Figma Positions Design as the Operating Layer for AI-Built Software
Figma chief executive Dylan Field says the company is investing in AI agents and code generation on the premise that cheaper, easier-to-generate code will make design the layer where teams define and steer software. Figma is already monetizing AI but is not including unproven products in its forecast, even as it reports accelerating revenue growth and a negative operating margin. Its aim is to connect design and code in the Figma canvas, allowing teams to build from design rather than treating it as an upstream handoff to engineering.
Figma is investing ahead of AI products it does not yet count on
Figma is reporting accelerating revenue while investing through a negative operating margin. Dylan Field said the company’s guidance is based on what it can observe and does not assume upside from efforts that have not yet proved themselves.
Figma reported $370.1 million in second-quarter 2026 revenue, up 48% year over year, against a Bloomberg estimate of $351.5 million. Its third-quarter revenue guidance of $373 million to $375 million also exceeded the $365.1 million estimate shown on screen. The same earnings graphic reported a negative 32% operating margin.
Field described the quarter as Figma’s third consecutive period of accelerating revenue growth and its first full quarter of AI monetization. He put net dollar retention, or NDR, at roughly 136%, saying the overall result left the company feeling “very strong” while it continued to invest for what is ahead.
The distinction in his guidance is central to Figma’s AI posture. The company is already monetizing AI, Field said, but prospective gains from products still being developed are not part of the observable base it uses to forecast. The investment period, in his telling, is meant to fund products and workflows that customers are only beginning to adopt.
Customers are moving between design and code rather than following one AI path
Asked how customers are incorporating AI into their work, Dylan Field said they are at different stages of figuring out where it belongs in their software-development workflows.
Some customers, whom Field characterized as early adopters, had identified their approach by the end of the prior year and were already doubling down on Figma. Others have spent the past six months working through the same question and are now also deepening their use of the product. Figma’s task, he said, is to meet customers where they are as those workflows change.
That means supporting work that begins in different places. Some teams start with design; others start in code. Field said Figma’s MCP lets customers pull work into Figma or read from Figma in order to build. He expects more of those movements between design and code to become encompassed directly in the Figma workflow through its agent.
The Figma footage shown alongside his comments illustrated the intended connection. One canvas displayed a designed site with navigation, event information, copy, images, and layout components. A Figma Make demonstration showed an “explore” page being brought to life after a request to add sorting and filtering; the generated interface included a visible .CSS indicator. The product presentation treated design and generated functionality as work occurring in the same environment rather than as a handoff between separate tools.
Field’s proposition is not that generated code eliminates the need for a place where teams define what they are making. As code becomes easier to produce, he said, Figma wants the design environment to remain connected to the work of building it.
Figma wants design to steer automated software production
Dylan Field said his immediate focus is on moving Figma’s agent toward monetization while improving how it serves customer needs and workflows. He also highlighted “code layers”: bringing code into the canvas.
As code becomes more of a commodity, design is the operating level and layer that you're able to work in in order to define what your team is doing.
Field’s argument is that easier code generation makes the design layer the place where a team specifies the product and directs what should be built. Figma’s agent would support that execution, while code layers bring the building process into the canvas. If the company captures that workflow, he said, teams will be able to build out from the design layer, with designers retaining a steering role.
The claim is also a statement about where Figma expects control to sit as AI changes software creation. Rather than treating design as an upstream artifact that engineers translate into a finished product, Field described it as the operating layer from which a team can define its work and build outward. The agent is intended to make that layer more capable, not to replace it.
When Ed Ludlow characterized the company as still trying to build, Field agreed. His emphasis was not on extending an established design tool set but on establishing an AI-assisted workflow in which designers can continue to steer the resulting software.
