Thousands converged in downtown San Francisco this week for HumanX, one of the year’s biggest AI conferences, proving people still prefer interacting with humans rather than AI agents.
Talking to VCs and founders inside San Francisco’s Moscone Center, the consensus was clear for most: Anthropic is the new Silicon Valley favorite. That sentiment is a sharp contrast with the first HumanX last year, held in a Las Vegas casino, where most VCs were placing their chips on OpenAI.
“In Vegas last year, it felt like OpenAI was the clear winner, and now it seems like Anthropic is miles ahead,” said Roseanne Winsek of Renegade Partners. “The Anthropic product is so good.”
Last year, Anthropic had not yet widely released Claude Code or Claude 4. Now, Claude Code is a phenomenon. Both companies are preparing to go public, and Anthropic is releasing models that are the envy of the industry. Valued at $380 billion, some VCs also see it as a better bargain compared to OpenAI’s rich $852 billion valuation, especially with Anthropic announcing this week that its run-rate revenue surpassed $30 billion, up from $9 billion at the end of 2025.
“They’re crushing it,” said Jared Quincy Davis, founder and CEO of Mithril, an AI cloud platform, referring to Anthropic. “It’s pretty clear that the focus that they had on enterprise, on frontier capabilities, on coding, and making deliberate decisions not to go into some consumer use cases, were great decisions”
OpenAI and Anthropic did not respond to requests for comment.
While there was universal praise for Anthropic, it was hard to find anyone saying good things about OpenAI, whether because of bewilderment over its recent acquisition of the internet talk show TBPN or questions about CEO Sam Altman’s deal with the Pentagon. Most founders and VCs were still reluctant to criticize OpenAI on the record.
“There are quite a few people who have disagreed with Sam and what he’s been doing,” said Andy Chen, a former partner at Coatue and Kleiner Perkins, who is expecting a brain drain of talent from OpenAI. “And Anthropic has tripled its revenue in the past three months.”
As if Anthropic did not already have enough momentum, midway through the conference, it announced its latest model, Mythos, which it said is so powerful it cannot yet be unleashed upon the general public because of the risk of cyber attacks.
“The Mythos model is a huge deal, said Tomasz Tunguz, founder and general partner of Theory Ventures. “There’s a tremendous amount of excitement.”
This year’s HumanX was twice as large as last year, with some 6700 attendees paying upward of $4,000 a ticket for a chance to rub shoulders with industry heavyweights like Lovable cofounder Anton Osika or billionaire venture capitalist Vinod Khosla.
On a vast exhibition floor, startups building AI agenetic security startups and autonomous workflows handed out branded hoodies, water bottles, and notebooks. Each day, attendees received an AI-generated list of people they should have lunch with and which sessions to attend. (The list I got was not very helpful, as it spit out a list of other VC journalists to meet.)
Robot humanoids and dogs roamed the floor, but nearby, the real dogs with their real fur at the “HumanX Dog Park” proved far more popular.
There was also a “retro lounge” with a pinball machine and jukebox, a mock New York City bodega, and a wellness lounge offering massages. The non-AI attractions seemed designed to calm the nerves of anxious attendees worried about what the AI future might hold.
“The mood I’m feeling is exuberance and existential terror,” said Stefan Weitz, a former Microsoft executive who is co-founder and CEO of HumanX. “I can’t reconcile the two.”
The rapid shift in sentiment from last year reflects the dizzying pace of AI advancement, leaving VCs exhausted.
“Every day you wake up and something has meaningfully changed,” said Tunguz. “Everyone is in a rush because everything is changing so fast. “
With how quickly models are advancing, no one was ready to count OpenAI out. When HumanX returns to Las Vegas next year, the betting favorite could have changed.
“These things change so fast,” Winsek said. “OpenAI will probably be back.”
Read the full article here
