Anthropic's model, capacity, security woes before IPO
SkimNews Take
Anthropic's rapid revenue growth and impending IPO are creating operational challenges, suggesting that the demands of scaling a leading AI company are distinct from those of pure research and development.
Get the Tech newsletter
Daily tech — startups, AI labs, chips, the launches that shape the next decade. Free.
- Anthropic's revenue has tripled to $30 billion this year, driven by its popular coding tools.
- Anthropic's flagship model Opus 4.6 faced criticism for perceived performance decline, while the newer Opus 4.7 delivered benchmark gains but higher token costs and bugs.
- Anthropic's compute capacity is strained, with surging demand causing tighter usage limits and occasional outages for Claude users.
- Anthropic inadvertently exposed internal Claude Code files in a software update, prompting concerns about internal safeguards and investigation into unauthorized access to its Mythos model.
- Claude Code was removed from the $20/month Pro plan, a limited test that sparked user confusion and fears of broader pricing changes.
- OpenAI is capitalizing on Anthropic's issues, with a leaked memo from CRO Denise Dresser accusing Anthropic of overstating revenue and CEO Sam Altman mocking its marketing.
- Anthropic is eyeing a potential IPO that could value it near $800 billion, while its product and security challenges risk eroding its reputation ahead of the race with OpenAI.
Why it matters: Developers and enterprise users lose confidence as Anthropic's model performance, capacity limits, and security breaches erode trust ahead of a potential $800 billion IPO, while OpenAI gains a competitive edge by highlighting these flaws and courting the disgruntled developer base in the AI market.
