Investor Jason Calacanis said on the All‑In podcast that companies such as ElevenLabs, Figma, and Lovable — reported top clients paying $50–100 million a year — are planning to leave OpenAI and Anthropic. The stated reason is loss of trust: these customers no longer believe their suppliers won’t become their largest competitors.
What triggered the concern
Observers pointed to Anthropic’s shift from being primarily a model supplier toward becoming a full‑stack product company, expanding into areas such as design, coding, legal, and finance. When Anthropic launched Claude Design, which was interpreted as a direct competitor to Figma, Figma’s stock fell, and Figma’s CEO characterized Anthropic’s communications as “not consistently candid.”
A concrete lesson for enterprise customers
The episode includes a practical demonstration of the risk: the supplier itself provided the lesson. According to the account, Figma paid Anthropic millions of dollars per month, and that relationship showed how a platform provider can learn a client’s business and then offer a similar product nearby.
The point made is not that Anthropic necessarily acted in bad faith, but that there is a structural problem: any platform that controls both the underlying model and the customer-facing storefront has inherent incentives to enter the markets of the customers who fund it. Pressure to grow revenues, including from potential IPO considerations, can reinforce those incentives.
Why this matters for the market
- For large customers (contracts worth tens to hundreds of millions annually), the central issue is whether reliance on a supplier is sustainable if that supplier might become a direct competitor.
- The dual role of being both platform and provider can create unavoidable conflicts of interest, regardless of individual companies’ intentions.
- The case highlights the need for contractual, technical, or corporate governance measures to mitigate the risk that a supplier will compete with its clients.
In summary: the Claude Design–Figma episode illustrates that trust between AI model platforms and large customers can be eroded by structural incentives, making it a material strategic consideration for enterprises and investors.
