San Francisco startup Paper closed a $34 million Series A on July 23 to build a design platform where HTML and CSS are the native language, not an export destination. The timing could not be sharper for Figma investors.
The round was led by Accel and ICONIQ, with participation from Designer Fund, WorkOS co-founder Michael Grinich, Lovable co-founder Anton Osika, and engineers from Anthropic and OpenAI. Paper was founded by CEO Stephen Haney, and its ARR has grown 25 times since launch. For four consecutive months, Ramp's Fastest Growing Companies list has included it. Customers include Ramp, Vercel, PostHog, and Y Combinator.
The product is straightforward in principle, difficult to execute. Paper renders every design element in real HTML and CSS, so what a designer touches on the canvas is literally the same code that ships to production. There's no handoff document, no Zeplin annotation layer, no developer spending a Tuesday afternoon re-implementing what the designer built in a proprietary format. The gap between design and engineering that has frustrated product teams for a decade is the thing Paper is trying to close structurally, not by adding better export presets.
Where it gets genuinely interesting is the AI integration. Paper has a bidirectional MCP server with 24 tools, letting agents read and write on the design canvas: creating artboards, updating styles, syncing design tokens, pulling JSX or Tailwind output from any node. In practice this means an AI coding agent can make a change in the codebase and have it reflected on the design canvas without anyone moving a file. That is the workflow that enterprise teams shipping software with agents actually need. Figma's architecture was not built for it.
Figma's stock fell 52% in the first half of 2026, according to The Motley Fool, a stunning number for a company that grew revenue 46% year over year to $333.4 million and posted net dollar retention of 139%, its best in two years. The fundamentals are strong. The market simply doesn't believe they'll stay that way. The catalyst investors keep citing is Anthropic launching Claude Design in April 2026, a tool that generates interactive prototypes and marketing materials from natural language prompts. The announcement alone sent Figma stock down roughly 6-7% in a single session.
Frankly, the fear isn't really about Claude Design specifically. It's about what Claude Design signals: that the most sophisticated AI labs in the world view design software as adjacent territory worth competing for. Figma's moat has always been its collaborative canvas and the lock-in that comes with design systems built inside it. That moat holds when the workflow runs through a human designer. It gets murkier when the agent is also a participant.
It's not trying to be a better Figma. It's built on the premise that Figma's data model is the problem: a proprietary vector format that must be translated before it touches production code. Paper skips the translation entirely. The backers who wrote checks this week, including engineers who have worked at the labs now entering the design space, are betting that this architecture matters more as agents take on more of the work.
What $34M buys when you're fighting upstream
Figma is not standing still. The company launched Figma Make in 2026, its own AI-to-code feature, attempting to bridge design and production from its side. It still processes tens of millions of design files. Its enterprise relationships are deep and its brand with designers is genuinely strong. A 3-year-old startup with $34 million does not beat that on distribution.
What Paper can do is win the next cohort. The companies building new products today, teams at Vercel, PostHog, and the Lovable-style vibe-coding shops where an AI agent is writing most of the frontend, have no particular loyalty to the Figma workflow because they didn't grow up inside it. For them, a canvas that speaks HTML natively and has an MCP server wired up for agents isn't a novelty. It's the obvious default.
The $34 million will go toward product development and team expansion, according to the company's announcement. The competitive window is what matters more. If Paper can entrench itself with the companies now defining how agentic software teams work, and its customer list suggests it already is, the question for Figma isn't whether Paper wins the whole market. It's whether the next generation of product teams ever forms the habit in the first place.
Figma declined to comment on the Paper funding. Fifty-two percent of its market cap this year said plenty already.
Also read: Nubank is buying a $6 million bank so it can keep calling itself a bank • How to Build a B2B SaaS Sales Funnel Before You Hire a Sales Team • Appeals court blocks Trump's $100,000 H-1B visa fee and a Supreme Court fight is now likely
