Figma Is Building AI Features for a User That Doesn't Exist Yet
Figma opened at $115 on its first day of trading. Nine months later it's at $21. Below its own IPO price.
In between, they shipped more AI features than any design tool in history. Make. Code to Canvas. MCP server. AI image tools. Slides AI. The list keeps growing.
And the stock keeps falling.
The question the market is asking
The market isn't punishing Figma for lack of innovation. It's asking a harder question: do these features create value or just cost?
40% revenue growth. A billion dollars in annual revenue. And still not profitable. Every AI feature requires compute. Every compute cycle costs money. If the features don't drive retention or expansion revenue, they're just burning cash to show the market that you're doing something. Sora had the same problem: spectacular technology, no reason to come back on day ten.
What I've seen from the inside
I use Figma every day. It's the best design tool ever built. The multiplayer model, the component system, the dev handoff workflow: all of it is genuinely world-class.
I've tested most of their AI features. Some save real time. Auto-layout suggestions, smart rename, content generation for mockups: these solve real friction points in my daily work. I use them.
Most of the others I stopped using after a week. Not because they don't work. Because they don't solve a problem I actually have frequently enough to build a habit around.
The user that doesn't exist yet
That's the part nobody wants to say out loud. Figma is building AI features for a user that doesn't exist yet.
The current Figma user base is professional designers who know exactly what they want to create and need precise control over every detail. AI features that generate layouts or suggest designs are solving a problem these users don't have. They already know what to build. They need better tools for building it, not tools that guess on their behalf.
The user these features are actually for, the non-designer who wants to create professional-quality designs with AI assistance, doesn't use Figma today. And there's no guarantee they'll choose Figma over the dozen other AI-first design tools being built specifically for them.
Figma is building features for a user they hope will arrive while potentially alienating the users who are already there.
The lesson I keep coming back to
You can't fix a weak problem space with better features. The SaaSpocalypse is proving this across the entire sector. If the behaviour you're designing for isn't real yet, the feature doesn't matter how good it is.
The best product companies don't ship what's possible. They ship what's needed. Right now, Figma is shipping possible.
A $60 billion valuation said the market believed in the vision. An $11 billion valuation says show us the users.
The gap between those two numbers is the cost of building for a user that doesn't exist yet. Every product team should study it carefully, because the temptation to build AI features for hypothetical users is stronger than it's ever been. And the market has started punishing it before the design community has even noticed. $2.5 trillion is being spent on AI and almost none of it is going into solving the actual usability problems.
---
