Figma battles activist pressure and platform risk after partner Anthropic launches competing tool

At the annual Config conference, Figma took the stage to showcase a dramatic reimagining of its design workspace, transforming the traditional canvas into an interactive environment where code, animation, custom shaders, and autonomous AI agents live together[1][2]. The announcements paint a picture of a company eagerly embracing the generative artificial intelligence revolution, aiming to turn raw, prompt-based user interface generation into a highly collaborative, multi-agent multiplayer experience[1][2]. Yet, beneath the flashy product demonstrations lies a profound strategic vulnerability. The core cognitive engines driving these new canvas features do not belong to Figma; they are rented via application programming interfaces from external frontier research labs[1]. This structural dependency places Figma in a highly precarious position, squeezing its financial margins just as one of its primary artificial intelligence partners has emerged as a direct, aggressive competitor in the digital design market[1][3][4].

This reliance on rented intelligence introduces a severe economic challenge for Figma as a newly public company[5][6]. While features like Figma Make, which generates functional prototypes from natural language prompts, have seen explosive adoption among the platform's largest enterprise customers, the cost of running these models remains a heavy burden on Figma's balance sheet[7]. Generating layouts, self-correcting visual hierarchies, and running multiple agents simultaneously on a multiplayer canvas requires massive amounts of real-time cloud computation[1][8]. Because Figma must pay third-party API providers for every single query and token processed, its cost of goods sold rises in tandem with user engagement. Financial disclosures from earlier this year revealed that while Figma experienced massive revenue growth, reaching hundreds of millions of dollars in a single quarter, it remains unprofitable, posting significant net losses due to skyrocketing research and development expenses alongside heavy stock-based compensation[7][9][10]. The financial reality highlights a broader dilemma in the enterprise software sector: without proprietary foundational models, software vendors risk turning their high-margin subscription models into low-margin reseller agreements for compute[4].

The risk of this platform dependency transitioned from a theoretical business concern to an immediate existential threat when Anthropic, a key model provider for Figma, launched a direct competitor called Claude Design[11][12]. This new tool allows users to generate polished user interfaces, design systems, slide decks, and prototypes through a simple conversational interface, integrating deeply with Anthropic's coding agent, Claude Code[13][14][15]. The launch exposed a deep rift between the two tech companies, catching the design software industry off guard[16]. Only days before Claude Design was officially announced, Anthropic Chief Product Officer Mike Krieger resigned from Figma’s board of directors, a seat he had occupied for only about a year[10][17]. The sequence of events prompted Figma Chief Executive Officer Dylan Field to comment publicly at an industry event that Anthropic had not been consistently candid in their communications[18][19]. This friction underscores a growing trend of platform risk, where frontier artificial intelligence labs initially act as infrastructure partners to vertical software companies, only to aggressively move up the stack and launch competing consumer-facing applications using the very technology their partners rely upon[20][21].

This sudden competitive threat has not escaped the notice of Wall Street, sparking sharp criticism and intense pressure from activist shareholders[3]. Shortly after the launch of Anthropic’s competing design tool, Findell Capital Management issued a public letter and a detailed report to Figma's board of directors and executive leadership, arguing that the company was significantly undervalued and suffering from severe governance and spending issues[22]. The activist investor urged Figma to radically streamline its product portfolio, advising the company to focus exclusively on four core pillars: Design, Dev Mode, FigJam, and Make[22][10]. Furthermore, Findell raised serious conflict-of-interest concerns regarding Figma’s board composition, pointing out that multiple board members remained material investors in Anthropic even after the resignation of Mike Krieger[23]. The investment firm demanded an independent investigation to determine whether Anthropic had improperly benefited from Figma's confidential product roadmaps during their partnership, emphasizing the need to cut research and development expenditures to industry standards to restore profitability[6][23].

Figma’s defensive strategy at Config relies not just on legal or financial maneuvering, but on a distinct philosophical pivot regarding the future of creative labor[24]. Rather than positioning its platform as a tool for total automation, Figma is betting heavily on the enduring value of human judgment and aesthetic taste[25][26]. Executive leadership has consistently argued that when artificial intelligence commoditizes code and basic layout generation, the true competitive differentiator for any product becomes the human craft, point of view, and moral reasoning that cannot be replicated by a machine[7][27]. The newly introduced AI agents are designed to function as collaborative sidekicks rather than autonomous replacements, handling tedious visual iterations, error states, and responsive scaling across breakpoints while leaving high-level strategic decision-making to the designer[1][28]. By anchoring the design process in human curation, Figma hopes to convince both creators and enterprises that a static, prompt-generated user interface is insufficient, and that the collaborative multiplayer canvas remains the essential arena where real product thinking happens[1][29].

Ultimately, the high-stakes chess match between Figma and its AI providers illustrates the defining conflict of the current technological era: the battle between those who own the underlying cognitive models and those who own the end-user workflow[12][20]. Figma possesses a massive, deeply entrenched moat of user workflows, enterprise integrations, and collaborative habits developed by millions of designers over a decade[22][4]. However, as frontier models gain higher-resolution visual reasoning and direct code-integration capabilities, the friction of generating entire software applications directly from a prompt continues to dissolve[30][8]. If Figma can successfully navigate its cost structures, pacify activist investors, and prove that workflow-embedded human judgment is more valuable than raw machine generation, it will establish a highly defensible blueprint for vertical software in the age of intelligence[25][4]. If the underlying models continue to commoditize the user interface itself, Figma’s struggle may stand as a prominent warning of how quickly the foundations of the software industry can be rewritten by the very partners that power them[20][21].