Figma

Priced at $33, closed day one at $115.50, bottomed at $16.84, and the business grew faster than ever the whole way down

Figma's July 31, 2025 debut was one of the largest first-day pops ever for a billion-dollar IPO: priced at $33.00, it closed at $115.50 — up 250% — and touched $122.00 the next day, a market value around $56 billion. The S-1's warning that a limited float would amplify volatility turned out to be the single most predictive sentence in the document, in both directions. From that peak the stock fell almost without interruption: $70.28 by the end of August 2025, $51.87 by the end of September, $35.97 by late November, and a trough of $16.84 on June 25, 2026 — down 49% from the IPO price and roughly 86% from the peak. As of August 28, 2026 it trades at $28.82, still 12.7% below the offer price, with the market cap around $15 billion.

The collapse was a multiple story, not a revenue story. Quarterly revenue marched from $249.6 million in Q2 2025 (+41%) to $274.2 million (+38%), $303.8 million (+40%), $333.4 million (+46%), and $370.1 million in Q2 2026 (+48%) — by mid-2026 Figma was reporting its third straight quarter of accelerating growth and raised full-year guidance to roughly $1.46 billion (+39%). What the income statement did show was the S-1's disclosed pain arriving on schedule: the IPO RSU trigger produced a $1.1 billion net loss in Q3 2025, and operating losses continued every quarter afterward ($135 million in Q1 2026, $117 million in Q2 2026) as stock compensation and AI infrastructure spending — the exact margin risk the filing flagged — ate the P&L. Supply mechanics compounded the fade: the stock spent the entire measurement window far above the $41.25 early-release threshold, unlocking employee shares on the accelerated schedule the prospectus described.

The AI double-edge defined the narrative war. Bears treated Figma as roadkill for prompt-to-app tools of the Cursor/v0/Lovable generation, and the market largely priced it that way — a sub-$10 billion valuation at the June 2026 trough for a company compounding above 40%. Figma's counterattack used the other edge: Figma Make built on Anthropic's Claude models, a $200+ million acquisition of Israeli AI startup Weavy in October 2025 (rebranded Figma Weave), Model Context Protocol integrations with Claude Code and OpenAI's Codex announced in early 2026, and AI-credit monetization that analysts credited for the growth reacceleration. The dependence risk sharpened too: in April 2026 Anthropic's Mike Krieger left Figma's board amid reports that Anthropic's own models would ship design tooling competitive with Figma's core product — the S-1's warning about relying on model providers 'in a highly competitive environment' made flesh.

Thirteen months in, the scoreboard is split. Dylan Field still controls the company and still runs it; retention and growth never cracked; the balance sheet never needed rescuing. But an investor who bought the first-day close has lost three-quarters of their money, and one who bought at the IPO price is down 12.7% while the S&P-grade question — does AI make Figma bigger or unnecessary? — remains genuinely open. The S-1 disclosed nearly everything that happened; what it could not disclose was which of its two AI stories the market would believe in any given quarter.

Figma is the browser-based, multiplayer design tool that became the default way software teams design products — and, after regulators killed Adobe's $20 billion acquisition and Figma pocketed a $1 billion breakup fee, it arrived at the IPO window as that rarest of 2025 offerings: a profitable, 46%-growth SaaS company with 132% net dollar retention, pitching itself as the emerging system of record for all of product development. The S-1 asks investors to believe that AI will expand, not destroy, the market for designing software, and to accept a founder-controlled share structure in exchange for the ride.

The best case runs through platform expansion and AI. Figma Design was only the wedge: FigJam, Dev Mode, Slides, and the four 2025 launches (Sites, Make, Buzz, Draw) turn a design tool into an end-to-end idea-to-product platform, and 76% of customers already use two or more products. Growth is bottom-up and organic — word-of-mouth adoption inside organizations that converts free users into paid seats and then whole-company deployments, with net dollar retention of 132-134% doing the compounding. Dev Mode showed the playbook works: launch a product for an adjacent persona (developers are ~30% of monthly active users) and retention expands. If AI genuinely multiplies the number of people who design and build software — Figma Make turning prompts into working prototypes — the company sells more seats to more kinds of collaborators, and the $1 billion of Adobe breakup-fee cash plus GAAP profitability funds the whole campaign without dilution.

With hindsight: The operating half of the best case has largely tracked: multi-product expansion and AI monetization drove revenue from $249.6M (Q2 2025) to $370.1M (Q2 2026), growth reaccelerated to 48% — as fast as at IPO — and guidance calls for ~39% growth to ~$1.46B in 2026. But the 'AI expands the market profitably' half is unproven: every post-IPO quarter showed operating losses as AI and stock-comp costs bit, and the market's verdict was brutal — from a $56B day-one valuation to under $10B at the June 2026 trough, with the stock still 12.7% below the $33 offer price 13 months later. The platform thesis is winning; the valuation thesis failed at the offered pop, and is roughly break-even against the actual IPO price.

• Figma has repeatedly expanded beyond its original design tool into a multi-product platform, and each new product broadens the base it can monetize.

  “we have evolved our business and platform significantly since publicly launching our initial product, Figma Design, in 2015, including through the introduction of new offerings” source ↗

• Growth is organic and bottom-up — users adopt Figma by word of mouth inside organizations, keeping acquisition costs structurally low.

  “Historically, a significant portion of our revenue growth has been derived from organic growth that occurs within organizations when new users decide to use our platform based on word-of-mouth recommendations, as opposed to management driven enterprise-wide procurement processes” source ↗

• The free Starter plan is a conversion funnel: free users become paying customers and paying customers buy more products.

  “our future success depends, in part, on our ability to convert users of our free plan into paying customers on a paid pricing plan and selling additional offerings to existing paying customers” source ↗

• New products demonstrably expand retention — Dev Mode's launch drove net dollar retention up through 2024.

  “we experienced an expansion in our Net Dollar Retention Rate throughout 2024 subsequent to our launch of Dev Mode in 2023” source ↗

• Figma is investing heavily to make AI an accelerant of its platform rather than a threat to it.

  “While we have made, and expect to continue to make, significant investments to integrate AI, including generative AI, into our platform” source ↗

1. AI could obsolete design tools or shrink the value of interface design

   partly came true

   competition · structural

   The S-1's most striking disclosure: AI may lower the cost of building competing products, change how people interact with software in ways that devalue interface design itself, or reduce the number of designers and developers who need Figma at all. The company selling an AI-accelerated future explicitly warns that the same force could make it obsolete.

   “altering how consumers and businesses interact with websites and apps and consume content in ways that may result in a reduction in the overall value of interface design, or by otherwise making aspects of our platform obsolete or decreasing the number of designers, developers, and other collaborators that utilize our platform” source ↗

   What happened: The fear materialized in the multiple, not (yet) the fundamentals: the AI-roadkill narrative helped drive the stock from a ~$56B day-one value to under $10B by June 2026 even as revenue growth reaccelerated to 48%. Demand for design work did not visibly collapse, and Figma monetized AI credits — but the obsolescence question remains open, and its own model suppliers began edging toward design tooling.

2. Short-term AI investments will compress gross and operating margins

   came true

   profitability · serious

   Figma states flatly that AI features like Figma Make will hurt both gross and operating margins in the near term, and concedes it cannot yet estimate the longer-term impact. Inference costs sit inside a per-seat subscription model that was not designed to pass them through.

   “in the short term, we expect that our AI investments and use of AI technologies, including as part of Figma Make and our Figma AI features, will negatively impact our gross margins and operating margins” source ↗

   What happened: Exactly as warned: Figma posted operating losses in every reported quarter after the IPO — including $135M in Q1 2026 and $117M in Q2 2026, well after the one-time RSU charge — and analysts repeatedly flagged elevated AI infrastructure spending as the cost of the growth reacceleration.

3. Intense competition from AI-driven and design-to-code entrants

   partly came true

   competition · structural

   Beyond legacy suites, the filing calls out a new class of competitor: AI-native tools that compress or skip the design step entirely and go straight from prompt to product. These entrants can be built cheaply precisely because AI lowered the barrier.

   “We face competition from a number of companies, including companies that cater to multiple stages of the design and development process, point tools that address individual parts of the process but can expand to cover more, and design-to-code and AI-driven companies and tools that compress or accelerate steps in the workflow” source ↗

   What happened: The AI-native competitive wave (prompt-to-app tools in the Cursor/v0/Lovable mold) arrived and dominated the bear case, but Figma's reported results show no share collapse — growth accelerated three straight quarters to 48% by Q2 2026. Figma also counterattacked, acquiring AI startup Weavy for over $200M in October 2025 and shipping its own AI products.

4. AI-driven efficiency could reduce seat counts under per-seat pricing

   unclear

   growth · structural

   Figma charges per user. If AI lets fewer people do the same design and development work, customers need fewer seats, and the revenue model erodes even while the product stays loved. The company admits it may have to rework its billing model in response.

   “there could be a decrease in the number of designers, developers, and other collaborators that use our platform if such individuals are able to significantly increase their efficiency through the use of AI capabilities alongside or instead of our platform” source ↗

   What happened: Too early to judge. Figma hedged the model by layering AI-credit monetization on top of seats, and revenue growth of 46-48% in the first half of 2026 suggests no seat erosion yet — but public disclosures don't break out seat counts, and the structural question of what AI does to per-collaborator pricing remains unresolved.

5. Dual-class structure concentrates control in Dylan Field, including a proxy over his co-founder's shares

   came true

   governance · structural

   Class B carries 15 votes per share, and co-founder Evan Wallace handed Field an irrevocable proxy over his entire stake, giving the CEO control over every stockholder matter — directors, charter amendments, and any sale of the company. Public investors are passengers.

   “Mr. Field will be able to control matters submitted to our stockholders for approval, including the election of directors, amendments of our organizational documents, and any merger, consolidation, sale of all or substantially all of our assets, or other major corporate transactions” source ↗

   What happened: The structure operates exactly as disclosed: Field controls the company through super-voting Class B shares plus the Wallace Proxy, and public holders had no governance recourse during a drawdown that took the stock down ~86% from its peak. No control challenge is possible by design.

6. A wave of post-IPO stock-based compensation will weigh on profitability

   came true

   financing · serious

   The offering triggers the performance condition on years of accumulated RSUs — roughly $809.8 million of stock-based compensation measured as of March 31, 2025, with $735.9 million more to follow — guaranteeing that reported results turn sharply negative right after listing.

   “Following this offering, our future cost of revenue and operating expenses, particularly during the quarter in which this offering is completed, will include a substantial amount of stock-based compensation expense with respect to these RSUs” source ↗

   What happened: On schedule: the IPO-triggered RSU charge produced a roughly $1.14B operating loss and $1.10B net loss in Q3 2025, swinging Figma from GAAP profitability to a deeply loss-making fiscal 2025, and heavy stock-comp expense kept every subsequent quarter in the red.

7. Lock-up terms allow early release of employee shares if the stock trades 25% above the IPO price

   came true

   market · serious

   The lock-up has a valve: if the stock sustains a price at least 25% above the IPO price around the first post-offering earnings release, a tranche of employee shares frees up early. The mechanism converts a hot aftermarket into fresh supply on an accelerated schedule.

   “if the closing price per share of our Class A common stock on NYSE for at least five trading days in any ten consecutive trading day period, with at least one of such five trading days occurring after the Initial Post-Offering Earnings Release Date is at least 25% greater than the initial public offering price” source ↗

   What happened: The stock spent the entire post-IPO measurement window far above the $41.25 early-release threshold (monthly closes of $115.50, $70.28, and $51.87 from July through September 2025), triggering the accelerated employee release described in the S-1. Fresh supply coincided with the relentless fade — from $70.28 at end-August to $35.97 by end-November 2025.

8. Limited float and IPO dynamics could produce a volatile stock price

   came true

   market · serious

   Only a small slice of the company floats at the IPO, and the filing warns that the thin float will amplify price swings in either direction, untethered from operating results.

   “In addition, the limited public float of our Class A common stock following this offering will tend to increase the volatility of the trading price of our Class A common stock” source ↗

   What happened: Spectacularly: +250% on day one (one of the largest pops ever for a $1B+ IPO), a peak of $122.00 the next day, an 86% peak-to-trough collapse to $16.84 by June 2026, and a 52-week range that still spans more than 4x. The thin-float warning was the most predictive sentence in the filing.

9. Revenue growth is expected to decelerate as the business matures

   didn't happen

   growth · serious

   Standard but consequential at a premium multiple: Figma tells investors not to extrapolate its 46-48% growth, and separately expects net dollar retention to fluctuate or decline off a growing revenue base.

   “Even if our revenue continues to increase, we expect our revenue growth rate to decline in the future as our business matures and our platform achieves more widespread adoption” source ↗

   What happened: The opposite, so far: after dipping to 38% in Q3 2025, growth reaccelerated for three straight quarters to 48% in Q2 2026 — matching the pre-IPO pace — with full-year 2026 guidance of ~39% growth to roughly $1.46B, aided by AI-credit monetization.

10. The March 2025 pricing and billing overhaul makes forecasting harder and could alienate customers

    unclear

    other · serious

    Months before the IPO, Figma ended user-driven seat upgrades, introduced admin approval and multi-product seats, and raised the top price. The company concedes the change muddies its own forecasting and makes historical comparisons less meaningful — at exactly the moment public investors must underwrite the trend.

    “The changes to our pricing and packaging plans introduced in March 2025 have made, and any further changes to components of our billing models that we may introduce in the future may make, forecasting our operating results more difficult” source ↗

    What happened: No visible customer revolt surfaced in reported numbers — growth held above 38% throughout — but churn and seat-level effects of the repricing aren't separately disclosed, so whether the overhaul cost goodwill or merely reshuffled revenue timing can't be judged from public data.

11. Dependence on third-party foundation models for AI features

    partly came true

    platform dependence · serious

    Figma's AI products, including Figma Make, run on rented off-the-shelf foundation models. It cannot control their availability or pricing, and the vendors it depends on are themselves racing up the stack in a highly competitive market.

    “We cannot control the availability or pricing of such third-party AI technologies, especially in a highly competitive environment, and we may be unable to negotiate favorable economic terms with the applicable providers” source ↗

    What happened: The dependence deepened rather than resolved: Figma Make launched on Anthropic's Claude models, and early 2026 brought MCP integrations with Claude Code and OpenAI's Codex. The competitive edge of the risk sharpened in April 2026 when Anthropic's Mike Krieger left Figma's board amid reports of Anthropic shipping design tools competitive with Figma — the supplier-turned-rival scenario the S-1 gestured at.

12. Heavy reliance on Dylan Field as CEO and product visionary

    didn't happen

    key person · serious

    Field is chair, CEO, president, controlling stockholder, and per the filing critical to technology, product, and strategy — while also actively investing in outside technology companies, which the S-1 admits could divert his attention and create conflicts.

    “we are highly dependent on the services of Dylan Field, our Chair of our Board of Directors, Chief Executive Officer, and President, who is critical to the development of our technology, products, platform, future vision, and strategic direction” source ↗

    What happened: Field remained chair and CEO throughout the first 13 months, personally fronting the AI pivot (Figma Make, the Weavy acquisition, model-provider partnerships). No succession event or damaging distraction materialized.

13. Single-cloud dependence on AWS

    didn't happen

    platform dependence · serious

    The platform was designed to run on AWS, and Amazon holds discretion over pricing, terms, and access; migration would be costly and slow. Hosting costs also rise mechanically with the free user base.

    “AWS may also take actions beyond our control that could seriously harm our business, including, but not limited to, discontinuing or limiting our access to one or more services, increasing pricing terms, terminating or seeking to terminate our contractual relationship altogether” source ↗

    What happened: No AWS disruption material to Figma was reported in the first 13 months as a public company; the dependence persists but stayed dormant.

14. Pending OFAC voluntary self-disclosure over potential sanctions violations

    unclear

    legal · serious

    Figma may have inadvertently let users in embargoed or sanctioned countries access the platform and self-reported to OFAC; the matter was unresolved at filing, with no loss accrued.

    “As a result, we have submitted to OFAC a voluntary self-disclosure concerning potential violations and the voluntary self-disclosure is still under review” source ↗

    What happened: No public resolution, fine, or enforcement action surfaced in the sources reviewed through August 2026; the matter appears to remain a dormant tail risk.

15. Security breaches and prior cybersecurity incidents

    didn't happen

    tech & security · boilerplate

    Figma hosts customers' unreleased product designs — commercially sensitive by definition — and discloses that it has already experienced cybersecurity incidents, though none material to date.

    “We have experienced, and may in the future experience, cybersecurity incidents; however, to date, these incidents have not had a material impact on our business, operating results, and financial condition” source ↗

    What happened: No material breach or security incident at Figma was publicly reported between the IPO and August 2026.

16. Treasury exposure to a Bitcoin ETF and alternative assets

    unclear

    market · serious

    Unusually for an enterprise software issuer, Figma parked part of its corporate treasury in a Bitcoin exchange-traded fund and reserves the right to buy more alternative assets, importing crypto volatility into quarterly results.

    “changes in the market value of our investments, including in our marketable securities, in particular as a result of volatility related to our investments in a Bitcoin exchange-traded fund or any future investments in alternative asset classes” source ↗

    What happened: The position drew IPO-season commentary but no disclosed material gain or loss; it remained small relative to Figma's post-breakup-fee cash pile, and no fetched source reports it affecting results through mid-2026.

How the S-1 reads

A founder-king IPO born from blocked M&A: like other second-act listings after failed acquisitions, Figma arrived cash-rich ($1B breakup fee) and control-locked (15:1 dual class plus a co-founder proxy), asking public investors for capital it didn't need on terms it wouldn't negotiate. The distinctive pattern is the self-refuting AI disclosure — the same technology is pitched as the growth engine in the front of the prospectus and named as a potential obsolescence event in the risk factors — paired with a per-seat pricing model whose core unit (the human collaborator) is exactly what AI threatens to economize. Add the disclosed supply mechanics (an $810M RSU trigger, an early-release lock-up valve tied to a 25% pop) and the S-1 reads, in hindsight, like a complete instruction manual for the boom-bust trading pattern that followed.

• “AI has become more prevalent in the markets in which we operate and may result in significant changes in the demand for our platform, including, but not limited to, reducing the difficulty and cost for competitors to build and launch competitive products” source ↗

  The S-1 names the exact mechanism of its own bear case: AI lowering the cost of building competitors and changing demand for the platform.

• “Mr. Wallace and his affiliated trust entered into the Wallace Proxy, granting Mr. Field complete and unlimited authority to act, in his sole discretion, on their behalf, to vote any number of shares of our capital stock” source ↗

  Co-founder Evan Wallace handed Dylan Field an irrevocable, sole-discretion proxy over his entire stake.

• “we have a number of current employees whose equity ownership in our company has given them a substantial amount of personal wealth, or will shortly after the completion of this offering result in them having substantial wealth” source ↗

  Figma disclosed the retention risk of minting hundreds of newly wealthy employees on IPO day.

• “in 2022, we entered into an agreement to be acquired by Adobe, Inc. (“Adobe”). However, based on our joint assessment that there was no clear path to obtain the required regulatory approvals for the transaction to close, and in 2023, we mutually agreed with Adobe to terminate the agreement” source ↗

  The filing's dry account of the blocked $20 billion Adobe deal that made this IPO possible.

• “we have observed a lengthening of the sales cycle recently for some prospective customers that we attribute to increased sensitivity to information technology security concerns, particularly with respect to products that include AI features or otherwise incorporate AI technologies” source ↗

  AI is already lengthening Figma's own sales cycles as buyers scrutinize AI-bearing products.

• “we have experienced instances of credential sharing, abuse of our Figma for Education offerings, credit card fraud, and other instances of misuse or fraud on our platform that result in bad debt, chargebacks, or other losses to us” source ↗

  Bottom-up adoption has a shadow side the company itemizes: freeloading and fraud at the platform's edges.

What this one teaches

• A first-day pop is a pricing error, not a verdict: the same company was worth $56B on day one and under $10B eleven months later while revenue grew every single quarter — the fundamentals barely moved, the multiple did all the work.
• When an S-1 pitches a technology as its growth engine in the summary and its obsolescence risk in the risk factors, the stock will trade on whichever story the market currently believes; the risk lives in the narrative and the multiple, not the income statement.
• Disclosed supply mechanics are a schedule, not a footnote: the $810M RSU trigger and the 25%-above-IPO early-release valve told you in advance that euphoria would manufacture both a loss-making first year and accelerating share supply into the pop.
• IPO-year GAAP is nearly unreadable after an RSU performance trigger — Figma swung from $44.9M quarterly net income to a $1.1B quarterly loss without the business changing; judge the operating loss ex-charge and the growth line instead.
• A blocked acquisition can be the best outcome for the target: the $1B Adobe breakup fee plus forced independence let Figma fight the AI transition with its own products and balance sheet — though shareholders who paid the pop, not the company, funded the market's education.

The paper trail

1. 2025-07-01 S-1 filing index ↗ document ↗
2. 2025-07-21 S-1/A filing index ↗ document ↗
3. 2025-07-28 S-1/A filing index ↗ document ↗
4. 2025-07-31 424B4 filing index ↗ document ↗

Filed as Figma, Inc.. All documents are public domain, served by SEC EDGAR.