Figma went public on July 31, 2025 at $33 a share, popped 250% on day one to close at $115.50, and got valued near $68 billion. It peaked at $142.92. This week it trades around $19, with a market cap near $10 billion.
That is down roughly 87% from the high. Down about 50% just in 2026. Down 43% from the IPO price itself.
What makes Figma different from every other busted software IPO is that the business got better the whole way down. Revenue growth didn’t slow as the stock fell. It accelerated. The question making the rounds is whether Figma is the single most oversold name in software.
The answer is more interesting than yes or no, and you only see it by putting Figma next to its peers: first against a profitable slow-grower like HubSpot, then against the whole cohort of software leaders still growing 30% or more.
The numbers the market is ignoring at Figma
Q1 2026, reported May 14:
- Revenue of $333.4M per quarter ($1.3B+ ARR), up 46% year over year
- That 46% accelerated from 40% the prior quarter and 38% the quarter before. Two straight quarters of acceleration, not deceleration
- Net dollar retention of 139% on paid customers over $10K ARR, the highest in over two years
- 82% gross margins
- Free cash flow of $89M, a 27% FCF margin
- $1.6B of cash on the balance sheet
- Accounts paying $100K+ a year up 48% to 1,525
- Paid customers up 54% to 690,000
- Full-year guidance raised to $1.422B to $1.428B, about 35% growth at the midpoint
A B2B company growing 46% with accelerating growth, 82% gross margins, 139% NDR, and a 27% free cash flow margin is a top-decile business. Back out the cash and it trades around 6x forward revenue. In 2021 that same business carried a 30x+ multiple.
The stock sits at $19 for three reasons, and only one of them is real.
First, the IPO pop was never real. A 250% day-one move is not price discovery, it is a small float meeting frantic demand. The $142 peak was a sentiment number. A big chunk of the decline is just air leaving a balloon that should never have inflated. Measuring “oversold” from a fake peak overstates the case.
Second, the entire software sector got repriced on AI fear. ServiceNow, SAP, Microsoft, Adobe, all down hard. When the index gets cut, the highest-multiple name gets cut most. Figma was the highest-multiple name.
Third, and this is the only one that matters: the market decided AI kills Figma. When Anthropic shipped Claude Design in April, the take spread fast that prompt-to-interface tools make a design platform redundant. Findell Capital, now running an activist campaign, has pushed the board to examine the Anthropic relationship.
That third thesis is exactly what Q1 contradicts. If AI were eating Figma, NDR would not be 139%. Seats would not be expanding across entire organizations. Weekly active users of Figma’s MCP, the feature that lets AI coding agents read and write Figma files, would not have grown 5x in a quarter. When code becomes a commodity, the design file becomes the system of record that the agents have to read from. The market is pricing Figma as a wrapper. The data says it is the source of truth.
First comparison: Figma vs HubSpot, the slow-and-profitable anchor
HubSpot is the opposite kind of company. Twenty years old. Profitable. The category-defining CRM and marketing platform for the mid-market. No one thinks AI makes HubSpot redundant overnight.
HubSpot’s Q1 2026 did $881M in revenue, up 23% as reported and 18% in constant currency, with ARR of $3.45B, nearly 300,000 customers, a 21% full-year operating margin target, and roughly $750M of free cash flow expected for the year. Full-year revenue guidance is $3.70B to $3.708B, up 18%.
HubSpot trades around $182, down from a 52-week high of $568. That is a 68% drawdown, and its market cap fell more than 50% over the trailing year to roughly $9.4 billion.
The odd part: Figma and HubSpot carry almost the same market cap right now, around $9 to $10 billion each. But HubSpot does 2.6x the revenue: $3.7B versus Figma’s $1.425B. The market values a $3.7B profitable platform and a $1.4B hypergrower at nearly the same absolute number, and pays Figma a 3x revenue-multiple premium for its growth rate. That premium is defensible. Faster growth and a stickier wedge are worth more.
But flip it. HubSpot, a profitable category leader still growing 18% to 23% with $750M of free cash flow, trades at 2x forward revenue and about 14x forward non-GAAP earnings. That is a value-investor multiple on a software compounder. By any standard of the last decade, both of these stocks are cheap. They just got cheap in different ways.
Second comparison: the 30%+ growth club, where Figma looks broken
The HubSpot setup tells you the whole sector got repriced. The sharper test is to line Figma up against the other software leaders still growing 30% or more, the names everyone agrees are winning the AI cycle, and look at what the market pays for each unit of growth.
Every one of these reported in the last two months. Multiples are EV to forward revenue, as of late June 2026:
Read that table top to bottom. Palantir grows 85% and the market pays 36x. Cloudflare grows 34% and gets 28x. Datadog grows 32% and gets 19x. Snowflake grows 34% and gets 15x, the cheapest of the infrastructure names.
Then Figma: it grows 46%, faster than every name on the list except Palantir, and trades at 6x, the lowest multiple in the group except HubSpot, which grows less than half as fast.
This is not a sector-wide discount anymore. HubSpot at 2x makes sense as the slowest grower in the group. Figma at 6x does not, because Figma is near the top of the growth table and near the bottom of the multiple table at the same time. Snowflake grows 12 points slower than Figma and trades at more than double the multiple. Datadog grows 14 points slower and trades at 3x the multiple.
Add the quality screen and it gets worse for the bears. Figma’s Rule of 40, growth plus free cash flow margin, lands around 73. That is elite, second in this group only to Palantir’s absurd 145, and ahead of Cloudflare, Snowflake, and CrowdStrike. The company with the second-best Rule of 40 in the cohort has the second-cheapest multiple in the cohort. Those two facts do not belong in the same row.
The bull and bear cases
The Street is split, and both sides have a real argument. Consensus sits at a Hold, with price targets clustered from the high-$30s to the mid-$40s. On a stock near $19, that range implies 90%+ upside if the bulls are right and real pain if the bears are.
The bull case:
- Growth and quality together: 46% accelerating growth, 139% NDR, 82% gross margins, a 27% free cash flow margin, and FY2026 guidance implying about 35% growth. Citi initiated at Buy with a $36 target on what it called ramping AI traction.
- AI shows up as a tailwind in the numbers, not a threat: MCP weekly actives up 5x, Figma Make and Weave driving seat expansion, AI credit monetization live since March.
- The losses are mostly optics. The 2025 GAAP loss was driven by IPO-related stock-based comp, a non-cash item, and consensus points to GAAP profitability in 2026.
- Insider conviction: a Figma board director bought roughly $36.5M of stock near the lows in February 2026.
- Valuation reset: about 6x forward revenue versus a far richer IPO multiple, and the cheapest multiple in its growth cohort.
The bear case, the serious version rather than the permabear one:
- It is still not “cheap” on GAAP. RBC Capital carries the Street-low target near $28 with a Hold, citing AI-driven margin pressure. On a trailing operating margin still deep in the red, paying up on revenue is a premium, not a bargain.
- The AI threat is structural, not just pricing. Anthropic, Google, and the prompt-to-app tools are going after the premise that a designer opens the canvas at all, the deeper risk laid out below. Morgan Stanley sits at Equal Weight with a $38 target.
- Dilution is real. Stock-based comp ran near $1B in 2025, and the final tranche of insider shares, roughly 35% of the total, unlocks in August 2026. That is a near-term supply overhang on top of steady share-count creep.
- An activist is circling. Findell Capital is publicly pressing the board to cut costs, sharpen product focus, and re-examine the Anthropic relationship, which tells you not everyone trusts current capital allocation.
- The long-term economics are unproven. Even sympathetic value investors note the wide spread in intrinsic-value estimates, which means there is no clean margin of safety at today’s price.
The bears are right about the setup and the bulls are right about the business. The Anthropic threat and the August unlock are real risks. But nothing in today’s numbers, a 46% grower holding 139% retention, explains a multiple one-third of Snowflake’s. The bear case arguably caps the upside. It has not yet justified the discount.
The Anthropic problem: design is the natural next step after code
The bear argument that matters most is not valuation, and it is not really Claude Design as a product either. It is gravity. Coding is the number one use case for large language models, and design sits one step upstream of code. A model that can write your app is a short hop from generating the interface that becomes the app. Every frontier lab gets pulled into Figma’s lane by the logic of its own best product, not by a side bet. Anthropic is simply first and most direct about it.
The dependency makes it sharper. Figma Make, Figma’s flagship AI feature, runs on Anthropic’s Claude models, so Figma routes AI spend to the same lab that, on April 17, shipped Claude Design straight at its core surface. Anthropic’s chief product officer, Mike Krieger, resigned from Figma’s board three days before the launch. The stock fell 6 to 8% that day. Figma is paying for the engine that powers its own competitor.
This is why the threat is different from every AI scare Figma has shrugged off. Adobe’s Firefly and Figma’s own AI tools assumed a trained designer in the loop, and Figma’s estimated 80 to 90% share of UI and UX design is built on that assumption. The labs’ tools do not assume a designer. Claude Design, Google’s Stitch, Lovable, and Vercel’s v0 all start from a prompt, not a canvas. The goal is not a better canvas. It is to make opening Figma unnecessary for landing pages, decks, and prototypes, the work non-designers used to file as design requests.
The resource gap is the other half. Anthropic went from roughly $9B in annualized revenue at the end of 2025 to north of $30B by early April 2026, and ships updates about every two weeks. Figma cannot out-model the labs whose models it resells, and several of them now want its market. This is what Findell Capital is pointing at when it presses the board to re-examine the Anthropic relationship: Figma is financing and validating its own disruptor.
The bear case has a real ceiling, though. Claude Design still cannot do what Figma’s pros depend on: no editable .fig export, no shared component libraries, no real-time collaboration, no developer handoff. Nobody moves a 500-component design system into a chat window in 2026. And none of it shows up in the numbers yet: NDR is 139%, seats are expanding, MCP usage grew 5x. The whole debate is timeline. The next 18 to 24 months decide whether Figma stays the system of record or becomes the upstream that AI tools route around. That is the risk the multiple is pricing, before the income statement has felt a dollar of it.
Figma is a Multiple Detached From Growth, But Fixated On a Tough Possible Agentic Future-to-Come
Figma has the most dramatic chart, down 87%, which is why people reach for “most oversold.” The comparison work says that frame is wrong.
If you want the safest oversold name, it is HubSpot: a profitable leader at 2x revenue and 14x earnings, with bounded downside. If you want the cleanest mispricing, it is Figma, because the multiple is completely detached from where the company sits on growth and quality versus its actual peers. A 46% grower with a Rule of 40 in the 70s should not trade at one-third of Snowflake’s multiple. The peer set is the proof the discount is specific, not just macro.
The reason both can be true at once is the thing founders should take from this. The 2026 repricing of B2B + AI was so violent it stopped distinguishing between companies. It compressed an 18% profitable compounder and a 46% accelerating design leader into the same $10 billion bucket, and it parked Figma three turns of revenue below software growing slower than it. The market is no longer pricing fundamentals. It is pricing one macro fear, that AI commoditizes all software, and applying it hardest to the names with “design” or “interface” in the description.
Figma’s Q1 said the opposite: AI is the tailwind, not the threat. Accelerating growth, 5x MCP adoption, AI credit monetization already live, seats expanding across whole organizations. The numbers are getting better in the exact place the bears say the company should be dying.
Back to the question. Is Figma the most oversold software stock of all? On the chart, no, plenty of names fell further from sillier peaks. On growth-and-quality-adjusted value versus its real peers, Figma is the strongest case on the board. The fastest grower in its cohort is priced like one of the slowest. That gap, not the 87% drawdown, is the signal.
