The HumanX AI conference packed San Francisco's Moscone Center this week with thousands of engineers, founders, and enterprise buyers. It is the kind of gathering where you can feel, in real time, which company the industry believes in. And this year's answer was unambiguous.

Claude. Claude. Claude.

Both TechCrunch and CNBC reported the same thing independently: Anthropic dominated the conversation. Claude Code was the tool on every vendor's lips. One exhibitor told TechCrunch he made a point of mentioning that his team used Claude extensively, then added that ChatGPT "fell off." Multiple conference panels gave Anthropic shoutouts. Bloomberg wrote an entire newsletter about how Anthropic owned the event.

The ChatGPT Problem

TechCrunch's correspondent nailed the diagnosis: OpenAI's problem is not that it is building bad products. It is that the company looks reactive rather than strategic, "as if it's simply responding to events rather than shaping them."

Consider the last two months of OpenAI decisions. It killed Sora. It killed the "sexy ChatGPT" project. It abandoned a UK Stargate expansion. It injected advertising into ChatGPT. It launched a $100 subscription tier clearly aimed at stealing Claude Code users. And a devastating New Yorker profile questioned whether Sam Altman can be trusted at all.

Each of those moves makes tactical sense in isolation. Together, they paint the picture of a company that does not know what it wants to be when it grows up.

The Numbers Tell the Same Story

Ramp's AI Index data, published this week, shows Anthropic growing from 24.4% to 30.6% of US businesses in a single month. That is a 6.3 percentage point gain, surpassing the previous monthly record. OpenAI still leads overall. But the trajectory is unmistakable, and the Wall Street Journal recently noted that both companies are "the fastest-growing businesses in the history of tech."

The difference is that Anthropic is gaining with a coherent strategy: build the best coding tools, win developers, let word of mouth do the marketing. OpenAI is spending billions to figure out whether it is a consumer product company, an enterprise platform, a media company, or an advertising business.

The Bret Taylor Defense

OpenAI board chair Bret Taylor defended Altman at HumanX when asked about the New Yorker profile: "If you want to seek out detractors for him, you'll find them, and they'll be very vocal about it. I think Sam's remarkable."

That is the kind of defense you give when you cannot deny the criticism. You redirect to character testimony. Which tells you everything about how seriously the board is taking the narrative shift.

The Bottom Line

Conferences are vibes. Vibes are not revenue. OpenAI is still generating massive revenue and about to IPO. But in an industry where developer loyalty determines who wins the next cycle, losing the room at HumanX is a leading indicator. Not a lagging one.

OpenAI did not lose HumanX because Claude is better. It lost HumanX because Anthropic looks like it knows exactly what it is building. And right now, OpenAI does not.