TechCrunch: At HumanX conference, everyone was talking about Claude

The HumanX conference, held at San Francisco’s Moscone Center in April 2026, brought together 6,500 executives, founders, and investors for several days of discussion on how AI is changing business operations. The defining observation across multiple reports from the event: OpenAI no longer dominates enterprise AI conversations the way it did a year ago. At HumanX, that attention had shifted to Anthropic, with Claude Code cited most frequently as the AI tool that engineering and product teams are actually using in production.

During the conference, Anthropic announced that its run-rate revenue has passed $30 billion and that the company is now valued at $380 billion. Midway through the event, Anthropic also unveiled a new model called Mythos — described by observers as significantly more capable than its current public models, and reportedly not yet released due to concerns about potential misuse in security-sensitive contexts.

Claude Code’s prominence at HumanX reflects a shift in how enterprise teams evaluate AI tools. While OpenAI captured early consumer attention with ChatGPT, Anthropic has built credibility with engineering and product teams who need AI systems that behave predictably across long-running, multi-step tasks. Claude Code’s agentic coding workflow — completing development tasks with relatively low supervision — has made it a practical daily tool for many teams, which translates into organizational familiarity and preference at the procurement level.

Why it matters for product managers

When 6,500 enterprise decision-makers at a major AI conference are talking about one company’s tools more than any other, that is a leading indicator of how vendor preference conversations will play out over the next several months. For PMs involved in AI platform selection, API choices, or evaluating which tools their engineering teams will want to use, the HumanX signal is worth tracking. It also reinforces that enterprise AI trust is being built less through marketing and more through consistent performance in production environments — which is a different evaluation criteria than benchmark scores.