) and the elimination of options/ESPP.

Significant acquisitions? No recent acquisitions — the posture is divestiture.

Change in accounting policies? Moved to single-segment reporting (Create/Grow now disaggregated revenue); revised intangible useful lives down (accelerating amortization); withdrew net-expansion and >$100k-customer KPI disclosure.

Recent changes — new markets, facilities, management? New CEO (Bromberg, May-2024), founder off the board (Feb-2026), Unity 6 and Unity AI launched, Vector rolled out. A genuinely new chapter executed by a ~2-year-old team on legacy assets.

APPENDIX B — Source Appendix

Unity Software Inc. (NYSE: U) — sources consulted, June 19, 2026. Primary sources prioritized; third-party/aggregator data labeled and reconciled to filings. All figures reconciled to SEC filings where the filing is authoritative.

Primary — SEC Filings (EDGAR, CIK 0001810806)

Primary — Earnings Calls / Management Commentary

• Unity Q1 2026 earnings call transcript, 2026-05-07 (via ROIC.ai transcript service) — Vector trajectory, ~70% mobile share + pricing power, sunsetting ironSource Ad Network / selling Supersonic, “strategic revenue” framing.
• Unity Q4/FY2025 results, investors.unity.com, 2026-02-11 — segment revenue, Vector = 56% of Grow.

Quantitative Data Sources (third-party, reconciled to filings)

• ROIC.ai — income statement, balance sheet, cash flow, profitability ratios, enterprise value, valuation multiples (FY2020–FY2025). Used for the multi-year financial table and EV build.
• Public market data — daily price history (2020-09-18 to 2026-06-18) and valuation-index own-history percentiles (P/B 44th, P/S 35th, composite 39th).
• FactorsToday — factor loadings (Market β ~1.65–1.99; anti-Quality, anti-Momentum, SmallSize tilt), leaderboard (5yr −23%/yr ann., −93% max drawdown, m3 +398% ann.), stock-info (β ~2.1, alpha −0.52, relative strength), related-stocks (C3.ai 0.99, JOBY, ARKK, ARKW, CWB).

Peer / Comparable Context (public filings and market data)

• AppLovin (APP) — AXON/MAX moat, ~76% EBITDA margin, ~$3.2B ad revenue, iOS share; the key competitive benchmark.
• Roblox (RBLX) — UGC platform valuation comp.
• Take-Two (TTWO) — game-publisher valuation comp.
• Electronic Arts (EA) — game-publisher context (take-private).

Third-Party Industry / Competitive Data (cited inline; access date 2026-06-19)

• Engine market share: 71% of top-1,000 mobile games on Unity; ~51% of 2024 Steam releases — cubix.co, worldmetrics.org.
• Unity Vector vs. AppLovin AXON: ainvest.com (2025-09), advisoranalyst.com (2025-07), naavik.co, mobilegamer.biz.
• Weta strategy unwind: fxguide.com, rnz.co.nz (Dec-2023), variety.com (2021).

Notes on Source Authority

SEC filings are authoritative for all financial and structural facts; where third-party aggregated data differed, the filing governed. FactorsToday and the valuation-percentile data are statistical/own-history context, not cross-sectional valuation conclusions. Management commentary (transcripts, releases) was treated as hypothesis and validated against filings and external competitive data. The analysis rests entirely on public primary sources and the third-party data noted above.