Figma: The Platform Role Is Real, but Dilution and the Last Unlock Leave No Margin of Safety

Gross margin is also no longer the effortless 90%-plus story investors imprinted on during the Adobe-deal period. In Q1 2026, GAAP gross margin fell to 79% from 91% a year earlier because cost of revenue rose 253%, driven chiefly by AI-related hosting and broader paid-platform usage. For full-year 2025, gross margin fell to 82% from 88% in 2024, with management explicitly pointing to higher technical infrastructure and hosting costs related to AI. So the bear case is real. It has yet to reach revenue, showing up first in margin structure and in the market multiple, not yet in revenue.

This leaves the central bull-bear disagreement in a cleaner form. Bulls argue that Figma is becoming the control plane for product intent in an AI-native development workflow, and that AI widens the user base by pulling in developers, PMs, and non-designers who still need an enterprise-safe visual system of record. Bears argue that AI lowers the skill threshold for design and pushes value into code models, developer environments, and cheaper generation tools, leaving Figma with seat compression, higher inference costs, and a slower-growing collaboration shell. The evidence today favors the bull side on product relevance and the bear side on valuation discipline. Figma looks more durable than the market’s harshest AI-obsolescence narrative implies, but the stock does not yet offer a clean margin of safety once dilution, lock-up supply, and still-unproven terminal economics are included.

Figma is best described as a re-rated high-quality growth company still earning the right to its long-run multiple, rather than “cheap growth” or “AI roadkill.” The business quality is real. The market’s repricing from the post-IPO peak was mostly justified. The mistake now would be to stop at that conclusion and assume there is no opportunity left. There may be one, but it is more likely to emerge after the final supply overhang clears and after at least one more quarter proves that AI credits are a margin-aware expansion lever rather than a revenue sugar high with infrastructure drag. The stock is no longer absurd. It is also not yet obviously mispriced.

Qualitative portrait label: re-rating. The company still shows the hallmarks of high-quality compounding growth in product usage and expansion metrics. The stock’s capital-market journey since listing has mainly been a correction from a bubble opening print into a probation period, where investors want proof on AI durability, lock-up absorption, and post-SBC economics before paying back a premium multiple.

3. Company vertical history

2.1 Origins

Figma existed because the dominant design workflow was broken in a way that desktop incumbents barely noticed. Product design lived in heavy local files, versioning was messy, feedback loops were slow, and engineering handoff was often an afterthought. Dylan Field and Evan Wallace began working on a browser-native alternative in 2012, using the web as the operating assumption for collaboration itself rather than as a thin wrapper around desktop software. Field has repeatedly described the company’s early ambition as building design tools “on (and for) the web,” and the company states that it was co-founded in 2012 and is headquartered in San Francisco. The founding background matters because it explains the opening move: Figma attacked the collaboration and workflow problem that desktop tools had left behind, instead of competing feature-for-feature with Adobe’s broad creative suite.

That starting problem also explains the early business model. Figma was born as a collaborative interface-design platform with a freemium motion baked into the product, not as an all-purpose creative cloud. The free layer created habitual use, the team layer created collaboration, and the paid seats monetized professional workflows once organizations standardized around the tool. Over time, the model widened from design seats to a broader product-development platform. The current product lineup includes Figma Design, Dev Mode, FigJam, Figma Make, Figma Slides, Figma Draw, Figma Sites, Figma Buzz, Figma Motion, and Figma Weave. That is a different surface area from the original wedge, but the commercial logic is continuous: start from shared digital product work, then increase the number of roles that need to live inside the same system.

2.2 Birth node and listing path

Figma’s public-market path was unusual because the company effectively lived through an aborted exit before listing. Adobe agreed in 2022 to acquire Figma for $20 billion. The transaction was terminated in December 2023, and Adobe recorded a $1 billion termination fee paid to Figma. That fee mattered financially and symbolically: financially because it strengthened Figma’s balance sheet and distorted 2023 earnings; symbolically because it confirmed how strategically important Figma had become to an incumbent willing to pay a very full price before regulators intervened.

When Figma finally came public, it sold 36,937,080 Class A shares at $33, including 12,472,657 shares sold by the company and 24,464,423 shares sold by existing stockholders, with a 30-day over-allotment option on another 5,540,561 shares. Gross proceeds were about $1.22 billion. Reuters reported that the IPO valued the company at about $18.8 billion before trading and that the stock then exploded in the open market, closing at $115.50 on its first day and implying roughly a $56.3 billion market capitalization on day one, then closing at $122 the next session. That sequence is crucial. The IPO story sold to institutions was a large, fast-growing, AI-adjacent software platform. The public market’s first interpretation was much more feverish: a scarce, high-growth software IPO in a reopened market, and therefore something to grab at almost any multiple.

2.3 Stage division

The first stage was product validation. From the 2012 founding to the broader commercial launch phase, Figma’s central task was to prove that browser-native design could be fast enough, collaborative enough, and credible enough for professionals. The lasting impact of that period is architectural. Because the file was born collaborative, later teamwork, commenting, design-system management, and AI-context features fit naturally into the product rather than being bolted on.

The second stage was workflow expansion. The company moved from a designer’s tool toward a platform for the broader product team: community, whiteboarding, design systems, and developer handoff all widened the product’s role inside organizations. By the time Figma launched Dev Mode and later the MCP server, the company had moved past selling only to designers and was trying to own the bridge between design intent and production code. This is the stage in which Figma’s moat became less about the drawing canvas and more about the shared structure around the canvas.

The third stage was the merger interlude and balance-sheet reset. Adobe’s bid and its collapse changed the company’s financial and narrative profile. The abandoned merger produced a $1 billion break fee, which helped drive 2023 net income and operating cash flow in ways that were not repeatable. Then 2024 brought a tender offer and large equity-related charges, which pushed GAAP results sharply the other way. The lesson from this stage is that Figma’s pre-IPO GAAP history contains unusually large non-operating distortions. That is why investors who compare 2023, 2024, and 2025 headline earnings without context will be misled.

The fourth stage is the current one: public-company scrutiny in the AI era. Since the IPO, Figma has broadened the product line again, rolled out Figma Make more aggressively, pushed MCP integration into agentic coding tools, and introduced AI-credit monetization. Financially, the business has accelerated. In the market, the stock has gone the other way, because the market no longer pays extreme scarcity multiples, because share supply is still normalizing, and because AI now cuts both for and against the story. That is the stage the company is in as of the base date.

2.4 Key-node deep dive

The Adobe termination was a genuine fate-changing event, not a footnote. Beyond keeping Figma independent, it left the company with a stronger balance sheet, clearer proof of strategic importance, and a cleaner argument to public investors that independence preserved a larger opportunity set. The market still references that abandoned $20 billion price, but it should be treated as historical context, not valuation truth. The relevant point today is that Figma stayed independent long enough to present itself as an AI-era platform rather than as a swallowed product line inside Adobe.

The IPO itself changed less about the business than about the accounting and share structure. Upon the IPO, Figma’s multi-class common-stock structure took effect, with Class B holders receiving fifteen votes per share versus one vote for Class A. The IPO also triggered large stock-based-compensation recognition tied to awards whose performance condition was the listing itself. In other words, the IPO simultaneously expanded the investor base and made GAAP optics much uglier. That tension still defines the stock.

The extended lock-up agreement was underrated at the time and still matters. The regular lock-up ending in November 2025 merely shifted the overhang into staged releases that ran through August 2026 instead of clearing it. By the base date, three release tranches had already occurred and one large final tranche remained. For a stock whose investor base was still adjusting from a euphoric debut to ordinary software valuation math, that mattered.

The product launches in 2024 and 2025 also changed the company’s strategic direction more than the market initially appreciated. Figma Make, Figma Draw, Figma Sites, Figma Buzz, expanded Dev Mode, and MCP server support were an attempt to control more of the path from concept to shipped product in a world where AI generation could compress the labor involved in each individual step. This node still has direct consequences today because current valuation depends on whether those moves can hold the workflow together as AI accelerates.

4. Financial vertical review

2.5 Financial vertical review

The cleanest way to read Figma’s financial history is to separate operating momentum from event-driven accounting. Revenue has been straightforward: $504.9 million in 2023, $749.0 million in 2024, and $1.056 billion in 2025, with Q1 2026 revenue at $333.4 million. The direction is unmistakable. Growth came first from new paid-customer additions, later from deeper expansion among larger customers, and more recently from broader product adoption and AI-related monetization. In Q1 2026, management said growth was driven by seat expansion, retention, enterprise adoption, new users, and early traction from AI credit monetization, while the 2025 10-K tied the annual revenue increase mostly to growth in paid customers over $10,000 and over $100,000 in ARR.

Gross margins tell the more complicated story. Figma’s gross margin was 91% in 2023, 88% in 2024, and 82% in 2025. In Q1 2026 it fell to 79% on a GAAP basis, while management reported an 82% non-GAAP gross margin in prepared remarks. The driver was cost of revenue, which rose much faster than revenue, chiefly from hosting and inference costs tied to AI and higher usage by paid customers. Pricing pressure was not the cause. For the full year 2025, cost of revenue rose 112%, and management explicitly attributed much of that increase to technical infrastructure and hosting costs related to AI. That is the financial hinge in the model. Figma still has software-like gross margins, but AI is making the business less frictionless than old SaaS narratives assumed.

Earnings quality is poor on a GAAP basis and serviceable on a cash basis, but neither number can be taken at face value. Net income was boosted in 2023 by the Adobe break fee. It then swung to a $732.1 million loss in 2024 and a $1.25 billion loss in 2025, followed by a $142.4 million Q1 2026 loss. At the same time, operating cash flow was positive in 2025 at $250.7 million and stayed positive in Q1 2026 at $97.3 million. That split is evidence that Figma’s first-year public-company numbers are being pulled in opposite directions by merger history on one side and equity compensation on the other. GAAP still matters here. The right lesson is “normalize both.”

The single biggest financial fact for common shareholders is dilution. Stock-based compensation was $1.364 billion in 2025, including $50.98 million in cost of revenue, $697.68 million in R&D, $218.82 million in sales and marketing, and $396.66 million in G&A. In Q1 2026, stock-based compensation was another $169.0 million. Even after the IPO-triggered vesting pulse, the backlog remains heavy: at March 31, 2026, unrecognized SBC totaled roughly $1.96 billion across RSAs, ordinary RSUs, and the two 2025 CEO awards. This is why the gap between GAAP and non-GAAP is economically real. The current business is producing real cash, but it is still paying a very large part of compensation in equity.

The balance sheet is strong. At March 31, 2026, Figma held about $1.638 billion in cash, cash equivalents, and marketable securities, had no outstanding balance on its $500 million revolving credit facility, and disclosed no other long-term debt with floating rates. Deferred revenue was $627.7 million, which is healthy for a subscription business. Lease liabilities were modest at $56.2 million. This leaves the company financially durable even if the next phase of AI monetization proves bumpier than management expects. Figma is fighting for valuation credibility, not for liquidity.

Free cash flow is positive and capex-light. In Q1 2026, free cash flow was $88.6 million after only $7.8 million of capex and $0.9 million of capitalized internal-use software development costs. In 2025, the business produced $242.7 million of adjusted free cash flow according to management’s annual presentation. That capex profile is typical of software, but the owner-earnings picture is less rosy once recurring dilution is treated as a true economic cost. This is the central accounting reality behind the stock: Figma already looks excellent on a cash-versus-capex basis, but not yet on a cash-after-dilution basis.

5. Price and valuation history

2.6 Price and valuation history

Figma’s public-market history has already run through four distinct valuation phases in less than a year. The first was the scarcity explosion. The stock priced at $33 and closed day one at $115.50, then $122 the next day. At those prices, the market was underwriting a rare, fast-growing, AI-tagged software IPO in a reopened listings market, not ordinary software assumptions. Against 2025 revenue guidance of roughly $1.02 billion at the time, the first-day close implied a price-to-sales multiple in the mid-50s. That was the valuation of a story stock, not the valuation of a business with still-unsettled AI economics and very large equity compensation.

The second phase was the first reality check. Just days after the IPO, Reuters noted that the stock had already dropped as low as $92.75 after the euphoria began to deflate. The business had not changed. The multiple had. This is why the peak should not be treated as a “fair value” reference point in any serious analysis of the current drawdown.

The third phase was the post-earnings and lock-up reset. After the September 2025 quarter, Reuters reported that the company’s first public-company earnings beat consensus modestly, but the stock still fell because expectations had been loftier and because part of the employee lock-up was about to expire. In November 2025, the ordinary lock-up ended, while the extended lock-up remained in place. By then the market had moved on from “rare IPO” toward “expensive software stock now facing share supply and slowing headline growth.”

The fourth phase is the current one: AI probation. Figma’s own results improved into Q4 2025 and Q1 2026, and the stock rallied sharply after both February and May earnings because annual guidance beat expectations and AI monetization looked better than feared. Reuters described both reactions as relief that AI could be a growth driver rather than a threat. Yet the broader trend remained down because the market still viewed AI as both a tailwind to demand and a threat to the design category’s value capture. By 2026-07-27, the stock had closed at $22.94, and public historical data showed a 52-week high of $142.92 and a 52-week low near the high teens, with the public history page showing the all-time high closing price at $122 on 2025-08-01.

Current valuation is far below the peak, but it is not distressed. With about a $12.0 billion market cap and roughly $1.64 billion of cash against no debt, Figma trades around 8.4 times FY2026 revenue on a market-cap basis and about 7.3 times on an enterprise-value basis using the midpoint of current guidance. That is a large compression from the IPO frenzy, but it is still a premium to more mature peers because the market is crediting faster growth and a wider TAM expansion path. The valuation center has shifted for two reasons at once: the business is no longer viewed as a scarce IPO trophy, and the market is now demanding proof that AI expands the company’s role faster than it compresses seat economics.

6. Business model and moat

3.1 Revenue structure

Figma reports one operating segment, which is appropriate because the product suite is sold as a connected platform rather than as separately disclosed businesses. The revenue model nonetheless has distinct layers. The first and still dominant layer is seat revenue: Full seats for builders and designers, Dev seats for developer workflows, and lower-cost collaboration seats. The second layer is add-on revenue, which includes governance-related offerings and advisory services. The third and newest layer is consumption revenue from AI credits, with included monthly credits per seat and the option for paid top-ups and, by plan, pay-as-you-go billing. The commercial significance of AI credits is that they push Figma away from pure headcount-linked monetization and toward usage-linked monetization without abandoning the seat base that anchors enterprise contracts.

The real profit source is still the subscription business. AI credit revenue is early, but management’s Q1 remarks make clear that it is already affecting outperformance, especially in larger organizations and Pro teams buying add-ons. So far this looks complementary rather than cannibalistic. Management said Q1 outperformance in NDR was driven by seat expansion and growth in non-seat offerings, including AI add-ons, not by one offsetting the other. Customers using MCP were growing Full seats faster, and over 60% of larger customers added Full seats relative to their prior renewal. That matters because the bear thesis requires AI to make seat demand structurally worse. So far the evidence shows AI expanding seat value for enterprise users even as it may commoditize work at the long tail.

Customer concentration does not appear to be a material risk. Figma stated in filings that no single customer accounted for 10% or more of revenue. The larger business dependence rests on one workflow assumption: that serious digital-product teams will still want a controlled visual system of record even if much more of the code is generated automatically. If that assumption breaks, revenue concentration would shift from customer risk to category risk.

3.2 Cost structure and operating leverage

The cost base is split between classic software operating leverage and newer AI-related variable cost. R&D, G&A, and most of the go-to-market organization are scale costs. They should show leverage over time. Hosting and inference costs are more variable, especially for AI-heavy features. Figma’s filings also show that not all technical infrastructure is booked in cost of revenue. Some hosting tied to free users and AI-related free usage sits in sales and marketing. That accounting detail matters because it can make the margin impact of AI features look more dispersed than it really is.

There is real operating leverage in the model, but only on the adjusted basis. In Q1 2026, non-GAAP operating margin was 16%, while GAAP operating margin was negative 41%. That gap means the business can scale operationally at the same time as common-share economics remain pressured by dilution. Investors should therefore think about two separate leverage curves: the software business itself is already showing leverage, while the publicly reported earnings power available to common holders is still catching up because equity cost remains extraordinarily high.

3.3 Moat

Figma’s first real moat is workflow standardization around collaboration. The product was built for multiplayer design and review from the start. That matters more than brand. Teams choose Figma because designers, PMs, marketers, and developers can work in the same shared object rather than passing around exported files. The scale of that standardization is visible in the company’s claim that 95% of the Fortune 500 uses Figma. Even allowing for the breadth of the word “uses,” that kind of enterprise reach signals that the company already sits inside most large software organizations.

The second moat is structured design context. The MCP server, Dev Mode, and Code Connect logic push Figma beyond being a display layer. Figma now argues, with some evidence, that coding agents produce better output when they can read components, variables, layout data, and design-system rules from the source rather than visually infer them from screenshots. That is a stronger position than “we also have AI tools.” It means Figma can become upstream context instead of downstream decoration. That is the main reason I reject the idea that Figma is becoming a commoditized front-end overall. The front-end drawing task is easier to commoditize than the shared design-system graph and approval layer.

The third moat is land-and-expand economics across roles. Figma is no longer selling only to designers, and that is exactly what the AI era should reward if the company executes. Q1 data showed Paid Customers up 54% year over year to about 690,000, larger-customer cohorts accelerating, and Pro team conversions up over 150%. If AI were merely replacing the need for the platform, it would be harder to explain why broader role adoption is accelerating at the same time. The moat is uneven, though. It is strong in enterprise and team workflows, weaker in low-end experimentation where AI-native builders and lower-cost alternatives can substitute more easily.

Brand on its own is not a real moat, and neither is “AI.” Every serious software vendor now has an AI story. The durable advantages are collaborative architecture, embedded design systems, and role expansion inside product teams. If those weaken, the marketing story will not save the economics.

3.4 Management and governance

Dylan Field remains the defining management fact. He is chair, chief executive officer, and president. The board says that combination is justified because, as co-founder, he is best positioned to set strategy and execute the business plan. On product and growth, execution has been strong. The governance issue is control. As of the 2026 proxy, Class B shares carried fifteen votes each, and Field held 48.5% of voting power directly plus another 23.8% through an irrevocable proxy over Evan Wallace’s trust, for total voting power of 72.3%. That is founder control in substance, not merely in style. Long-term investors may accept it, but they should not pretend it carries no discount.

Capital allocation has been sensible in the narrow sense: the balance sheet is strong, there is no balance-sheet leverage problem, and the company has kept investing in product rather than trying to window-dress short-term margins. But compensation policy has been aggressive. The 2025 CEO stock-price award covered 14.5 million Class B shares across milestones from $60 to $130, and three of those tranches were already achieved in 2025. The separate 2025 CEO service award covered another 14.5 million Class B shares. That may align the founder with long-run stock performance, but it also reinforces why investors have to treat dilution as an essential part of the valuation story rather than an afterthought.

On governance hygiene, I did not find evidence in the public filings reviewed of a major accounting dispute, fraud issue, or auditor problem. Mike Krieger resigned from the board in April 2026, and the company stated that the resignation was not due to any disagreement on operations, policies, or practices. That does not erase all governance questions, but it does remove one easy source of speculation.

7. Industry and cycle

4.1 Industry structure

Figma sits at the intersection of collaborative design software, product-development workflow, and AI-assisted software creation. The industry is still in growth, but its profit pools are split. Historically, Adobe captured the lion’s share of the broader creative-software profit pool with deep specialist tools and huge distribution. Figma’s opening was narrower but powerful: collaborative interface and product design, then developer handoff, then adjacent ideation and prototyping. The current opportunity is larger than pure design seats because product managers, developers, marketers, and now AI agents can all consume the same design context. That is why Figma’s self-described platform has become wider over time.

Industry growth comes from several different sources at once. One is continued displacement of desktop and file-based workflows by cloud collaboration. Another is seat expansion into adjacent roles. A third is AI lowering the cost of trying things, which can increase workflow volume even if it reduces the manual effort per artifact. The profit pool does not automatically stay with the original design vendor, though. In an AI-heavy workflow, value can shift into foundation models, developer environments, code repositories, or workflow orchestration. Figma’s strategic task is to keep enough state, approval, system rules, and context inside its platform that AI increases volume on its turf instead of bypassing it.

The bargaining power picture is favorable downstream and mixed upstream. Enterprise customers can pressure vendors if design tools are seen as replaceable, but the pain of ripping out a standardized design system across large organizations is meaningful. Upstream, Figma is exposed to model providers and cloud infrastructure costs as AI usage scales. Management’s emphasis on routing queries across models and investing in first-party models is an attempt to regain bargaining power there. That is strategically sensible, but it is an early response, not a solved problem.

4.2 Cycle attributes

Figma is exposed to three softer cycles rather than to the deep cyclicality of commodity or inventory businesses. The first is enterprise software spending, which affects seat expansion and the speed of enterprise rollouts. The second is the rate cycle, because high-growth software valuation multiples remain sensitive to discount rates. The third is the technology-iteration cycle, and that one matters most now. When platform shifts are slow, collaboration software compounds steadily. When a platform shift is fast, the market tends to ask which layer will capture value and which layer will be flattened. Figma is in exactly that kind of technology-iteration cycle today.

In an upcycle for AI-enabled software creation, the biggest beneficiary should be whichever platform becomes the trusted context layer. For Figma, that means more Full and Dev seats, more AI-credit consumption, and better retention as teams standardize. In a downcycle for the thesis, the fragile point is slower seat growth and weaker willingness to pay as low-end generation tools make the visual layer feel cheaper, not immediate customer churn. That is why NDR and seat mix matter more than quarterly logo counts.

4.3 Policy, regulation, and geopolitics

The most important regulatory precedent is antitrust. Figma’s largest strategic transaction was blocked, and that ended any assumption that the company could rely on eventual takeout as a valuation floor. That matters for investors because it removes a convenient but lazy backstop from the thesis. There are also standard software-company risks around data security, AI training disputes, privacy law, and cross-border data handling, all of which Figma discloses in its filings. These are real risks, but today they are second-order compared with the bigger questions of monetization, margin structure, and competitive position.

8. Horizontal competitor analysis

5.1 Judge the competitive landscape first

This is Scenario C: ample competitors and substitutes, but only a few truly matter as valuation peers. Adobe, Atlassian, and GitLab are the best public reference set because investors actually use them to frame different parts of the debate. Adobe is the direct creative and design incumbent. Atlassian is the broader system-of-work and product-team collaboration benchmark. GitLab is the nearest public analogue on developer workflow, AI-assisted software creation, and the “system-of-record-for-builders” question. Private rivals such as Canva, Cursor, Lovable, Replit, and other AI-native builders matter competitively, but they do not solve the public-market comparison problem on their own.

5.2 Horizontal comparison dimensions

Adobe became a broad creative suite first and a collaborative cloud second. Customers choose Adobe for depth in image, vector, video, document, and creative-AI workflows. Figma customers choose Figma for speed of collaboration, component-driven product design, and shared context across design and engineering. Figma’s strongest territory is the product-team workflow around digital interfaces, while Adobe’s strongest territory remains specialist creative depth across many media types. The comparison is about territory rather than universal displacement. That is why Adobe’s AI anxiety and Figma’s AI anxiety rhyme without being identical. Adobe’s problem is whether creation tools get cheaper. Figma’s problem is whether the design layer remains the right control point.

Atlassian became the planning, ticketing, documentation, and coordination spine for software and business teams. Customers choose Jira, Confluence, Loom, and Rovo because they want work tracked, knowledge organized, and teams coordinated around execution. Figma sits earlier in the chain. It is stronger where product intent, interface systems, and visual specification are the bottlenecks. Atlassian is stronger where project orchestration and organizational knowledge are the bottlenecks. The competitive overlap runs through budget share and workflow gravity, not through direct feature conflict. If product teams begin in Figma and stay there longer, Figma gets room to expand. If the system of work becomes centered somewhere else and design becomes a generated commodity, Atlassian-like layers gain relative leverage.

GitLab became the code-and-delivery system of record. Customers choose GitLab for repository, CI/CD, security, and increasingly AI-assisted software development inside a single application. Figma is competing for the context that instructs what should be built before code lands in the repo, not for the repository itself. That distinction matters for terminal value. If AI makes raw code cheaper, the winning layers could be the repo, the issue system, and the design system. Figma needs to remain indispensable before GitLab begins to matter. Beating GitLab is not the test. The encouraging signal here is Figma’s MCP strategy, which explicitly aims to keep design context inside developer workflows rather than outside them.

Financially, the peer picture is revealing. Adobe is much larger and far more profitable. Atlassian is also much larger, growing strongly, and carries its own large gap between GAAP and non-GAAP, though nowhere near Figma’s current dilution intensity. GitLab is smaller than Atlassian and Adobe but already more mature on public-market multiple normalization. Against those three, Figma’s premium rests almost entirely on faster growth and on the market’s willingness to believe that its niche can widen from design into product creation. That premium narrows sharply if growth falls into the 20s or if gross margins continue to deteriorate. It widens only if AI expands revenue at least as fast as it raises cost and competitive uncertainty.

5.3 Ecological-niche analysis

Figma’s niche is better described as the leader in collaborative product-design workflow and a challenger platform for AI-native software creation context, rather than as “Adobe challenger” in the simple sense. It takes profit most directly from fragmented point tools, from manual handoff work, and from specialized design seats that are hard to collaborate around. The entities most likely to take its profit pool run past Adobe and Canva to AI-native builders that can make low-end prototypes and websites feel good enough without a rigorous design layer.

If the industry hits a price war in low-end creation, Figma’s position gets weaker in the long tail and stronger in the enterprise core, because governance, shared libraries, approval workflow, and design-system integrity matter more as software creation speeds up. That is why I think the company’s long-run role is real but narrower than the most enthusiastic bulls imply. Figma is unlikely to own all AI-generated creation. It has a good chance to own the enterprise design-system checkpoint through which a meaningful share of that creation must pass.

9. Current fundamentals and bull-bear divergence

6.1 Last 4 quarters

Figma’s last four reported quarters show a business that has become stronger even as the stock became weaker. Q2 2025 revenue rose 41% to $249.6 million, with positive operating income and 129% NDR. Q3 2025 revenue rose 38% to $274.2 million and management raised full-year guidance, though GAAP results were swamped by $975.7 million of one-time IPO-related SBC. Q4 2025 revenue rose 40% to $303.8 million, NDR reached 136%, and 2026 revenue guidance came in above expectations. Q1 2026 then accelerated again to 46% growth at $333.4 million, with another guidance raise. On the simple operating facts, Figma has executed better, not worse, since listing.

The tension is that margin structure has not kept up with the top line. AI-related hosting and free-user support costs have risen fast, and Q1 2026 GAAP gross margin fell to 79% from 91% a year earlier. Management’s own Q1 remarks made clear that broader and deeper AI adoption improved engagement and retention, but they also spent unusual time explaining model routing, provider optimization, and first-party models to manage inference cost. Companies do not make that kind of operating point unless cost discipline has become a real part of the story.

6.2 What is the market trading right now?

The stock is trading a conflict between real fundamentals and a punitive narrative. The real fundamentals are accelerating revenue, rising NDR, strong enterprise cohort growth, and a healthy net-cash balance sheet. The narrative is that AI might hollow out the design layer, compress seats, and turn Figma’s new AI features into a costly defense rather than a new monetization engine. Lock-up supply and dilution intensify that skepticism because investors have not yet seen a clean, common-share version of the earnings power. The result is a stock priced as if the business is strong today but suspect tomorrow.

6.3 Bull-bear divergence

The bull case rests on evidence already visible in usage and monetization. NDR at 139%, 48% growth in customers over $100,000, weekly Make usage above 60% in large customers, and the faster Full-seat growth among MCP users all suggest that AI is broadening Figma’s relevance rather than bypassing it. The most persuasive bull point is that enterprise customers are using AI products in ways that appear to increase, not reduce, the need for shared design context.

The bear case rests on three harder facts. First, gross margin has already taken a hit from AI-related infrastructure. Second, GAAP profitability for common shareholders remains poor because dilution is still enormous. Third, the final 77.7 million-share extended-lock-up release had not yet occurred by the base date. None of those facts disproves the product thesis, but all three explain why the stock can stay cheap even if the business stays strong.

My judgment is that the market is overstating immediate AI obsolescence and understating the degree to which Figma may become the design-system layer for AI-generated software. At the same time, many bulls are understating dilution and cost drag. So the market is directionally wrong on the product role and directionally right on demanding a lower multiple than the IPO frenzy implied. That combination leads to a nuanced but firm conclusion: the collapse from the peak was mostly justified, but the current narrative is too bearish on strategic relevance and not bearish enough on common-share economics.

10. Valuation analysis

7.1 Historical valuation

Public valuation history is short and distorted, so percentile language has to be used carefully. Within its own public life, Figma’s current valuation is much closer to the low end than the high end. But that says more about how extreme the IPO was than about whether the stock is cheap now. A move from roughly mid-50s sales multiples at the first-day close to about 8.4 times market-cap-to-guided-sales today is a collapse in expectations, not evidence by itself of undervaluation.

7.2 Peer valuation

Relative to peers, Figma still trades at a premium on forward sales. Adobe’s updated FY2026 target implies a much lower sales multiple despite strong cash generation. Atlassian also trades at a lower multiple despite faster recent revenue growth than many investors expected. GitLab, though slower-growing than Figma, is also in the same general neighborhood rather than miles away. The premium for Figma is therefore understandable, but it is no longer a “pay anything” premium. It requires the company to keep proving two things: that growth remains well above the peer set and that AI expands lifetime customer value instead of merely raising variable cost.

7.3 Absolute valuation

7.3.0 Cash-flow passthrough

Operating cash flow versus net income is not analytically clean over 2023-2025 because 2023 included the Adobe break fee and 2024-2026 included extraordinary equity-accounting effects. Still, two things are clear. First, capex is low: Q1 2026 capex plus capitalized internal-use software was under $9 million, and 2025 capex intensity was also low. Second, reported free cash flow materially overstates owner earnings because stock-based compensation remains massive and recurring. With roughly $1.96 billion of unrecognized compensation expense still on the books at March 31, 2026, a normalized common-share earnings view is far below headline free cash flow. I therefore default to enterprise value to forward revenue, with an explicit dilution assumption, rather than to EPS or unadjusted FCF.

The valuation below is scenario analysis within a research framework, not investment advice.

Using diluted share assumptions in the low-530 million range and the net-cash assumptions above, these scenarios imply per-share values of roughly $17 to $18 in the conservative case, around $24 in the base case, and around $32 to $33 in the optimistic case. The current price is therefore not demanding if one believes the optimistic case, but it offers little margin of safety against the conservative one.

7.4 Expectation-gap analysis

The market is implicitly pricing a mixed outcome: Figma remains relevant, but growth decelerates enough and dilution remains high enough that the stock does not deserve a premium anywhere near its IPO-era multiple. The biggest expectation gap at the next print is whether AI monetization carries gross margin and seat mix the right way, not headline revenue. Investors will care most about Q2 growth versus the $348 million to $350 million guide, gross margin, commentary on AI-credit attach rates, NDR, and what happens when the last 77.7 million extended-lock-up shares become eligible for sale.

7.5 Margin-of-safety recheck

Current price sits above the value implied by the conservative scenario, so margin of safety is zero on that yardstick. The most fragile assumption in the base case is the belief that Figma remains the design-system-of-record in enterprise AI workflows and therefore holds NDR comfortably above 130, rather than the FY2027 revenue number by itself. If that assumption is cut to about 70% of the base conviction level, the likely result is lower growth and a lower multiple, which pulls base-case fair value down into the high teens.

If earnings power were flat for the next three years, current return prospects would likely be low-single-digit and inferior to a reasonable bond hurdle. On that basis, this is a good company but not yet a clear bargain. It is worth waiting for a better price rather than treating the post-IPO collapse as sufficient proof of value.

Margin-of-safety sufficiency verdict: none.

11. Risk analysis

8.1 Business risk

The first major risk is AI-driven seat compression. Probability: medium. Impact: high. The observable indicators are lower Full-seat growth, weakening Dev-seat upgrade commentary, a drop in Pro team conversion momentum, and NDR rolling over toward the low 120s or below. The transmission path is straightforward. If AI-native tools make rough prototyping good enough for more teams, Figma can keep usage high while monetization strength fades, especially in the long tail. Revenue growth then slows, and the market stops paying a premium multiple for strategic relevance.

The second business risk is that AI becomes revenue-accretive but economically thinner than investors hope. Probability: medium-high. Impact: high. The indicators are GAAP gross margin staying below about 78%–79%, cost of revenue rising faster than revenue, and management spending more time on cost controls than on attach rates. The transmission path starts with lower gross profit per unit of usage, moves into weaker operating leverage, and ends in a multiple ceiling even if top-line growth remains decent. This is already the place where the bear case has the strongest evidence.

8.2 Financial risk

The biggest financial risk is dilution, not debt. Probability: high. Impact: medium-high. Figma has no long-term debt problem and more than $1.6 billion of cash, but it also has nearly $2.0 billion of unrecognized stock-based compensation expense to run through over the coming years. If the share price recovers materially, the economic transfer to employees will be more visible, not less. The transmission path runs through share count, per-share economics, and the market’s willingness to believe non-GAAP margins.

8.3 Valuation risk

The largest valuation risk is that even the current multiple still prespends some success on AI. Probability: medium. Impact: high. Figma at around 7.3 times EV to FY2026 revenue is no longer a bubble stock, but it is still priced above slower-growing peers because investors see a bigger runway. If Q2 and Q3 show merely “good” execution rather than clear evidence of durable AI monetization and controlled margins, the multiple can still compress. To fall, the stock only needs the story to shift from “strategic winner” to “solid tool with unresolved economics.” No disaster required.

8.4 Governance and external risk

Founder control is a real governance risk. Probability: medium. Impact: medium. With 72.3% voting power effectively under Dylan Field, ordinary shareholders cannot realistically change direction if they dislike compensation policy, acquisition strategy, or AI investment intensity. That control can be positive in long-cycle product transitions, but it also means governance protection is weak if the strategic bet is wrong. The observable indicator is whether compensation, equity issuance, and major product bets become harder to justify against common-share outcomes, not voting drama.

A final external risk is post-lock-up selling pressure around the last extended-lock-up release. Probability: high in timing, medium in lasting damage. Impact: medium. The remaining 77.7 million eligible shares after the Q2 2026 release are large enough to affect technical price action and sentiment even if they do not alter intrinsic value. In a stock already associated with AI skepticism, supply can become narrative.

12. Catalysts and tracking indicators

9.1 Positive catalysts

The clearest positive catalyst is another quarter in which AI monetization shows both breadth and discipline. That means Q2 revenue at or above guidance, NDR holding in the mid-130s, larger-customer growth staying strong, and gross margin not deteriorating further despite the first full quarter of credit monetization. A second positive catalyst would be the final extended-lock-up release passing without extraordinary selling pressure. A third would be evidence that MCP and Make continue pulling developers and PMs into paid seat upgrades rather than merely increasing feature usage among existing designers.

9.2 Negative catalysts

The obvious negative catalyst is a guidance reset tied to slowing seat growth or weaker AI-credit attach. More subtle but equally important would be a quarter where revenue still beats but gross margin breaks lower again, because that would confirm the fear that AI revenue is arriving with poor incremental economics. Another negative catalyst would be heavy insider or venture-holder selling once the final extended-lock-up tranche is released.

9.3 Tracking dashboard

The dashboard matters because Figma’s story can look healthy on one metric while eroding on another. Revenue without retention is not enough, and neither is retention without gross-margin control, or positive free cash flow without dilution discipline. Investors should pay closest attention to the combination of NDR, gross margin, and seat-mix commentary. Those three together will tell you whether AI is raising the platform’s value or merely raising usage costs. The Q2 2026 earnings date is already set for August 5, 2026.

13. Cross-synthesis summary

10.1 Bull and bear reasons

Bull reasons:

• Figma’s core operating metrics are strengthening, not weakening, with Q1 2026 revenue up 46%, NDR at 139%, and large-customer cohorts accelerating.
• MCP and Make usage suggest Figma is becoming more central to AI-native software workflows, with MCP-using large customers growing Full seats materially faster.
• The balance sheet is strong, with about $1.64 billion in cash and marketable securities and no long-term debt outstanding.
• The stock has already undergone a major multiple reset from an unsustainable IPO frenzy, reducing the risk that investors are still paying bubble prices.

Bear reasons:

• GAAP economics are still weak for common shareholders because SBC remains very large and recurring, with nearly $2.0 billion of unrecognized expense still to be recognized as of March 31, 2026.
• AI already pressured gross margin, with Q1 2026 GAAP gross margin falling to 79% from 91% a year earlier as hosting and inference costs jumped.
• A large final lock-up overhang remained in place at the base date, with about 77.7 million shares still due for release after Q2 2026 or by August 31, 2026.
• Founder voting control is absolute enough to warrant a governance discount, with Dylan Field effectively controlling 72.3% of voting power through direct holdings and the Wallace proxy.

10.2 Pre-mortem: where I might be wrong

Script one: by mid-2027, AI-native builders and design-generation tools make low-end prototyping cheap enough that Figma’s long-tail paid conversions slow materially. Revenue growth drops toward 20%, NDR falls below 125%, and the market stops paying even 6 times forward sales. If EV/Sales compresses toward 3.5 to 4.0 times while dilution continues and the company still lacks clean GAAP profitability, the stock could trade in the low teens and be down roughly 50% from the current price.

Script two: Figma keeps growing revenue, but AI monetization proves gross-margin destructive. Cost of revenue keeps rising faster than planned, GAAP gross margin falls into the mid-70s, and investors conclude that Figma is winning usage but losing economics. The multiple then compresses even without a revenue collapse, because the market reclassifies the company from software-quality compounding to AI-assisted growth with weak passthrough. That combination could also easily halve the share price if it happens during a broader software multiple contraction.

10.3 Final research conclusion

Figma has already proved something important. It proved that collaborative, browser-native product design was a new center of gravity for digital-product teams, not a feature. It also proved, in the first year after the IPO, that AI has not yet broken that gravity. The numbers through Q1 2026 argue the opposite: customers are deepening their use of the platform, larger accounts are expanding quickly, and AI-related products are helping rather than hurting adoption.

What Figma has not yet proved is the part that matters for public-market returns: that this AI-era relevance will convert into clean, durable, common-share economics. The market’s harsh re-rating from the post-IPO peak was mostly justified because that peak was built on scarcity and euphoria, not on sober cash-flow math. Today’s price is much more defensible, but it still assumes enough success that I do not see a clear margin of safety. My view is therefore straightforward: Figma is more durable than the market’s most bearish AI-disruption narrative implies, but the stock is not yet compelling enough to buy aggressively before the last supply overhang clears and before another quarter tests whether AI credits support margins rather than simply support growth.

【Company-profile scores】

• Fundamental quality: high
• Growth: high
• Moat: medium
• Financial soundness: strong
• Management credibility: high
• Valuation attractiveness: low
• Risk level: high
• Suitable investor type: long-term growth

【Investment rating】

• Rating: Watch

• One-line thesis: Strong enterprise and AI-era usage trends are real, but dilution, gross-margin pressure, and the last lock-up release leave too little margin of safety at today’s price.

• Three price signals:

  • 【Ideal Buy Price】12–14 USD Basis: at least a 20% margin of safety below the value implied by the conservative scenario, with allowance for dilution and post-unlock volatility.
  • Acceptable hold price: 21–27 USD
  • Clearly overvalued price: 36 USD and above

• Current-price classification: acceptable hold

• Whether to wait for a better price: yes. A more attractive entry would be in the mid-teens or after Q2/Q3 confirm that AI-credit revenue is scaling without another leg down in gross margin. The opportunity cost of waiting is that the stock could rerate quickly if Q2 beats cleanly and the final unlock passes quietly.

• Target holding horizon: 3–5 years

• Expected annualized return:

  • Conservative scenario: about -24% to -20%
  • Base scenario: about +3% to +6%
  • Optimistic scenario: about +38% to +44%

• Max-loss risk: roughly 50% in a script where growth slows toward 20%, NDR drops below 125%, and the multiple compresses to 3.5–4.0 times EV/Sales.

• Reassessment-trigger signals:

  • if net dollar retention falls below 125% for two consecutive quarters
  • if GAAP gross margin falls below 76% for two consecutive quarters
  • if customers over $100,000 in ARR growth drops below 25% year over year
  • if stock-based compensation shows no meaningful decline as a percentage of revenue through 2027
  • if post-Q2 2026 unlock selling causes persistent abnormal volume and management cannot demonstrate stable institutional absorption

【Valuation Range】

• current: 22.94 (close as of 2026-07-27)
• bear (conservative · ideal buy zone): [12, 14]
• base (fair · acceptable hold zone): [21, 27]
• bull (optimistic · above the clearly-overvalued line): [36, 40]

14. Key data tables

Figma’s top line has been excellent, but the economic story is bifurcated. Reported cash generation is strong; common-share economics are still heavily shaped by dilution. The 2023 operating cash flow figure is distorted by the Adobe break-fee-related tax effects and should not be treated as a run-rate.

† 2023 and 2024 operating cash-flow comparability is distorted by the abandoned Adobe merger and related tax/payment items.

This is the core peer picture. Figma is the fastest-growing of the group but not the cheapest. The current premium is therefore a growth and strategic-role premium, not a quality-premium based on mature earnings. That is why the stock can look optically “down a lot” and still not be obviously undervalued.

15. Research uncertainties

The biggest blind spot is the true seat-versus-consumption mix of revenue after Q1 2026. Management has provided encouraging directional commentary on AI credits, but not enough disclosure yet to model the steady-state economics with confidence.

A second uncertainty is how much of the current margin hit is transitional and how much is structural. Management described several levers to control inference costs, but investors have only one full quarter of partial monetization and no full-quarter post-launch history at the base date.

A third uncertainty is the magnitude of the final extended-lock-up selling once the last 77.7 million shares become eligible. The legal supply schedule is disclosed clearly; the actual behavioral outcome is not.

A fourth uncertainty is peer-set evolution. The most important long-run threats to Figma may come from private AI-native companies rather than from its current public comparables, which makes relative valuation inherently incomplete.

16. Sources

• Figma investor relations quarterly results and earnings releases for Q2 2025, Q3 2025, Q4 2025, and Q1 2026.
• Figma Form 10-K for FY2025 and Form 10-Q for Q1 2026.
• Figma Q1 2026 prepared remarks and product/help documentation on pricing, AI credits, and MCP server workflows.
• Figma proxy statement and April 2026 8-K for governance, voting control, and board changes.
• Figma IPO pricing announcement, prospectus-related SEC materials, and lock-up disclosures.
• Reuters reporting on the IPO, post-IPO trading, quarterly reactions, AI-related market narrative, and Adobe deal context.
• Peer-company primary materials: Adobe Q2 FY2026 release, Atlassian Q3 FY2026 release, GitLab Q1 FY2027 release.
• Current market data and public price history pages used for as-of-date pricing and historical stock-path verification.

17. Other tickers mentioned

• ADBE.US: primary creative-software incumbent and the abandoned acquirer, central to the design-software and AI-valuation comparison
• TEAM.US: system-of-work peer used to frame product-team workflow valuation and AI-era collaboration economics
• GTLB.US: developer-workflow peer used to test whether Figma can remain the design-system layer in AI-native software creation
• MSFT.US: referenced indirectly through VS Code and GitHub-style developer workflows that interact with Figma’s MCP strategy

This report is based on public information and does not constitute investment advice. Markets carry risk; invest with caution.