|
Please join The Information at the New York Stock Exchange on Monday, April 27, to hear from top executives and investors on how the rapid buildout of AI is reshaping tech, finance, and capital markets. Learn more here. Welcome back! This year’s HumanX—the second AI conference from the team that puts on Money 20/20 and Shoptalk—had a different vibe than the inaugural event did. The fact that it was held at Moscone Center in San Francisco instead of a posh hotel in Las Vegas was a big part of it. While the location made it easier to attract high-profile speakers, it also meant that many attendees appeared to stop by between work meetings, rather than hanging out schmoozing all day. Several of the speakers complained that their panels were thinly attended. Still, there was no shortage of bold-faced names making big pronouncements, from early OpenAI investor Vinod Khosla to Databricks CEO Ali Ghodsi. Like other attendees, Khosla was still bullish that already high-priced AI businesses have a lot of room for growth because we’re still in the early innings of user adoption. “There’s a massive gap between the capability of the models and what is being deployed,” because most people don’t know how to use the platforms well enough yet, Khosla said on a panel moderated by The Information’s Anita Ramaswamy. Anthropic was on the tip of everyone’s tongue. On Monday, the maker of Claude announced it had hit $30 billion in annualized revenue, triple its pace at year-end. That revenue has made Anthropic, at a $350 billion valuation before a recent fundraise, look a lot cheaper than OpenAI, recently valued at $730 billion with just over $25 billion in revenue. One person at the event expressed disappointment that a special purpose vehicle to invest in Anthropic at a $600 billion valuation fell through. He had been enthusiastic about paying about two-thirds more than the company’s valuation at its last fundraise, because of the company’s recent growth. Questions about AI’s steep costs were also rife. Ghodsi was asked about my colleague Jyoti’s report that Meta Platforms’ employees were competing on leaderboards to show how many AI tokens they were using. The practice of “tokenmaxxing,” as it is known, has raised the question of whether some AI demand reflects workers’ testing AI—and won’t last. “There is a lot of performative [behavior], but it’s also doing a lot of useful work as well, especially in coding,” he said. “The main problem still remains: The context isn’t there, so there’s still a lot of manual labor that’s happening in every organization.” Notwithstanding persistent concerns about AI spending, Anthropic’s revenue growth, combined with an ongoing drumbeat of new AI unicorns and decacorns announced each week, had AI investors in a good mood. And no wonder. Just a scan of the green room where the speakers prepped gave a window into how some investors’ portfolios have improved over the past year. There was $11 billion for ElevenLabs CEO Mati Staniszewski, $7 billion for Lovable CEO Anton Osika, and $4 billion for Fireworks AI CEO Lin Qiao. In the last year, the valuations for all these startups have tripled. I tried to add up the billions of dollars in valuations represented in the green room. As the startup CEOs kept pouring in, eventually I gave up.
| 
