Figma CEO Dylan Field: Anthropic Lacks Consistent Transparency in Product Communication
Dylan Field, CEO of Figma, stated during a closed-door discussion aimed at AI entrepreneurs and investors that Anthropic, when launching its new design product, "did not maintain consistent transparency" in communication, implying a gap between the two parties regarding product direction and collaboration expectations. English reports and second-hand accounts indicate that Figma previously integrated Claude into its own products, viewing Anthropic as an ecosystem partner, while Anthropic subsequently launched Claude Design, a potentially competing tool, which the market interpreted as a shift from collaboration to direct competition.
Discussions surrounding Claude Design have rapidly gained traction in design and investment circles, with several English commentaries positioning it as an AI-native design entry point that allows non-designers to generate prototypes, presentations, and pages through text, posing a structural threat to traditional tools like Figma. Following the release of the related product, Figma's stock price briefly fell, and multiple analysts and practitioners began discussing the impact of "model providers entering the application space" on SaaS ecosystem partnerships.
Source: Public Information
ABAB AI Insight
Dylan Field's remark about "not maintaining consistent transparency" points to a structural change in the relationship between AI foundational model companies and application-layer SaaS: shifting from "bottom supply + top ecosystem" to "integrated from bottom to top." When Anthropic serves as both a model supplier for Figma and launches Claude Design, which directly overlaps with Figma's functionalities, the traditional "platform-developer" cooperation logic is disrupted, and partners begin to turn into potential substitutes.
From a business structure perspective, this reflects the AI era's version of the "in-house application development" issue: model companies control computing power, distribution, and the speed of capability iteration. Once they decide to move up to the application layer, they can easily squeeze the profit margins of SaaS companies that rely on their APIs. For tools like Figma, which are built on high human design processes, Claude Design's approach of "directly generating usable drafts" effectively bypasses traditional design workflows, shifting the value chain focus from "design process" to "describing intent + machine generation," thus re-evaluating the profit position of design tools.
On a deeper level, this represents a "trust structure reconfiguration" in the entire software ecosystem under the impact of AI. In the past, SaaS could view cloud vendors and API providers as relatively stable infrastructures, but now, foundational model companies inherently possess a strong strategic impulse to "consume the ecosystem upward," blurring cooperation boundaries and diminishing the credibility of verbal commitments and long-term roadmaps. In such a landscape, "whether to honestly disclose the real product roadmap" is no longer just a matter of business ethics, but a core variable in how partners price their own risks and whether to continue building critical functionalities on external models.
From a historical and financial perspective, this is akin to the "Sherlock" effect of Apple and Google on third-party apps during the mobile internet era, but this time it occurs at a higher level of intelligence: model providers can quickly replicate most application value with minimal manpower and strong models. For companies like Figma, the challenge is not just an additional competitor, but a change in the entire industry rules—whoever controls general intelligence and computing power has the authority to rewrite "who is the platform and who is the plugin" at any time.