AI is killing Figma.

Figma's stock has dropped 85% in nine months. In the same period, its revenue grew 40%. That gap is the whole story: the market isn't betting against the company, it's betting against the way of working that Figma was built for. For years, the expensive part of design was turning a picture of a product into real, working code. That part now takes an afternoon and an AI tool. The designers who could read code already left the old workflow years ago. The falling stock price is just the slowest people in the room finally noticing.

This essay is about why the market is right, what it means for the "show off your mockups" culture that grew up around Figma, and what's actually left for the tool once the workflow it owned disappears. It's written for designers, founders, and the people who lead design teams — because all three are about to make decisions based on whether they believe this.

The chart

Figma closed around $21. Its high was $143.

That's an 85% drop in nine months — from a company that just reported quarterly revenue up 40% year over year, full-year revenue over $1 billion, and guidance for 30% growth next year. As a business, Figma is doing fine. The stock market is voting that it won't stay that way.

What makes the chart interesting is that each big drop lines up with a specific competitor announcement, not with anything Figma did wrong:

• Google ships a Stitch update — its prompt-to-UI tool. Figma drops 12% in two days.
• Anthropic announces a new frontier model. The stock falls 14% over three days.
• Anthropic launches Claude Design — prompts to polished prototypes. The stock drops another 7%.

This isn't a normal post-IPO cool-down. It's a direct reaction to tools that all share one trait: none of them asks you to start with a blank Figma canvas. Stitch makes a UI from a prompt. Claude Design makes a prototype. Lovable, Bolt, and v0 make working apps from a description.

Every announcement that shrinks the gap between "idea" and "shipped product" wipes another chunk off Figma's value. The market is pricing in the end of an entire way of working.

What Figma was actually for

Figma's product is excellent. That's not the argument. The argument is about what designers used it for, and whether that job still exists.

The job was simple. A designer made a screen in Figma — the pixels, the spacing, the components, the states. An engineer then turned that screen into code. Figma was the canvas where the design lived. The code was where the real product lived. And Figma's shareable link was the handoff between those two people.

That middle step — translating the design into code — was expensive. It needed an engineer, an interpretation, a handoff, and a few rounds of back-and-forth review. And here's the key point: the only reason a high-polish Figma file was worth making was because that translation step was expensive. You needed a detailed source to hand off. Figma was the best tool ever built for making that source.

Now the translation step is no longer expensive. v0 takes a screenshot and ships working code. Cursor takes a Figma frame and writes the components. Claude Code takes a written description and deploys a real app. When translation costs almost nothing, the detailed source file becomes optional. The work moves straight into code.

Figma's product hasn't gotten worse. What disappeared is the job that made Figma essential.

The years of fake work

I owe part of my career to Dribbble and Behance. I posted there for years. I won small things. I got hired off them.

So I can say this honestly: about ninety percent of what I posted was theatre. Polished mockups of products that didn't exist and would never ship. A "Spotify redesign" by someone who'd never seen Spotify's actual numbers. An "Air India rebrand" by someone who'd never set foot in their offices. A "case study" of a fintech app that was really just thirty Figma screens with nothing behind them.

We built a whole performance economy on top of Figma. Figma was the stage. Dribbble was the audience. LinkedIn was the applause. Every screen was made for an imaginary CEO — never a real user.

And the hiring world took it seriously. Companies hired off pretty mockups. Design courses taught students to build portfolios that were mockups with paragraphs attached. A whole generation learned to optimize for looking like they'd solved a problem, instead of actually solving one.

When AI made polished mockups cheap, it exposed how hollow this layer was. A prompt now produces a "Spotify redesign" in fifteen seconds. The skill that used to take fifteen hours in Figma isn't scarce anymore — and the only people who ever valued it were designers showing off to other designers.

So the market isn't only repricing Figma. It's repricing the entire culture of showing off in Figma.

What I did instead

I made that choice early: I stopped using Figma for anything that would actually ship. I started designing in code — first with Cursor, then Claude Code. Every case study on this site since then was built in the real, running product, with version control and a live link, with real users on it within the same week.

The reason was just math. Figma added a step, and that step cost calendar days, not hours. Three rounds of handoff feedback take three weeks. Three rounds of "build it, ship it, watch someone use it, fix it" take three days.

The Figma step lost the math.

This wasn't some clever insight. By then it was the obvious move for any designer who could read code. The ones who couldn't are the ones who stayed — and they're the ones now reading threads asking "is design dying?" while the stock chart answers them every single day.

It isn't design that's dying. It's the version of design that needed Figma to function. Designers who can think — who can hold a user's mindset, pick the right system, and decide what to build — are more valuable than ever. What's gone is the translation step in the middle. And for the last decade, translation was most of a product designer's week.

The honest counterargument

The strongest case for Figma: AI tools make a mess of scattered pieces. v0 makes a component, Lovable makes an app, Stitch makes a screen — and someone still has to decide which one is right, what the design system is, who owns the spec. Figma, the argument goes, is where all those pieces get reconciled into one true version.

That's a real job. It's also the bull case for the stock at $21. Two responses.

One: "source of truth for design" is a smaller job than "the design tool for everyone." It only matters at companies big enough to need design governance — Google, Atlassian, Shopify. That doesn't support the valuation Figma had at its peak. The market shrinks.

Two: in the long run, the source of truth is the running product itself — not a Figma file describing what the product should look like. Design tokens in code, component libraries, AI keeping the design system in sync — these are eating the "source of truth" job too. Reconciliation happens in the code, not in a separate design tool beside it.

Figma at $21 is priced for a future where it's the design-governance tool for big enterprises. That future is real. It's just smaller than the future Figma was priced for at $143.

If you live in Figma

Two specific moves.

Build something in code this week. Open Cursor or Claude Code, pick the smallest design problem you have — one card, one screen, a marketing page — and build it directly in code. Don't start in Figma. The first time will feel awkward. By the third time it'll feel obvious.

Stop building mockup portfolios. Build things that run. Now the portfolio that gets you hired is a list of working links, not a grid of pretty images. One shipped login flow, browser extension, or working tool beats fifty mockups.

Try both. If your sense of design genuinely hasn't shifted afterward, fine — go back to Figma. Most people who try won't.

If you lead a team

The same shift changes what you hire for and how you run the team. Stop screening for mockup portfolios; ask candidates for things that run. Stop measuring designers by how clean their files are; measure them by what shipped and whether users actually used it. And take the math seriously — if your team still routes every decision through a Figma handoff, you're paying in calendar weeks for a step that competitors are now doing in days. The teams that win the next two years are the ones designing in the real product, not in a picture of it.

The close

The chart isn't saying AI is replacing designers. It's saying the workflow Figma was built for is being replaced. The designer is fine. The tool the work used to flow through is in trouble.

Most designers haven't noticed, because Figma is the air they breathe. It'll become obvious to them the same way it became obvious to investors — slowly, in a series of 14% drops, until one day they realize the workflow they trained for has been redundant for two years and they were the last to know.

Figma will survive. It'll be a smaller, governance-focused tool for companies that can afford a design-ops team. The designers and leaders who do well are the ones who already see it coming.

The ones who don't yet — open Cursor.