Figma, Inc. (NYSE: FIG) — The Category-Killer the Market Marked Down From Bubble to Coin-Flip on Whether AI Builds Its Moat or Buries It

Independent Equity Research Report date: 2026-06-20 · Coverage: Application Software (Collaborative Design) Subject: Figma, Inc. · NYSE: FIG · CIK 0001579878 · FY-end Dec 31 · HQ San Francisco, CA Price at writing: ~$18.88 (6/18/26 close) · Market cap: ~$10.0B · Net cash: ~$1.6B · EV: ~$8.4B

⚡ Claude’s Take

This block is the author’s own independent opinion and is general information only — not investment advice and not a recommendation to buy or sell any security. The analytical body of this report (Sections 1–15, below) takes no position and names no price target; only this block does otherwise. Do your own research and consult a licensed adviser before investing.

Verdict: HOLD here / accumulate-on-weakness toward the low-to-mid teens (~$13–17) / NOT-a-short. Fair-value zone ~$22–30. Conviction: MEDIUM.

Figma is a genuine category-killing franchise — it killed Adobe XD, marginalized Sketch, buried InVision, and was good enough that Adobe agreed to pay $20B for it and then handed over a $1B check just to walk away. It compounds revenue at 35–46% with a 139% net-dollar-retention rate (a 2-year high, rising into the very deceleration the market fears), generates real cash, and sits on a fortress ~$1.6B-net-cash balance sheet with effectively no debt. The IPO priced this at ~76x sales and ~$60B; eleven months later the market has marked it down ~85% to ~7x sales and ~$9B. That is not a fundamentals collapse — revenue accelerated the entire way down. It is a multiple round-trip from bubble to abandonment. At ~7x EV/sales for a 35%+ grower with 80% gross margins, the bubble premium is gone; this is no longer expensive on growth.

So why only HOLD, and why accumulate lower rather than here? Because two things keep this from being a layup. First, the bear case is the company’s own 10-K: Figma concedes in writing that AI may let customers “reduce the number of seats purchased or renewed even if output increases,” and that “switching costs may decline as AI tools enable automated recreation of design systems.” The single question that decides this stock — does generative AI make Figma the indispensable human-and-agent control surface, or does it dissolve the design layer and route around the canvas — is genuinely unresolved, and the leading indicators (NDR, seat growth, the new AI-credit monetization S-curve) won’t settle it for several more quarters. Second, the headline “FCF-positive” story is half an illusion: stock-based comp ran ~$1.36B in FY25 (129% of revenue, an IPO catch-up) and still runs ~50% of revenue at the Q1-26 run-rate, so owner FCF (FCF minus SBC) is negative ~$320M annualized, and dilution is ~6–7%/yr against a dual-class structure where founder Dylan Field controls 72.3% of the votes on ~10–13% of the economics. The de-rate removed the bubble, but at $18.88 the tape is pricing the low end of base — you are not being paid a margin of safety for a live existential risk. Framing: a quality-compounder-at-a-fair-price whose terminal value is a real AI coin-flip — a de-rated franchise, not yet a bargain, and decidedly not a short given the cash, the moat, and the accelerating top line. Tag: “They marked the category-killer down 85% — from a bubble to a coin-flip on AI.”

• Conviction: Medium. The franchise quality is high-confidence; the terminal AI outcome and the SBC-normalization path are genuinely uncertain, which caps conviction.
• What flips it bullish: Two-to-three quarters of NDR holding >130% with growing seat counts and materially disclosed double-digit-% AI-credit/consumption revenue — i.e., proof AI is additive, not substitutive — alongside SBC visibly tracking toward ~30% of revenue.
• What flips it bearish: NDR rolling toward ~115%, net seat adds stalling, gross margin drifting below ~75% on inference costs, or SBC stuck above ~40% past FY27. Any one would confirm “AI compresses the seat pool and there is no real profit underneath.”

📈 Stock Price Action — Five-Year Event Map

Factual price history, not a recommendation. Figma IPO’d 2025-07-30, so the “five-year” map is necessarily an ~11-month post-IPO arc. Price moves are FACT; attributed drivers are INTERPRETATION. No price target, no support/resistance, no chart-pattern reading.

Arc. Figma’s entire public life is a single round-trip from euphoria to abandonment. The IPO priced at $33 (7/30/25); the stock opened $85 on its first trading day, printed an intraday high of $142.92, and closed $122 on 8/1/25 (the all-time high) — roughly 76x EV/sales. It then bled almost monotonically for ~9 months to an all-time-low close of $16.86 (4/29/26), bounced to ~$27 (6/1/26), and gave it back to $18.88 (6/18/26). The stock is now −84.5% off its ATH, below its $33 IPO price, in a 52-week range of roughly $16.60–$142.92. Crucially, revenue growth accelerated (38% → 40% → 46% YoY) across the entire decline — this was a multiple de-rate, not a fundamental miss.

Cycle narrative. (1) IPO pop — a textbook 2025 hot-IPO blow-off; scarcity plus design-category hype carried it to ~76x sales. (2) The de-rate — the multiple, not the business, drove the ~60% fall; EV/sales went 76x → ~26x while revenue compounded 40%+. (3) Q3 print — a strong growth quarter that did nothing to halt the re-pricing. (4) Mid-cycle bleed — another leg down as EV/sales ground toward single digits, tracking the broad AI-software de-rate. (5) Q4 print — the +35% FY26 guide and accelerating revenue sparked a relief bounce. (6) Knife to the ATL — the bounce failed; the design-layer-disruption thesis hit peak pessimism at $16.86. (7) Q1 print — +46% revenue (accelerating) snapped the stock ~28% higher. (8) Give-back — the rally faded below $20 as the “$50B wipeout” framing crystallized; the AI-overhang discount persisted despite accelerating growth.

1. Executive Summary

Figma is the dominant collaborative product-design platform — a browser-based, real-time multiplayer canvas that has become the system of record for how the world’s best software teams design, prototype, and hand off to code. It is a high-quality, fast-growing franchise: revenue grew from $505M (FY23) to $749M (FY24, +48%) to $1,056M (FY25, +41%), and re-accelerated to +46% in Q1-26 ($333M) even as the deceleration narrative gripped the stock. Net dollar retention is 139% (a 2-year high), the paid-customer base is ~690,000 (+54% YoY), gross margins are software-grade (though compressing), and the balance sheet is a fortress: ~$1.6B net cash, no funded debt.

The investment question is not whether Figma is a good business — it plainly is. The question is whether generative AI is Figma’s engine or its undertaker, and that question is genuinely unresolved. The bull case is that when execution becomes cheap (AI writes the code), human design taste and a shared “source of truth” become the scarce, defensible edge, and Figma becomes the human-plus-agent control surface for all software creation. The bear case — articulated, remarkably, in Figma’s own 10-K risk factors — is that AI compresses the number of design seats organizations need “even if output increases,” that AI can enable “automated recreation of design systems” (eroding the switching-cost moat), and that AI inference costs structurally cap gross margins. The market has voted with the bears: a ~85% de-rate from a ~$60B IPO peak to a ~$9B market cap, from ~76x sales to ~7x.

Two financial truths sit underneath. First, the GAAP losses are stock-based compensation, not cash burn — FY25’s $1.29B operating loss is dwarfed by $1.36B of SBC (an IPO RSU catch-up), and the business generated +$245M of headline free cash flow. Second, that headline FCF flatters the economics: SBC still runs ~50% of revenue at the Q1-26 run-rate, so owner FCF (FCF minus SBC) is negative ~$320M annualized, and ~6–7%/yr dilution is a real per-share leak. Figma is genuinely improving but is not yet a cash machine for shareholders.

At ~7x EV/sales the bubble premium is gone, and on a growth-adjusted basis the multiple is reasonable versus slower-growing peers (ADSK ~6.7x, NOW ~7.6x at half the growth). But the price embeds a demanding-but-not-heroic bar — roughly high-20s%/low-30s% revenue CAGR for 5–6 years and SBC halving toward ~25% of revenue — and gives little credit either to AI as a tailwind or to AI as a margin/seat threat. The honest read: a de-rated, category-defining franchise priced near the low end of a reasonable base case, with a live existential question still open. Great business; fair, not screaming, price; terminal value a coin-flip the next several prints will adjudicate.

2. Business Overview

What Figma sells. Figma is a single, browser-based, multiplayer (real-time collaborative) platform for designing and building software, sold by subscription on a per-seat basis. Its defining technical achievement was bringing Google-Docs-style simultaneous multi-user editing to professional vector design inside a web browser — a “performant multiplayer canvas” that incumbents built on native desktop apps could not match. One connected canvas now spans the product-development lifecycle, from ideation through design, prototyping, developer handoff, and increasingly publishing and AI-generated build. (FACT — 10-K FY2025 business section; S-1, accessed 2026-06-20.)

The product portfolio. The core is Figma Design (the vector design tool). Around it sits a portfolio that roughly doubled in 2025:

• FigJam (2021) — online whiteboard for ideation/brainstorming.
• Dev Mode (2023) — a design-to-code handoff workspace for developers, now including a Model Context Protocol (MCP) server (2025) that lets an AI coding agent read Figma designs and convert them to code. (MCP weekly active users in Design grew 5x QoQ in Q1-26 — Q1-26 call.)
• Figma Slides (2024) — collaborative presentations.
• 2025 launches that “doubled the portfolio”: Figma Sites (design and publish a website to a custom URL), Figma Make (AI prompt-to-functional-prototype/app), Figma Buzz (marketing/asset content generation at scale), and Figma Draw (illustration).
• Figma Weave (from the Oct-2025 acquisition of Weavy Inc.) — AI image/video generation with leading third-party models embedded in Figma; and an AI Assistant (alpha). (FACT — 10-K; S-1.)

Seat types (five) and pricing tiers (four). Seats: Viewer (free; view + comment), Collab (FigJam + Slides), Content (Buzz, Sites CMS, FigJam, Slides), Dev (Dev Mode + Content), and Full (everything, incl. Figma Design and Make). Tiers: Starter (free), Professional, Organization, and Enterprise (multi-product/brand, design-system management, enterprise controls/compliance). Crucially, the seat roles deliberately “blur,” so non-designers (PMs, engineers, marketers) buy lower-tier seats — the mechanism by which a designer-led land turns into a whole-org expand. (FACT — 10-K; S-1.)

How it makes money — seat subscriptions, with an emerging AI-consumption layer. Revenue is overwhelmingly recurring per-seat subscription, typically billed annually. In 2025 Figma introduced AI Credits across all seats; beginning March 18, 2026 it started enforcing AI-credit limits and rolling out add-on credit subscriptions / pay-as-you-go overage billing — the first material non-seat (usage-based) monetization line, and Q2-26 is its first full quarter. (FACT — 10-Q Q1-26 MD&A; 10-K “AI Credits.”) This consumption layer is strategically central: it is the designed hedge against AI seat-compression (if output-per-seat rises via AI, usage billing can recapture value even if seat counts plateau), but the 10-K’s own risk factors warn it can “reduce predictability of revenue and adversely affect margins.”

Customers, scale, and geography. Figma reports ~13M+ monthly active users, of whom roughly two-thirds are non-designers — confirming the platform has escaped the “designer tool” box into a whole-org collaboration surface. Paid customers spending >$10K ARR: 7,233 (FY23) → 10,517 (FY24) → 13,861 (FY25) → 15,218 (Q1-26, +37% YoY). Paid customers >$100K ARR: 630 → 963 → 1,405 → 1,525 (Q1-26, +48% YoY). 95% of the Fortune 500 and 78% of the Forbes Global 2000 used Figma (Mar-2025). The business is genuinely global: international was 53.4% of FY25 revenue ($564.2M of $1,055.8M) and grew faster than the US (+44.8% vs +36.8%); ~85% of MAUs are outside the US, distributed bottoms-up via URL-shareable, six-language-localized product-led growth. No single customer is >10% of revenue. An ecosystem of >250,000 community resources and >10,000 plugins/widgets sits on Figma’s public APIs. (FACT — 10-K Key Business Metrics and geographic note; S-1.)

Verdict (Business Overview): A high-quality, recurring-revenue, product-led software franchise that has expanded from a designer tool into the connected platform on which whole product organizations build software. The model is asset-light, globally diversified, and broad-based (no customer concentration), with a structurally important and as-yet-unproven pivot toward usage-based AI monetization layered on top of seats.

3. Industry Dynamics

Where Figma sits. Figma occupies the design-and-handoff layer of the software-creation value chain — upstream of code, downstream of ideation. Historically this was a structurally attractive niche: high-gross-margin SaaS, strong customer captivity via embedded design systems, and (until Figma) a market fragmented across single-player point tools (Adobe XD, Sketch, InVision, Zeplin, Abstract). Figma consolidated it.

The market it already won. Real-time browser multiplayer let Figma displace the incumbents of the 2015–2020 era decisively. Adobe wound down Adobe XD; Sketch (macOS-only, single-player) was marginalized; InVision shut down its core product in 2024. By the Greenwald market-share-stability test, Figma won the prior competitive war outright. The ultimate validation: Adobe agreed to acquire Figma for ~$20B in 2022, then both parties terminated in December 2023 citing “no clear path” to EU/UK/US antitrust approval — with Adobe paying Figma a $1.0B termination fee. The regulators’ implicit finding (and Adobe’s willingness to pay $20B plus a $1B break fee) is third-party evidence that Figma’s position was dominant and hard to replicate. (FACT — 10-K Recent Developments; widely reported.)

Market size. Figma sizes its TAM at $33B today, based on a commissioned IDC study projecting the global “software-design workforce” expanding to 144M by 2029 and >1B new apps by 2028. (FACT-as-claim — S-1 “Market Opportunity.”) Caveat (ASSUMPTION): this is a self-commissioned, pricing-derived TAM (“design workforce × our ASP”) — directional, not independent, and it inflates precisely if AI compresses the number of design seats (the core bear case). For cross-reference, adjacent design-software pools run roughly ~$24B (global CAD) and ~$10–15B (BIM/AEC) — i.e., Figma’s claimed TAM is of the right order of magnitude for a broad “software-design” definition but leans on the optimistic interpretation of how AI affects seat counts.

The capital-cycle warning (Marathon lens). This is the key supply-side concern. The adjacent “AI app-builder / vibe-coding” space — Lovable, Vercel v0, Cursor/Anysphere, Replit, Bolt, plus hyperscaler entrants (Google Stitch) and frontier labs (Anthropic, OpenAI) building design+code tooling — has attracted enormous venture capital at rich valuations across 2024–2026. By Marathon’s framework, a rash of well-funded entrants, abundant capital, analyst extrapolation of an AI-design boom, and IPO/secondary activity (Figma’s own 2025 IPO) are classic markers of a sector in its boom/over-investment phase. And the entry barrier at the AI layer is low: foundation models are off-the-shelf and available to all (a Greenwald “third-party IT innovation confers advantage on none”). High returns plus abundant capital plus low barriers predict supply-side erosion of the design profit pool.

Verdict (Industry): The legacy collaborative-design niche is a structurally good industry — high-margin, captive, Figma-dominated. But the relevant industry is being redefined in real time by AI software-creation, which on a Marathon view is structurally less attractive: a flood of well-funded entrants, foundation models available to all, and a credible path for the design layer to be disintermediated by prompt-to-code. Net: a good business in an industry whose supply side has just turned hostile. The structural attractiveness verdict is therefore bifurcated and trending less favorable — exactly the tension the stock is wrestling with.

4. Competitive Position

Name the moat. Figma’s advantage is primarily demand-side customer captivity reinforced by genuine network effects, plus a modest scale/ecosystem layer — Greenwald’s more durable demand-side category, not an economies-of-scale fortress.

1. Network effects (real, multi-sided). The multiplayer canvas is worked on by designers and the ~two-thirds non-designers (PMs, engineers, marketers, execs). A Figma file is the shared workspace of an entire product org; value rises with each added collaborator. Design systems and shared libraries become the org’s single source of truth, and Dev Mode’s MCP makes that source-of-truth the structured blueprint feeding code. This is a genuine within-account network effect, not merely a large user base.
2. Switching costs (org-wide). Teams invest deeply in design systems (reusable components, tokens, conventions). Ripping these out, retraining muscle memory, and rebuilding the 250k+ community resources/10k plugins is costly. Standardization on Figma equals high captivity.
3. Scale / ecosystem. Largest user base + plugin ecosystem + the “system of record” brand. Modest by Greenwald standards — there is no high-fixed-cost moat like a chip fab — and the supply-side cost advantage is weak (foundation models are third-party, available to all).
4. Patents are not a barrier. Only 15 issued US patents (+126 pending). (FACT — S-1 IP section.)

Pressure-test the network effect. It is real and currently strengthening: NDR 136% → 139%, paid-customer growth +37–48%, revenue re-accelerating, and the two-thirds-non-designer mix means the collaboration loop is widening to the whole org. But it is an intra-organization moat, not an industry-wide one — there is no consumer-style “all my friends are here” lock-in across companies; a competitor only needs to win one organization’s adoption decision, not overcome a global network. And the captivity rests on the premise that human, hand-crafted, multiplayer design work remains the dominant workflow.

The AI disruption debate — the entire thesis.

Bear (moat dissolving). Generative AI (“vibe coding”) routes around the design layer. Tools that go prompt → UI → code (Lovable, v0, Cursor, Bolt, Replit, Google Stitch, Anthropic/OpenAI) make the visual design artifact a generated, disposable intermediate rather than a hand-crafted source of truth. If output is generated, orgs need fewer designer seats and may never build or maintain a Figma design system. Figma’s own 10-K risk factors articulate this precisely: customers may “purchase from vendors that provide end-to-end design-to-code or prompt-to-product experiences”; organizations “may reduce the number of designers, developers, or collaborators… reducing the number of seats purchased or renewed even if output increases”; and — most damaging — “switching costs may decline as AI tools improve interoperability, automated conversion of files, or automated recreation of design systems across platforms.” That last clause is Figma conceding the central pillar of its switching-cost moat is AI-erodable. The margin evidence is already visible: Q1-26 cost of revenue +253% YoY (+$33.7M AI inference/hosting), gross margin 91.5% (Q1-25) → 79.4% (Q1-26), and a binding $50M minimum AI-compute purchase commitment (May-2026). (FACT — 10-Q Q1-26; 10-K risk factors.)

Bull (moat strengthening). “When execution is cheap, taste/judgment/craft is the scarce edge.” Figma positions as the human-plus-agent control surface: the canvas where humans and AI agents co-work, anchored by deep product context (design systems, tokens, intent) that ungrounded LLMs lack. The MCP server makes Figma the structured source of truth that feeds agentic coding — i.e., it benefits from vibe-coding tools rather than only competing with them. Live evidence: NDR re-accelerating to 139%, revenue re-accelerating to +46%, new AI surfaces (Make/Weave/MCP) expanding consumption, and ~60% of >$100K-ARR customers using Figma Make weekly (up from >50%). Management’s framing (Q1-26 call): a Google product team can “only get to a level of detail that we want in Figma — that’s not possible with vibe coding.”

Weighing it. The leading indicators (NDR, paid-customer growth, revenue re-accel, gross retention) say the moat is intact and the flywheel still widening — there is no evidence yet of seat erosion or churn to prompt-to-code tools. But the moat’s composition is shifting from a fortress (multiplayer + design-system lock-in) toward a contested control-surface bet whose durability depends on an unproven proposition (that taste/orchestration stays human-and-Figma-centric while execution commoditizes). The economics are already eroding at the gross-margin line, and Figma’s own filing concedes the switching-cost pillar is AI-erodable. The honest read: a moat under active assault, defended so far by superior product velocity (Greenwald’s emulable “operational effectiveness”) layered on a genuine-but-narrowing structural moat.

Direct competitor comparison. Adobe — XD is dead; Adobe lost the head-to-head and tried to buy Figma; residual threat is Firefly/Express AI plus Creative Cloud budget bundling, but Adobe is not winning the collaborative-design workflow. Canva — mass-market, AI-heavy, moving up-market; the most credible volume threat to lower/mid tiers (Buzz/Sites/Content), less so to enterprise product design; asymmetric (Canva attacks marketing/content; Figma’s core is product/UX). Sketch — legacy, marginalized. Microsoft — bundling risk (Designer/Copilot/Power Platform) into enterprise seats, a budget/consolidation threat more than a product-parity one. AI-native entrants (the real risk) — Lovable, v0, Cursor, Bolt, Replit, Google Stitch, Anthropic/OpenAI: these don’t compete on Figma’s canvas, they try to make the canvas optional. That is the disintermediation vector.

Verdict (Competitive Position): Figma has a durable advantage today — a real demand-side moat (network effects + design-system switching costs) that produced the textbook Greenwald signature of a winner: dominant share, 136–139% NDR, every prior incumbent displaced, and a $20B + $1B Adobe validation. It is not a crowded market with weak differentiation in the legacy workflow. But it is a moat under live dissolution pressure: the supply side has turned hostile, the cost advantage is nil, gross margin is compressing ~12 pts/yr on inference, and the company’s own 10-K concedes seats may shrink “even if output increases.” The verdict is a genuine, currently-intact moat whose durability is the entire investment question — strengthening on today’s KPIs, structurally more contestable than at any point in its history.

5. Growth History and Forward Opportunities

Historical growth — fast and essentially all organic. Revenue: FY23 $504.9M → FY24 $749.0M (+48.3%) → FY25 $1,055.8M (+41.0%). The 2025 tuck-in acquisitions (Weavy, Payload, Modyfi) were acqui-hires for tech/talent, not revenue — so growth is ~100% organic. (FACT — 10-K.) Most striking, quarterly YoY growth is accelerating into the deceleration narrative: Q3-25 +38% → Q4-25 +40% → Q1-26 +46% ($333M, revenue +$105M YoY). The re-acceleration is real, not a base effect. (FACT — 10-Q Q1-26 MD&A.)

The guide bridge. Q2-26 guidance is ~+40%; the FY26 guide of $1.422–1.428B (~+35%) was raised $55M. The full-year guide implies a 2H deceleration off the Q1 +46% — consistent with management not expecting +46% to persist, and embedding AI-credit-monetization uncertainty and tougher comps. The ~$369M of incremental FY26 revenue is bridged by: (1) seat expansion within existing orgs (the NDR engine, which carries the bulk), (2) new paid-customer adds (~690K, +54% YoY), (3) up-market $100K+ ARR mix, and (4) the new AI-credit consumption S-curve (first full quarter Q2-26) — with new logos and AI credits the swing factors between the top and bottom of guide.

What is driving the re-acceleration. NDR is 139% (a 2-year high, up from 132% a year prior — expansion accelerating, not just steady). The up-market cohorts are the fastest-growing: >$10K ARR customers +37%, >$100K ARR +48%. International ($564M, 53.4% of revenue) is growing faster than the US. And the product surface has expanded the buyer base beyond designers — Dev Mode/MCP (engineers), FigJam (PMs), Slides (whole org), Make/Sites/Buzz (developers + marketers), Weave (creative/AI) — structurally supporting more seats per org, which is the NDR mechanism. Pro-team conversions (a leading indicator) were up >150% YoY in Q1-26.

Forward vectors (in rough order of optionality).

1. AI-credit consumption monetization — the only usage-based line and the partial hedge against seat compression; enforcement began Mar-2026, Q2-26 is the first full quarter. Pro teams buying AI add-ons already spend ~3x those that don’t. Open question: will credit pricing stick without churning users, or does enforcing limits trigger downgrade/backlash? One quarter of data at most by the next print.
2. Figma Make / Weave / Sites / Buzz — extend the monetizable surface to developers and marketers.
3. Enterprise up-market — the $100K+ cohort (+48%) is the durable, stickiest, least-AI-disruptable revenue.
4. International — 53%+ of revenue and outpacing the US; long, under-penetrated runway.
5. Dev Mode + MCP — pulls engineers into the paid surface, expanding seats per org.

The counter-argument, taken seriously. The re-acceleration could be partly a post-IPO sugar-high: the March-2025 pricing/packaging overhaul, pent-up enterprise budget, and IPO-driven brand salience can flatter near-term cohorts, and the FY26 guide’s implied 2H decel is consistent with that. The structural risk is the existential one (see Risk Analysis) — Figma’s own 10-K warns AI may reduce seats “even if output increases,” and a seat-based model is exposed to AI labor compression; the AI-credit S-curve that hedges it has ~one quarter of evidence.

Verdict (Growth): High-quality growth on the evidence to date — organic, broad-based (no customer >10%), NDR-driven (139%, rising), up-market-mixing, internationally diversified, and re-accelerating. The caveat is forward, not historical: the durability of seat-based ARR in an AI-labor-compression world is unresolved, and the hedge (credit monetization) is one quarter old. Quality of growth-to-date: high. Confidence in its persistence past ~12 months: medium.

6. Financial Quality

Revenue and retention. Covered above — $1,056M FY25 (+41%), $333M Q1-26 (+46%), NDR 139%, RPO $682.3M at 3/31/26 (+5% q/q), deferred revenue $628M. Unit economics (expansion) are genuinely elite; 139% NDR is top-decile SaaS. (FACT — 10-K/10-Q.)

Gross margin — the compression story. GAAP gross margin: 91.2% (FY23) → 88.3% (FY24) → 82.4% (FY25), and 79.4% in Q1-26 (cost of revenue jumped to 20.6% of revenue from 8.5% a year earlier). Cost of revenue grew +97% in FY24 and +112% in FY25 — outpacing revenue — driven by AI/generative-AI inference compute and third-party AI-partner hosting. Non-cancelable hosting/AI-infra purchase commitments now total $498.3M through 2030 (vs a ~$50M-class figure pre-IPO). (FACT — 10-K Note 11; 10-Q.) This is the central quality risk to the software multiple: AI features carry variable inference COGS that scale with usage, structurally capping gross margin below the classic 85–90% software range. Management declines to commit to a floor, instead emphasizing “gross profit dollars” and its levers (model routing by task complexity, model-agnostic provider arbitrage, first-party models trained on Figma’s design corpus). Open question: does steady-state GM settle in the high-70s (AI-heavy) or recover toward the low-80s as inference costs fall?

SBC — the quality-of-earnings crux. FY25 total SBC was $1,364M = 129% of revenue (CoR $51M / R&D $698M / S&M $219M / G&A $397M). This is not run-rate: it is the IPO RSU liquidity-event catch-up — the August 2025 IPO satisfied a performance/liquidity vesting condition on a large RSU pool, front-loading expense (the biggest slug landed in Q3-25, where the operating loss was −$1,137M). FY24’s $948M of SBC was similarly dominated by a one-time May-2024 RSU Release ($801M) and tender-offer options ($88M). The honest steady-state read is Q1-26 SBC of $169.0M ≈ 50.7% of revenue (~$676M annualized, ~46–50% of FY26 revenue). So SBC has fallen from 129% to ~50%, and management guides it lower still — but ~50% is extraordinary versus mature SaaS (8.5% at CRM, 14.7% at NOW, ~26% at the high-SBC end with TEAM). Even a “normalized” ~45–50% means GAAP profitability is years out and dilution stays elevated. (FACT — 10-K stmt-of-ops note; 10-Q.)

“FCF” quality and owner-economic FCF. FY25 CFO was +$250.7M; FCF (after ~$8M total capex/cap-SW) was +$242.7M (23% margin), asset-light. Q1-26 FCF was $88.6M (27% margin) after a one-time $56.1M annual-bonus payout (a 17-pt drag). The crux: headline FCF is real cash but flatters the economics, because the largest non-cash add-back is SBC. Owner-economic FCF (FCF minus SBC) was negative ~$1,121M in FY25, and ~−$320M annualized on the Q1-26 run-rate. On an owner-economics basis — treating SBC as the real cost it is — Figma does not yet generate cash for shareholders; the headline FCF is substantially a transfer of value to employees via dilution. This turns positive only over a multi-year path as SBC normalizes and revenue grows. This is the single most important QoE caveat in the report. (FACT/INTERPRETATION — 10-K non-GAAP recon; 10-Q.)

Dilution. Shares outstanding went ~492.7M (Q3-25) → 513M (12/31/25) → 526.2M (3/31/26) → 528.4M (5/11/26) — roughly 6–7%/yr gross share growth from RSU vesting/options/ESPP/M&A stock, with no buyback to offset. The 2025 Equity Plan carries a 5% automatic annual evergreen increase to the reserve, institutionalizing ongoing dilution. (FACT — 10-Q; DEF 14A.)

Profitability frame. GAAP ROE/ROIC are not meaningful (losses are SBC-driven; FY25 operating margin −122%). The right SaaS frame: non-GAAP operating margin was 16% in Q1-26 (FY26 guide ~9% on the year), FCF margin 23–27%, and Rule of 40 ≈ 62 (46% growth + 16% non-GAAP op margin) — comfortably above 40. On the metrics the market uses for hyper-growth SaaS, Figma screens excellently; the GAAP losses and the owner-FCF gap are the skeptic’s counterweight.

Balance sheet. At 3/31/26: ~$1.65B cash + marketable securities (including a small, idiosyncratic ~$57M Bitcoin-ETF / ~$15M BTC treasury position), no funded debt (only ~$56M operating-lease liabilities), a $500M undrawn revolver, current ratio 2.5x, positive tangible equity ~$1.34B. Net cash ~$1.6B. A genuine fortress with ample runway. (FACT — 10-K Notes 3/7/11; 10-Q.)

Verdict (Financial Quality): Economics do improve with scale on the SaaS frame — 139% NDR, Rule-of-40 in the low-60s, 23–27% FCF margin, asset-light, fortress net cash. But two real flaws temper it: (1) gross margin is structurally compressing (91% → 79%) under AI-inference COGS with $0.5B of committed hosting spend, capping the software-multiple thesis; and (2) SBC, even past the IPO catch-up, runs ~50% of revenue, so owner-economic FCF is still negative (~−$320M run-rate) and dilution is ~6–7%/yr. The business is genuinely high-quality and improving — but headline GAAP losses overstate the problem while headline FCF overstates the cash actually accruing to owners. The truth is in between, and improving, but not yet a cash machine for shareholders.

7. Capital Allocation

The IPO was mostly a liquidity event, not a growth raise. At the August 2025 IPO ($33/sh), Figma issued only 12.5M primary shares (net proceeds ~$393M to the company), while selling stockholders sold 30.0M secondary shares (~71% of the 42.5M total) and received the proceeds; 246.0M preferred shares converted to Class A. The primary proceeds largely repaid a revolver drawn to fund RSU tax withholding on IPO-vesting RSUs. (FACT — 10-K MD&A; 424B4.) This followed a 2024 tender offer that cashed out $566.7M from employees/investors at $23.19/sh. Interpretation: Figma raised only $393M into a company already sitting on >$1.5B and generating cash — the IPO and the prior tender were primarily liquidity programs for pre-IPO holders and employees, not growth capital.

The $1B Adobe breakup fee. Received 12/20/23, booked in other income (it is the reason FY23 was GAAP-profitable). Figma has never paid a dividend; the $1B was retained on the balance sheet (underpinning the ~$1.6B net-cash position) and partly funded the 2024 employee/investor tender. Interpretation: retaining the fee for balance-sheet strength and employee liquidity is defensible, but no value was returned to (future public) shareholders, nor is any expected.

M&A — small, sensible tuck-ins. FY25 cash for business combinations was $58.3M across two deals: Weavy Inc. (10/3/25, AI media editing → Figma Weave; ~$85.3M total consideration, $76.9M goodwill, plus $43.8M of retention stock) and Payload CMS (4/17/25, headless CMS into Sites/Make; ~$10.4M, plus $22.2M retention stock), with Modyfi an additional acqui-hire. Interpretation: classic capability-buys, not revenue-buys — reasonable, immaterial individually, but heavy on retention stock (more dilution). Not empire-building.

R&D / S&M intensity. FY25 GAAP opex ratios look alarming (R&D 98% of revenue, S&M 55%, G&A 53%) but are SBC artifacts. Ex-SBC: R&D ~31% of revenue (product-heavy, founder-led, appropriate for an AI arms race), S&M ~34% (efficient for 41% growth with 139% NDR — product-led growth lowers CAC), G&A ~15%. The ex-SBC cost structure is rational.

Buybacks / dividends. None, and none expected. Appropriate for a growth company, but it means per-share value depends entirely on outrunning the ~6–7% dilution.

Comp structure — the alignment red flag. CEO Dylan Field’s 2025 total compensation was $864.4M, driven by an $861.9M stock award (a 2025 “CEO Service Award” plus a stock-price-based award; grant-date FV ~$27.45/sh, service- and stock-price-vested); CTO Kris Rasmussen’s 2025 total was $175.0M. The annual cash bonus is tied to “company performance metrics determined by the Compensation Committee,” which are not disclosed as ROIC/FCF; equity is service- and stock-price-based. Interpretation: incentives are growth/stock-price-linked, not ROIC/owner-FCF-linked — typical of founder-led hyper-growth but not the ROIC-aligned comp a long-term owner would prefer. The ~$862M founder mega-grant aligns Field to the share price but adds materially to SBC/dilution and is a governance demerit on magnitude. (FACT — DEF 14A Summary Comp Table.)

Dual-class control. Class B carries 15 votes/share. As of 3/31/26, Field directly owns Class B equal to 48.5% of total voting power; with voting proxies he controls 98.1% of Class B = 72.3% of total voting power on a ~10–13% economic stake. Figma is a NYSE “controlled company,” exempt from certain board-independence requirements; all officers/directors as a group hold 77.5% of votes. Interpretation: Field has absolute, entrenched control; public Class A holders have negligible governance recourse. Standard founder-control risk, mitigated only by his track record.

Insider behavior. The local corpus did not download the Form 4/3/144 bodies, so transaction-by-transaction sale figures could not be precisely quantified (an EDGAR pull would be required for exact dollars). But the structure is clear: insider/VC selling is substantial by design — 30.0M secondary shares in the IPO (~$990M), the prior 2024 tender ($566.7M at $23.19), and 82 post-IPO Form 144s as lockups expired. Large VC holders remain (Index 14.1%, Greylock 13.2%, Kleiner 10.7%, Sequoia 7.3%). There are no open-market purchases (code P) and none would be expected. Interpretation: the insider tape is one-directional selling — normal post-IPO VC monetization, but not a bullish conviction signal, and there is no insider-buying support under the stock.

Verdict (Capital Allocation): Mixed, lean cautious. Positives: asset-light, disciplined small tuck-ins (no empire-building), retained the $1B Adobe fee into a fortress balance sheet, efficient product-led GTM. Negatives that matter: (1) the IPO was ~71% secondary — a liquidity event, not growth capital; (2) no ROIC/FCF-linked incentives; (3) a staggering ~$862M 2025 CEO equity grant plus a 5% evergreen plan institutionalize heavy dilution; (4) entrenched 72.3% founder voting control with controlled-company carve-outs. Capital allocation is not destructive, but incentive alignment falls short of a strict owner-economics standard, and the dilution/control structure tilts per-share economics toward insiders.

8. Changes and Headwinds — Last Two Years

A recent-events timeline (oldest → newest):

• Dec 17, 2023 — Adobe deal collapse. Figma and Adobe mutually terminated the ~$20B merger on “no clear path” to EU/UK/US antitrust approval; Adobe paid a $1.0B breakup fee. Interpretation: vindicated Figma as standalone-fundable and an independent competitor; the cash fortified the pre-IPO balance sheet. Adobe subsequently sunset XD. (FACT — 10-K.)
• March 2025 — pricing/packaging overhaul. “Significant changes to pricing, packaging, and billing models” plus AI credits introduced across all seats — the rails for the seat/ARR re-acceleration and for later credit monetization. (FACT — 10-Q Q1-26 MD&A.)
• 2025 — product expansion (Config-era launches): Make (prompt-to-app GA), Sites + Buzz, Dev Mode, Slides, MCP server, AI Assistant (alpha) — TAM expansion from designers to developers, marketers, and the whole product org.
• 2025 — M&A (all tuck-in/acqui-hire, organic-growth-neutral): Payload CMS, Modyfi, and Weavy (→ Figma Weave, 10/3/25). Capability-buys, not revenue-buys.
• July/August 2025 — IPO (NYSE: FIG). Priced ~$33; a massive first-day pop to ~$122–143 intraday, valuing the company near ~$59–60B at peak. Generated the large IPO-related RSU SBC charge behind the FY25 ~$1.3B GAAP net loss. (FACT — 10-K; Benzinga 2026-06-14.)
• March 18, 2026 — AI credit monetization goes live. Began enforcing AI-credit limits plus add-ons/pay-as-you-go overage billing; Q2-26 is the first full quarter — the new usage-based revenue lever. (FACT — 10-Q Q1-26.)
• Through June 2026 — the “$50B wipeout.” Stock fell from the ~$143 post-IPO high to ~$18; market cap ~$59B → ~$9B (≈−84.5% off the $122 ATH). Interpretation: the driver is a sector-wide “AI subsumes software” de-rating (Adobe/Intuit/ServiceNow all slumped on the same fear), amplified by Figma’s seat-based exposure, the Q1-26 SBC-driven GAAP loss, and post-IPO multiple compression off an extreme first-day valuation — not a fundamentals collapse; revenue accelerated throughout. (FACT/INTERP — Benzinga 2026-06-14; baseline.)

Verdict (Changes/Headwinds): The operating changes strengthen the thesis (pricing overhaul + product TAM expansion + credit-monetization rails + a balance-sheet-fortifying $1B Adobe fee + vindicated independence). The market changes (IPO pop → 84% collapse, “$50B wipeout”) reflect valuation/sentiment, not deteriorating fundamentals — a price-context reset, not a business-quality reset. The divergence between accelerating revenue and a −84% stock is the central tension this memo must adjudicate.

9. Risk Analysis

The matrix is dominated by one risk (AI seat-compression) that is simultaneously the bull’s and the bear’s pivot. Risks #2–#3 are its financial expression; everything else is conventional and survivable.

Key matrix read. Risks #2/#3 (margin + monetization) are the financial expression of #1 — they determine whether AI is a net tailwind (consumption capture) or a net headwind (margin + seat erosion). Competition, valuation, SBC, lockup, and macro are conventional and survivable. Concentration and balance-sheet/total-loss risk are genuinely low. The whole risk picture collapses to a single axis: does AI add to or subtract from the seat pool?

10. Valuation Discussion (Embedded Expectations)

No price target, no buy/sell — embedded expectations and scenarios only.

Where the multiple sits. At $18.88, ~527M shares ≈ $10.0B market cap; net cash ~$1.6B → EV ~$8.4B on ~$1.16B TTM sales ≈ ~7x EV/sales. The de-rate trajectory is the load-bearing fact: EV/TTM-sales went 76x (peak, $122) → 25.8x (Q3-25) → 16.6x (Q4-25) → 8.2x (Q1-26) → ~7x now. Headline P/FCF is ~39x, but owner-FCF (FCF − SBC) is negative, so P/owner-FCF is not meaningful while the SBC normalizes. EV/gross-profit is ~9x — relevant because AI inference COGS is the bear’s margin lever.

Comp set (ROIC TTM multiples).

Read. At ~7x FIG trades above ADBE (4.3x), TEAM (2.9x), and CRM (4.5x), and roughly in line with ADSK (6.7x) and NOW (7.6x) — but it grows 35–46% versus their 10–20%, so on a growth-adjusted basis ~7x is not obviously expensive. The catch is profit quality: ADBE/ADSK/NOW/CRM are GAAP-profitable with real owner-FCF, whereas FIG’s “FCF” is an SBC add-back artifact and owner-FCF is negative. FIG is therefore “cheap on growth, rich on real profit.”

Embedded expectations (reverse-DCF). To justify the ~$8.4B EV at a terminal ~25x FCF on a normalized ~25% true-FCF margin, the business needs roughly $1.35B of normalized FCF eventually → ~$5.4B of revenue → a ~4.6x scaling from $1.16B TTM. Reaching ~$5.4B in ~6 years implies a ~29% revenue CAGR while SBC roughly halves from ~50% to ~25% of revenue (so reported FCF becomes real owner-FCF). Interpretation: at ~7x the market is underwriting roughly high-20s%/low-30s% revenue CAGR for 5–6 years AND SBC normalization to deliver ~25% true FCF margin. That is a demanding-but-not-heroic bar for a 139%-NDR, 35–46% grower — the de-rate from 76x has removed most of the bubble premium. Critically, the bar is not priced for AI to be a tailwind; it is priced for growth to stay healthy and for the dilution to actually normalize. The single biggest embedded assumption is SBC normalization, not the growth rate. What the market is pricing correctly: that the 76x IPO multiple was untethered, and that ~7x is a reasonable midpoint. What it may be pricing incorrectly (both directions): it gives little credit to credit/usage monetization as a new S-curve (bull), but it may be too generous on SBC normalizing and on whether 80% gross margins survive inference COGS (bear).

Scenario analysis (per-share zones; ~527M shares; net cash ≈ $3/sh).

• Bear (~$12–16) — AI dissolves the design seat: growth fades to mid-teens within 2 years; GM 80% → ~70% on inference COGS; SBC stays ~40%+, owner-FCF stays negative; multiple compresses to 3–4x sales (the ADBE/TEAM zone) on ~$1.6–1.7B forward sales. EV ~$5–7B + $1.6B cash → ~$12–16/sh.
• Base (~$20–26) — 30%+ growth holds 2–3 years then decelerates to ~20%; GM holds ~78–80%; SBC normalizes toward ~30%, owner-FCF turns positive in ~2–3 years; multiple ~6–7x forward sales on ~$1.42B FY26 / ~$1.8B FY27. EV ~$9–12B + $1.6B → ~$20–26/sh. (Brackets the current $18.88.)
• Bull (~$33–45) — AI is a tailwind: agent/credit usage compounds on seats, growth holds 35%+, GM stable, SBC normalizes to ~25% delivering ~25–30% true FCF margin; re-rate to 10–12x forward sales. EV ~$16–22B + $1.6B → ~$33–45/sh — i.e., back to ~the IPO price, still well below the $122 ATH and below the $20B Adobe deal value.

The current $18.88 sits at the low end of the base zone, between bear and base — the tape is pricing closer to “growth fades / SBC never normalizes” than to “base case holds.”

11. Variant Perception

Consensus. “A great product and a real franchise, but an over-hyped IPO that got AI-de-rated; the design layer is the most AI-exposed surface in software, so even a 35–46% grower deserves only a ~7x multiple until it proves AI is additive, and until SBC stops eating all the profit.” Post-IPO orphan: −84.5% off ATH, below the IPO price, the “$50B wipeout” the dominant narrative.

Strongest bull case. Figma is the human-plus-agent control surface for software creation — the canvas where humans and AI agents converge, anchored by design-system context AI lacks. Design is the scarce, defensible edge AI does not commoditize; it amplifies the need for a shared source of truth. The 139% NDR proves stickiness even through the de-rate — customers are spending more, not churning. Credit/usage monetization is a new S-curve layered on seats, for which the market gives ~zero credit at ~7x. And the de-rate has done the work: ~7x for a 35%+ grower with 80% GM is reasonable versus ADSK 6.7x / NOW 7.6x at half the growth, with downside cushioned by ~$1.6B net cash and a brand Adobe valued at ~$20B.

Strongest bear case. AI dissolves the design layer and compresses seats — if agents generate UI directly, you need fewer Figma seats, and the seat-based model is structurally challenged; growth fades to teens. Gross-margin erosion is structural (inference pushes 80% → 70%), permanently lowering terminal FCF margin and the multiple it deserves. ~50% SBC means there is no real profit — reported FCF is an add-back illusion, owner-FCF is ~−$320M run-rate, and ~6–7%/yr dilution is a permanent leak. Post-lockup VC overhang hits the thin float, and dual-class control (Field 72.3% of votes) leaves public holders as price-takers. And the ~$20B Adobe price was the ceiling, not a floor — set pre-AI-fear; today’s strategic value may be lower.

The 3–5 assumptions that decide it (and what falsifies each).

1. Does AI add to or subtract from seat demand? Falsify bear: NDR stays >130% and seat count keeps growing alongside credit revenue for 3–4 quarters. Falsify bull: NDR rolls below ~115% / net seat adds stall.
2. Does gross margin hold ~78–80% as AI scales? Falsify bear: GM stable/expands as credits scale. Falsify bull: GM drifts below ~75%.
3. Does SBC normalize toward ~25% (owner-FCF turns positive)? Falsify bull: SBC stays >40% and dilution ~6–7% past FY27. Falsify bear: SBC drops to ~30% within 2 years and reported FCF starts to equal owner-FCF.
4. Does growth hold 30%+ for 2–3 years? Falsify bear: FY27 guide still ≥30%. Falsify bull: deceleration below 25% with no credit offset.
5. Does credit/usage monetization become material? Falsify bear: company discloses double-digit-% of revenue from usage/credits, growing fast. Falsify bull: credit revenue stays immaterial/undisclosed.

Factor-positioning read (heavily caveated). FactorsToday stock-loadings and leaderboard are empty — the IPO has <252 trading days, so there is no clean factor identity yet; treat FIG as an IPO orphan with no factor home. The measured beta of 0.59 is a short-history, noisy estimate dominated by the idiosyncratic post-IPO crash — do not treat FIG as a low-beta name. The only clean signal is relative strength, and it is decisively negative (rs_peak −84.5%, rs_6m −47.5%, rs_12m −42.8%, alpha −0.59) — a textbook falling knife on relative strength; the recent $16.86 → $27 bounce already gave back to ~$18.88. AZI’s own-history valuation percentile is null (no multi-year range), so the de-rate must be judged against the SaaS comp set, not FIG’s own history. Take-away: the tape says “abandoned post-IPO falling knife with no factor identity,” not a quality-momentum darling and not a clean low-vol value name; consensus leans bear (priced at the low end of base). The contrarian variant is that ~7x already discounts much of the AI-seat fear — so the asymmetry depends entirely on assumptions 1–3 resolving favorably.

12. Fact vs. Interpretation

13. Open Questions

1. Does SBC actually normalize from ~50% to ~25–30% of revenue, or is the IPO “catch-up” framing masking a structurally high run-rate? (Determines whether base-case owner-FCF ever turns positive.)
2. Is AI-credit/consumption revenue material yet, and growing? Company disclosure needed — Q2-26 is the first full quarter of enforcement.
3. Where does steady-state gross margin settle — high-70s (AI-heavy) or recovering toward low-80s as inference costs fall and first-party models scale?
4. Is the +46% Q1 a durable trajectory or a pricing-change + post-IPO pull-forward that the +35% FY guide implicitly walks down?
5. How much post-lockup VC/insider float overhang remains, and is it pressuring the tape independent of fundamentals? (Exact Form-4/144 dollars not quantified from the local corpus — would require an EDGAR pull.)
6. What are the 2026-04-15 and 2026-06-04 8-Ks (governance/guidance specifics)? Not fully read from the local set.

14. What Must Be True

Bull case — what must be true (and its falsification test). Figma is the durable human-plus-agent control surface: AI is additive, expanding seats and consumption rather than compressing them; gross margin holds ~78–80% as credit monetization offsets inference COGS; SBC normalizes toward ~25–30% so owner-FCF turns durably positive; and growth holds 30%+ for several years, re-rating the multiple toward 10x+. Falsification test: within 2–3 quarters, NDR rolls below ~115%, net seat adds stall, gross margin drifts below ~75%, or SBC stays above ~40% past FY27. Any one breaks the bull case — it would confirm AI is compressing the seat pool and/or that there is no real profit underneath.

Bear case — what must be true (and its falsification test). Generative AI routes around the design layer: orgs need fewer seats “even if output increases,” design systems are AI-recreatable (eroding switching costs), inference structurally caps margins in the low-70s, and ~50% SBC plus dilution means the headline FCF never becomes owner cash — so growth fades to the teens and the multiple compresses to 3–4x sales. Falsification test: NDR holds >130% with growing seat counts for 3–4 quarters, AI-credit/consumption revenue is disclosed as a materially-growing double-digit-% of revenue, gross margin stabilizes ≥78%, and FY27 guidance is still ≥30%. That combination would demonstrate AI is a net tailwind and the seat-compression fear is mis-priced.

The current price (~$18.88, low end of base) implies the market leans toward the bear’s seat-compression read while crediting little of either the credit-monetization upside or the catastrophic downside. The next 2–4 earnings prints — the first full quarters of AI-credit monetization — are the adjudicating evidence for both falsification tests.

15. Source Appendix

See the separate Source Appendix (Appendix B in the combined report) for the full, dated citation list. Primary sources relied upon:

• Figma, Inc. Form 10-K (FY2025), filed 2026-02-18 (period 2025-12-31) — business, risk factors, MD&A, financial statements, notes (revenue, SBC, commitments, M&A, dividend policy).
• Figma, Inc. Form 10-Q (Q1-2026), filed 2026-05-14 (period 2026-03-31) — Q1 results, gross-margin/COGS detail, KPIs, AI-credit monetization.
• Figma, Inc. Forms 10-Q (Q2/Q3-2025), filed 2025-09-03 / 2025-11-05.
• Figma, Inc. Form S-1 / S-1/A and 424B4 prospectus (2025-07) — market opportunity/TAM, cohort metrics, IPO structure (primary vs secondary).
• Figma, Inc. DEF 14A (2026) — executive compensation, dual-class voting power, beneficial ownership.
• Figma Q1-2026 earnings call transcript (2026-05-14, via ROIC.ai) — management commentary on AI monetization, gross margin, competition, guidance.
• AZI price history (5-yr OHLCV CSV) — price arc, EMAs, beta.
• ROIC.ai — three statements, enterprise value, valuation multiples (cross-checked to filings).
• FactorsToday — stock-info (relative strength, beta); loadings/leaderboard empty (IPO <252-day history).
• Comparable public companies — Adobe (ADBE), Atlassian (TEAM), Autodesk (ADSK), Salesforce (CRM), ServiceNow (NOW): public filings and disclosures used for comp multiples, SBC benchmarks, and AI-disruption framing.
• Benzinga (2026-06-14) — “After Buzzy IPO, Figma Stock Just Had a $50 Billion Wipeout.”

The analytical body of this article (Sections 1–15) carries no investment recommendation and no price target; the only opinion expressed is the clearly-labeled Claude’s Take block above, which is the author’s own independent view and general information only — not investment advice.

APPENDIX A — Standard Diligence Questionnaire

Figma, Inc. (NYSE: FIG) — as of 2026-06-20

Supplemental diligence record. Fact / Interpretation / Assumption labels applied where material. Where a question does not map to the business model, the correct analog is given.

General

What thoughtful questions have other investors asked about this company? The dominant question is binary and existential: does generative AI make Figma indispensable (the human-plus-agent “control surface,” design-as-the-scarce-edge) or obsolete (prompt-to-code “vibe coding” routes around the design layer and compresses seats)? Secondary lines: (a) is the +46% Q1-26 re-acceleration durable or a pricing-change/post-IPO sugar-high (FY26 guide implies 2H decel to +35%)?; (b) is the company actually profitable, or is “FCF-positive” an SBC mirage (owner-FCF negative)?; © where does gross margin settle as AI-inference COGS scales (91% → 79% in one year)?; (d) will AI-credit consumption monetization (live 3/18/26) stick without churning users?; (e) is the ~$20B Adobe price a floor or a pre-AI-era ceiling? Most investors agree the product is excellent; they disagree on terminal value.

Cyclicality & Earnings Nature

Are earnings at a cyclical high or low? Neither in the conventional sense — Figma is a hyper-growth software company with GAAP losses (SBC-driven), not cyclical earnings. Revenue is at an all-time high and accelerating (+46% Q1-26). If anything, margins are artificially depressed by the IPO-year SBC catch-up (FY25 SBC 129% of revenue), so reported GAAP earnings are at an artificial low. (FACT/INTERPRETATION.) Driven by external environment or internal actions? Predominantly internal — product-led growth, NDR-driven expansion, pricing/packaging overhaul (Mar-2025), new product surfaces. External overlays: enterprise IT-budget cycle and the sector-wide AI-software de-rate (which hit the stock, not the fundamentals). (INTERPRETATION.) How stable are revenues? Highly recurring per-seat subscription, NDR 139%, RPO $682M, deferred revenue $628M, no customer >10%. Among the more stable revenue bases in software — if the seat-based model survives AI. (FACT.) Outlook for products/services? Re-accelerating core, with new monetizable surfaces (Make/Weave/Sites/Buzz/MCP) and a new usage-based AI-credit line. (FACT for the launches; INTERPRETATION for trajectory.) How big is this market — growing/shrinking, domestic/international? Company/IDC TAM $33B today → 144M-person “software-design workforce” by 2029 (self-commissioned, pricing-derived — directional). 53.4% of revenue is already international and growing faster than the US. The bear caveat: AI could shrink the seat-based TAM even as software creation expands. (FACT-as-claim / ASSUMPTION.)

Business Quality & Competitive Moat

Is the industry getting more or less competitive? More. Capital is flooding AI software-creation (Lovable, v0, Cursor, Bolt, Replit, Google Stitch, Anthropic/OpenAI) — a Marathon capital-cycle boom; foundation models are off-the-shelf (no supply-side moat). (INTERPRETATION.) How profitable is the business (ROIC, ROE)? GAAP ROIC/ROE are not meaningful (SBC-driven losses). On a SaaS frame: Rule of 40 ≈ 62, FCF margin 23–27% (headline), non-GAAP op margin 16% (Q1-26). But owner-FCF (FCF − SBC) is negative ~$320M annualized — so true economic profitability to shareholders is not yet positive. (FACT/INTERPRETATION.) How profitable is the industry — competitors, barriers to entry? Legacy design SaaS: high-margin (75–91% GM), strong design-system captivity, oligopolistic — a good industry. The AI-redefined version: low barriers at the AI layer, hostile supply side. (INTERPRETATION.) Can the business be easily understood? Yes — a per-seat design-software subscription with an emerging usage layer. The terminal AI outcome is what is hard, not the model. (INTERPRETATION.) Can it be undermined by foreign low-cost labor? Not directly (software IP/network business). The analog threat is AI labor substitution — automating the design work itself. (INTERPRETATION.) Do brands matter? Yes — Figma is the “system of record” brand for product design; 95% of Fortune 500 used it. Brand + design-system standardization is a real switching-cost contributor. (FACT/INTERPRETATION.) Nature of competition / switching costs? Competition is shifting from feature-rivalry (vs XD/Sketch, won) to disintermediation (AI app-builders making the canvas optional). Switching costs are org-wide design-system lock-in — real today, but Figma’s own 10-K concedes AI may enable “automated recreation of design systems,” eroding them. (FACT/INTERPRETATION.)

Financial Condition & Balance Sheet

Assets not fully recognized on the balance sheet? The brand, the design-system network, and the data corpus (used to train first-party design models) are unrecognized intangible value. (INTERPRETATION.) Off-balance-sheet liabilities? $498M of non-cancelable hosting/AI-compute purchase commitments through 2030, plus operating leases — disclosed in the notes. (FACT.) How conservative is the accounting? Reasonably clean — revenue recognition standard SaaS; the complexity is in SBC and the IPO-related catch-up (well-disclosed). The idiosyncratic ~$57M Bitcoin-ETF / ~$15M BTC treasury position is a minor governance curiosity. (FACT/INTERPRETATION.) How CapEx-hungry is the business? Barely — capex ~$8M/yr on >$1B revenue (asset-light). The “capex equivalent” is AI-inference COGS (in cost of revenue) and R&D (heavily SBC). (FACT.)

Capital Allocation & Management

How much FCF, and how is it used? Headline FCF +$243M FY25 (23% margin), but owner-FCF negative. No dividends, no buybacks; cash retained on the balance sheet (~$1.6B net cash) and partly recycled into employee liquidity (2024 tender $566.7M) and small tuck-ins. (FACT.) Significant acquisitions recently? Only small acqui-hires: Weavy ($85M → Figma Weave), Payload CMS ($10M), Modyfi. Capability-buys, not revenue-buys, heavy on retention stock. (FACT.) Buying back shares? No — share count is rising ~6–7%/yr; a 5% automatic annual evergreen entrenches dilution. (FACT.) Issuing large amounts of stock to insiders? Yes — FY25 SBC $1,364M; a ~$862M 2025 CEO equity grant (founder mega-grant). (FACT.) Compensation policy of directors/management? Service- and stock-price-based equity; cash bonus on undisclosed “company performance metrics” — not ROIC/FCF-linked. Falls short of a strict owner-economics standard. (FACT/INTERPRETATION.) Motivations of management? Founder-CEO Dylan Field controls 72.3% of votes (15:1 dual-class) on ~10–13% economics — entrenched, mission/product-driven, aligned to share price via stock-price-vested equity but not to per-share owner economics. (FACT/INTERPRETATION.)

Valuation & Market Data

ADR, MLP, or K-1 issuer? No — US C-corp common stock, NYSE-listed, dual-class (Class A public, Class B super-voting). No K-1. (FACT.) Dividend policy? None; never paid; none expected. (FACT.) How profitable is the business? GAAP-unprofitable (SBC); cash-generative on headline FCF; not yet profitable on an owner-economics (post-SBC) basis. (FACT/INTERPRETATION.) Is net income diverging from cash from operations? Yes, massively and benignly — FY25 net loss −$1.25B vs CFO +$251M, the gap is non-cash SBC ($1.36B). This is the favorable direction of divergence (cash > earnings), but the SBC it adds back is a real dilution cost. (FACT/INTERPRETATION.)

Risks & Downside

What factors would cause the stock to decline? NDR rolling toward ~115%; net seat adds stalling (AI seat-compression confirmed); gross margin below ~75%; SBC stuck above ~40%; AI-credit monetization failing to materialize; growth decelerating below ~25%; post-lockup VC supply; broad software/AI de-rate continuing. (INTERPRETATION.) Risk of catastrophic loss? Low in the near term — net cash, no debt, accelerating revenue. The only catastrophic path is a multi-year structural AI displacement of the design layer (risk #1), which would be gradual, not sudden. (INTERPRETATION.) Chance of a total loss? Very low — fortress balance sheet, category-defining franchise, ~$1.6B net cash. Even the bear scenario is a lower multiple on a still-large revenue base, not a zero. (INTERPRETATION.)

Recent News & Events

Has the business environment changed recently? Yes — AI-credit monetization went live (3/18/26, first full quarter Q2-26); the sector-wide “AI subsumes software” de-rate drove the stock −84.5% off its ATH (“$50B wipeout,” Benzinga 6/14/26) even as revenue accelerated. (FACT.) Significant acquisitions? Weavy (10/3/25 → Figma Weave); Payload, Modyfi (2025). (FACT.) Change in accounting policies? None material; IPO-driven SBC recognition is the main FY25 distortion (well-disclosed). (FACT.) Recent changes — new markets, facilities, management? Product TAM expansion (Make/Sites/Buzz/Weave/Slides/MCP, 2025); March-2025 pricing/packaging overhaul; July-2025 IPO; CFO Praveer Melwani and co-founder/CEO Dylan Field in place; board slate stable. (FACT.)

APPENDIX B — Source Appendix

Figma, Inc. (NYSE: FIG) — Research Sources (as of 2026-06-20)

Primary sources before secondary; every non-obvious memo claim traces to one of these. Facts reconciled to SEC filings where the figure drives a verdict.

Primary — SEC filings (Figma, Inc., CIK 0001579878 — all available free on SEC EDGAR)

1. Form 10-K, FY2025 — filed 2026-02-18 (period 2025-12-31). Business description; risk factors (AI seat-compression, switching-cost erosion, margin pressure verbatim); MD&A; consolidated statements; Notes — revenue/RPO (Note 2), SBC composition (Note 13), purchase commitments $498.3M through 2030 (Note 11), M&A Weavy/Payload (Note 8), Adobe $1B fee / dividend policy, geographic revenue (US $491.5M / Intl $564.2M). URL: https://www.sec.gov/cgi-bin/browse-edgar?action=getcompany&CIK=0001579878&type=10-K
2. Form 10-Q, Q1-2026 — filed 2026-05-14 (period 2026-03-31). Q1 revenue $333.4M (+46%); cost of revenue +253% YoY / GM 79.4%; KPIs (NDR 139%, >$10K ARR 15,218, >$100K ARR 1,525); AI-credit monetization (enforced 3/18/26); $50M compute commitment; balance sheet ($1.65B cash+securities, 526.2M shares).
3. Form 10-Q, Q3-2025 — filed 2025-11-05 (period 2025-09-30). The IPO-SBC-catch-up quarter (op loss −$1,137M).
4. Form 10-Q, Q2-2025 — filed 2025-09-03 (period 2025-06-30). First quarterly report as a public company.
5. Form S-1 / S-1/A and 424B4 prospectus — 2025-07. Market opportunity/TAM ($33B; IDC 144M design workforce by 2029); cohort/KPI history; IPO structure (12.5M primary / 30.0M secondary at $33.00); IP (15 patents).
6. DEF 14A proxy statement — 2026. Executive compensation (Field 2025 total $864.4M incl. $861.9M stock award; Rasmussen $175.0M); dual-class voting (Class B 15 votes; Field 72.3% of total voting power; controlled-company status); 5% evergreen equity plan; beneficial ownership (Index 14.1%, Greylock 13.2%, Kleiner 10.7%, Sequoia 7.3%).
7. Form 8-K corpus (6 filings since IPO) — Earnings 8-Ks (2025-09-03, 2025-11-05, 2026-02-18, 2026-05-14) and governance/other items (2026-04-15, 2026-06-04 — not fully read).
8. Form 3/4/144 corpus — (135 Form 4s, 18 Form 3s, 82 Form 144s on EDGAR). Insider read derived from IPO/tender structure and public ownership tables; exact per-transaction sale dollars would require a detailed Form 4 review.

Primary — management commentary

9. Figma Q1-2026 earnings call transcript — 2026-05-14 (via ROIC.ai get_latest_earnings_call). CEO Dylan Field / CFO Praveer Melwani: AI-credit monetization early signals (75%+ of over-limit org/enterprise users kept consuming credits in April; Pro AI-add-on teams ~3x spend); gross-margin levers (model routing, model-agnostic, first-party models); FY26 guide raise ($1.422–1.428B revenue, $125–135M non-GAAP op income); competitive framing on Google Stitch/Anthropic/vibe-coding; NDR 139%; ~690K paid customers; Config (June) cadence. (Treated as hypothesis, validated against filings.)

Secondary — quantitative data services

10. AZI price history — https://azitrading.com/controls/download-data.php?t=FIG (5-yr OHLCV). Price arc: IPO $33 → $122 ATH (8/1/25) → $16.86 ATL (4/29/26) → $18.88 (6/18/26); EMAs; beta. AZI news feed (14 items) and valuation_index (null — IPO <1yr, no own-history percentile).
11. ROIC.ai — income statement, balance sheet, cash flow, enterprise value, valuation multiples (multi-period). EV ~$8.4–9.5B; EV/TTM-sales de-rate 76x → ~7x; SBC $1,364M FY25; CFO/FCF series. (Third-party aggregated; reconciled to filings; EDGAR/10-K primary where they differ.)
12. FactorsToday — stock-info (beta 0.59 [noisy/short-history], rs_peak −84.5%, rs_6m −47.5%, rs_12m −42.8%, alpha −0.59). Stock-loadings, leaderboard, and related-stocks empty (IPO <252-day history — documented universe gap).

Secondary — comparable public companies (for comps and framing)

13. Adobe (ADBE) — Figma’s near-acquirer and #1 incumbent; $20B deal blocked + $1B fee; Adobe ~4.3x EV/sales / ~10x P/E (cheapest in a decade) on the same AI fear; the “AI-disrupts-design-software” multiple-compression precedent.
14. Atlassian (TEAM) — Collaboration-SaaS NDR/Rule-of-40/SBC (26%) benchmarks; owner-FCF discipline.
15. Autodesk (ADSK) — Design-software industry structure; ~6.7x EV/sales; AI-seat-compression framing; CAD/BIM TAM.
16. Salesforce (CRM) / ServiceNow (NOW) — output/CRM_2026-06-10_full_report.md, Mature-SaaS SBC benchmarks (CRM 8.5%, NOW 14.7%); consumption-pricing pivot (NOW: 50% of Q1-26 net-new non-seat-based); EV/sales (CRM 4.5x, NOW 7.6x).

Secondary — news / trade press

17. Benzinga — “After Buzzy IPO, Figma Stock Just Had a $50 Billion Wipeout” (2026-06-14; AZI news id 402661). The abandonment narrative; ~$59B → ~$9B market-cap round-trip.
18. Benzinga / sector — software-sector rally/selloff items (AZI news ids incl. Snowflake-driven rally 5/28, payroll/bond-yield selloffs) — context for the macro/sector de-rate, not company-specific.

Frameworks applied

19. investment-research-frameworks skill — Greenwald & Kahn Competition Demystified (moat taxonomy: demand-captivity/network-effects classification; market-share-stability test; supply-side cost advantage = nil for off-the-shelf foundation models) and Chancellor/Marathon Capital Returns (capital-cycle boom in AI software-creation; supply-side erosion of the design profit pool).