Figma’s Canvas Becomes a Checkpoint

Figma’s Q1 revenue rose 46% to $333.4 million. After Google updated Stitch, Figma’s stock fell 7.98% in a day and was reported roughly 80% below its August 2025 IPO level. Customers were spending more as investors priced in a weaker hold on where design begins. If an agent can make the first draft, what are teams still paying Figma to control?

Figma made the design file a coordination point

Figma entered the market by changing the architecture around design work. In 2015, it emerged as a browser-based collaborative alternative to Adobe’s desktop software. Browser access made the design artifact a shared workspace where specialists could work together.

Figma’s browser architecture made simultaneous participation, feedback, and iteration native to the object itself. Its 2019 Community release extended that logic with public profiles, sharing, remixing, and discovery of work in progress. The canvas accumulated people, conventions, and reusable work around it.

Figma made the design file the coordination point connecting designers, developers, and the rest of an organization. As teams gathered decisions and dependencies around that object, moving the pixels elsewhere left the workflow in place.

AI agents cheapen screen generation. Figma continues to hold the shared language, review process, organizational memory, and production handoff around each screen. Teams can begin in an agent and gather around Figma for coordination.

Enterprise teams keep buying Figma

Figma reported Q4 revenue of $303.8 million, up 40% year over year and above the $293.15 million estimate. Q1 revenue then reached $333.4 million, up 46% and above the $313.2 million estimate. Management raised its annual forecast and guided Q2 above estimates, citing AI monetization and adoption.

Enterprise teams keep buying seats for compatibility, review, shared context, and handoff, and Figma is monetizing AI inside that demand. They can maintain those seats after they begin new work in a model platform or coding agent. Google, Anthropic, or OpenAI can gain discovery and orchestration before Figma’s seat demand weakens.

The first draft is becoming an abundant intermediate

Google’s Stitch update made the shift explicit. Natural-language prompts can produce high-fidelity UI in an AI-native design canvas, with a reasoning design agent helping create and iterate on the result. A user starts by describing the desired interface.

Figma’s Code to Canvas, launched with Anthropic, accepts work that begins outside Figma. Users can import code generated in Claude Code and convert it into an editable, collaborative design.

Designers can move from code to design, and prompts can produce either. Figma Make likewise turns a prompt into an app.

Anthropic then revealed Claude Design, a product for creating visuals, prototypes, slides, and other artifacts. Weeks earlier, it had asked firms including Figma for input. Anthropic was preparing to originate work that its partner’s software would later receive.

Teams increasingly trade in executable design intent: a request that can become a mockup, code, or an agent-revised artifact. The same intent can move into whichever environment comes next, allowing Figma to remain the shared destination without receiving the initial request.

Agents can turn Figma into a callable capability

An agent can become the customer interface and call Figma as a tool. The user states an objective, the agent coordinates services and evaluates output, and specialist applications operate behind that interaction.

OpenAI launched Codex plugins with more than 20 initial integrations, including Figma, Notion, Gmail, and Slack, to standardize repeatable workflows. Earlier ChatGPT app pilots had already placed Figma and Canva inside a conversational interface. In both cases, the agent owns task initiation and routing.

Codex can choose tools, carry context between them, and return completed work. That makes specialist applications callable capabilities inside OpenAI’s interface.

Google and Anthropic increasingly receive the initial request, and OpenAI routes work among tools. Figma retains leverage when enterprise teams require a shared, inspectable artifact that outlives the prompt session.

Investors are paying for the front door

Adobe showed the same gap between operating performance and valuation. It reported Q2 revenue of $6.62 billion, up 13% year over year, and raised its annual revenue and profit forecasts. Its shares still fell more than 5% after hours.

Adobe, Canva, and CapCut integrated their editing tools into Gemini. Google later enabled U.S. AI Mode users to link to and interact with Canva. Gemini can now serve as the front door to specialist editors.

Investors sent Figma shares down 7.98% after the Stitch update, leaving the stock reportedly roughly 80% below its August 2025 IPO level. That equity reaction priced Google’s new distribution position immediately, while Figma’s revenue continued to record rising customer demand.

Investors are pricing distribution above feature breadth. Adobe, Canva, and Figma monetize seats, workflows, and organizational standards, so they add AI inside existing products. Google, Anthropic, and OpenAI benefit when users start and finish more requests on their platforms, so they turn specialist applications into tools. Each side can cooperate through integrations while competing for the customer relationship.

Enterprise design leads need an approval gate

In 2025, 84% of developers used or planned to use AI tools, while only 33% trusted their accuracy, down from 43% in 2024. Developers adopted the tools faster even as confidence fell.

Figma can build durable value in that gap. A design lead reviewing a Claude-generated checkout flow needs to know whether its buttons use approved components, whether interactions meet accessibility rules, who approved deviations, and which code will ship. The design lead decides whether generation saves time or pushes cleanup downstream.

Code to Canvas gives Figma a route to make code-originated work editable, legible, and collaborative while preserving the decisions around it.

Figma can encode those checks in approved components, permissions, and review decisions. Design leads can inspect agent-generated work against one authoritative set of standards before handoff.

Figma’s integrations must preserve organizational context across tools. If Stitch, Claude, Codex, or another agent can carry an interface into production without that context, embedded workflows become transition costs that slow movement without preventing it.

The moat moves to the checkpoint

Customers drove 46% revenue growth by paying for collaboration and organizational memory. Investors sent the stock down 7.98% in one day as they questioned whether Figma would still govern work that begins in Stitch, Claude, or Codex.

Figma can turn eleven years of collaborative habit into enforceable workflow authority: the place where generated intent acquires structure, provenance, approval, and production meaning. When agents must pass through its components and permissions before shipping, the canvas governs drafts made anywhere. First drafts are becoming abundant. The checkpoint is becoming scarce.