Figma used its Config conference on June 24 to move code, motion, and generative tooling onto the design canvas: code layers that turn any frame into editable source, a Motion timeline with keyframes and AI-generated transitions, promptable shader fills built on WebGPU, generative plugins, and more than twenty Figma Weave tools wired into the canvas. [1][2][3][4]
The demo version of the day was bright. Designers animate components once and watch the motion travel through design systems like typography does. Developers inspect timelines in Dev Mode and copy animation as CSS, JSON, or React. Teams hand work across fewer walls because the walls themselves moved into the file. Chief executive Dylan Field framed the theme as ceiling-raising: AI has lowered the floor of creation, he argued, but people still raise the ceiling. [2]
The business version of the day lives elsewhere — in the help pages. They name the beta states, the plan eligibility, the seat treatment, the government-plan exclusions, and the boundaries of AI credits that govern who actually gets each capability and at what tier. [4] Code layers ship first to a waitlist starting in July. [2] That gap between stage announcement and entitlement table is where Figma's real product strategy is legible. The company is not merely adding features. It is deciding which parts of code, motion, agents, and collaboration live inside one paid control plane, and pricing the doorway accordingly.
MSM product coverage can celebrate the creative-tool expansion on its merits; The Verge's account of the announcements treats the reimagined canvas as the story, and it is a real one. [1] X will run the older argument — that AI replaces designers or empowers them — under clips of shader demos. Both frames skip the structural fact. A canvas that hosts code, motion, agents, and repositories becomes infrastructure, and infrastructure earns money through seats, credits, and enterprise contracts long after the keynote ends.
The details reward attention precisely because they are dull. Motion exports to MP4, WebM, animated SVG, and GIF, and is MCP-compatible, meaning an animated frame can pass straight to a coding agent. [3] Generative plugins let non-developers turn prompts into shareable tools. Agent skills let teams bottle repetitive work into reusable context with connectors, web search, and file attachments. [1] Each of those is a workflow decision with a billing consequence somewhere in the plan matrix.
For Figma's enterprise customers, the questions are procurement questions. Does a motion-heavy team need different seats than a prototyping team. How do AI credits accrue across an organization, and what happens when they run out mid-sprint. Which capabilities stay beta-labeled for government and regulated customers, and for how long. None of these answers appeared on the Config stage; all of them sit in the help center. [4]
The competitive subtext is equally material. Adobe, Canva, and the code-editor vendors all sell adjacent control planes. By pulling implementation artifacts onto the canvas, Figma is betting designers will pay for proximity between intent and artifact. If the bet lands, the design tool becomes the delivery mechanism; if it does not, the credits go unused and the seats get audited at renewal.
Config's sizzle reel will age in days. The seat table will govern budgets for years. Read the second document, not the first.