TL;DR
- New Product: Anthropic launched Claude Design, a conversational tool that turns prompts into prototypes, decks, and UI mockups on top of Claude Opus 4.7.
- Competitive Target: The product directly challenges Figma’s UI-design surface and extends a two-year partnership with Canva into product-level integration.
- Access: Claude Design is included for Pro, Max, Team, and Enterprise subscribers at claude.ai/design, drawing from weekly token limits with pay-as-you-go beyond the cap.
- Market Impact: Figma shares dropped another 5% on launch day, three days after CPO Mike Krieger resigned from Figma’s board on April 14.
- Strategic Shift: Anthropic is moving from foundation-model provider into the application layer as annualized revenue surpassed $30 billion by early April.
Anthropic is pushing Claude Opus 4.7 directly into Figma and Canva territory. Claude Design, launched yesterday by Anthropic Labs, turns natural-language prompts into prototypes, decks, and UI mockups for Pro, Max, Team, and Enterprise subscribers. Rollout proceeds gradually to paying Claude plans over the course of the day, with access included at no extra cost within existing subscription limits.
With the launch, as Anthropic Labs detailed, Anthropic makes its clearest move from foundation-model provider into the application layer owned by Figma, which the company says commands an estimated 80 to 90% market share in UI and UX design, and by Canva, which now counts 265 million monthly active users.
Moreover, it lands three days after CPO Mike Krieger resigned from Figma’s board. Figma’s stock, already down roughly 50% over the past year, lost another 5% after the launch.
Meanwhile, unsolicited investor offers valuing Anthropic at around $800 billion, more than double the $380 billion mark set in a February 2026 funding round, have sharpened the stakes further.
What Claude Design Does
Claude Design lives at claude.ai/design and operates as a conversational surface on top of Claude Opus 4.7, which Anthropic describes as its leading generally available vision model.
During onboarding, Claude builds a design system for the team by reading its codebase and design files, then applies that team’s colors, typography, and components to every subsequent project. Teams can maintain more than one design system in parallel.
Inputs span text prompts, uploaded images and documents in DOCX, PPTX, and XLSX formats, codebase references, and a web capture tool that pulls elements directly from a live website. Users can adjust output through chat, inline comments, draw-on-design edits, or custom sliders that manipulate images, colors, and data in real time. A “Frontier Design” mode extends the canvas to code-powered prototypes with voice, video, shaders, 3D, and built-in AI.
Completed work can be exported to Canva, PDF, PPTX, or standalone HTML, shared via an organization-scoped URL, or packaged into a handoff bundle that Claude Code can turn into implementation with a single instruction. Anthropic says Claude Design is built for founders and product managers who need to share ideas visually, while giving experienced designers faster paths to explore multiple directions.
Beta partners describe sharp drops in the effort required to turn static mockups into interactive prototypes.
“Brilliant’s intricate interactivity and animations are historically painful to prototype, but Claude Design’s ability to turn static designs into interactive prototypes with sophisticated animations, and to iterate on them through natural conversation, is a genuine step-change for our team.”
Olivia Xu, Senior Product Designer at Brilliant (via Anthropic)
Building on this, Brilliant’s team paired that capability claim with a quantified one: the company’s senior product designer reported that complex pages needed 20 or more prompts to recreate in competing tools but only 2 in Claude Design. Such a 10-to-1 ratio is a handpicked beta figure, but it names the metric Anthropic is competing on: compressing prompt iterations is what turns a chatbot surface into a design tool serious teams will use for real work.
Research Preview and Access
Access is included with Claude Pro, Max, Team, and Enterprise paid plans and draws from those subscriptions’ weekly token limits, with pay-as-you-go usage available beyond the cap. For Enterprise organizations, Claude Design ships off by default and must be enabled by an administrator in organization settings, with granular controls over who receives access. When users link a local codebase, the files themselves are not uploaded to or stored on Anthropic’s servers; Claude Design retains only the design-system representation it generates.
Initial research-preview access covers Canva, Datadog, and Brilliant. Anthropic says deeper integrations with third-party tools are due in the coming weeks.
However, distribution itself is a strategic choice. Rather than launching as a standalone freemium product and chasing Canva’s scale, Anthropic is anchoring Claude Design inside plans already populated with developers, founders, and engineering-adjacent teams: the exact cohort many expect to bring prompt-to-prototype habits into an organization’s design pipeline.
The Figma and Canva Flashpoint
Anthropic’s competitive framing is deliberately asymmetric. Anthropic told TechCrunch that the product is meant to complement Canva rather than replace it, aimed at users who aren’t starting inside a design tool.
Meanwhile, Canva itself is leaning into the framing. Already two years old, that partnership now threads Claude Design into Canva’s Design Engine and Visual Suite so that output can land in Canva as fully editable, on-brand assets. Anthropic timed the announcement to Canva AI 2.0 at the Canva Create event in Los Angeles on April 16.
Canva framed the launch as an extension of the partnership. Co-founder and chief executive Melanie Perkins said drafts from Claude Design land in Canva as “fully editable and collaborative designs ready to refine, share, and publish.”
In contrast, Figma’s picture is sharper. Krieger’s exit from Figma’s board on April 14 removed an awkwardness that had been building for months. Figma’s own counter, Code to Canvas, shipped in February 2026 and converts output from tools like Claude Code into editable Figma designs. That move now looks defensive rather than preemptive.
Prior coverage tracked Figma Make reaching general availability last summer as the incumbent’s own prompt-to-prototype play.
Signals cumulate against Figma: shares down $25 from their IPO price, another 5% drop on launch day, and a two-year Anthropic-Canva partnership suddenly thickening into product integration, positioning Figma as the incumbent whose editor surface Anthropic is specifically pricing against.
Datadog’s product team framed the workflow impact in concrete terms.
“Claude Design has made prototyping dramatically faster for our team, enabling live design during conversations. We’ve gone from a rough idea to a working prototype before anyone leaves the room, and the output stays true to our brand and design guidelines.”
Aneesh Kethini, Product Manager at Datadog (via Anthropic)
As a result, partner testimony cuts in two directions. It validates the capability claim and it underscores which review cycles Anthropic is trying to collapse: the same ones Figma’s editor has monetized for a decade.
Anthropic’s Application-Layer Push
Claude Design sits inside a broader strategic shift. Annualized revenue hit roughly $20 billion in early March and surpassed $30 billion annualized revenue by early April, up from $9 billion at the end of 2025.
Moreover, Anthropic is in early talks with Goldman Sachs, JPMorgan, and Morgan Stanley about a potential IPO as soon as October 2026.
The just released Opus 4.7 engine powering Claude Design reached 64.3% on SWE-bench Pro and delivered a 13% resolution improvement over Opus 4.6 on Anthropic’s internal coding benchmark. That upgrade also roughly tripled image resolution to 2,576 pixels on the long edge, the vision-model headroom that makes a prompt-to-prototype product practical in the first place.
Opus 4.7 arrived just one a day before Claude Design, a cadence that positions Anthropic Labs as the productization arm for frontier-model gains. Anthropic’s launch post framed the strategy as keeping the margin that foundation-model capability throws off in-house, rather than letting incumbents absorb that value into their existing editors.
Anthropic’s trajectory is unmistakable. It is no longer just supplying the model that lives inside somebody else’s design tool. It is building the tool itself and shipping it into paying enterprise hands.