Six months ago, the AI industry looked like a classic ai bubble. Companies were pouring hundreds of billions of dollars, much of it borrowed, into new data centers, and the comparison to the railroad boom of the 1800s and the dot-com surge of the 1990s was becoming routine.
Now the story has turned. Software developers are adopting AI tools en masse, and Anthropic’s revenue is growing faster than Zoom’s did during the pandemic, Google’s did in the early 2000s and Standard Oil’s did in the Gilded Age. If that pace holds, the company would be taking in more money than any company in the world by early next year.
That shift has a face inside the product itself. Anthropic released an update to Claude in November, and with Claude Code a team of autonomous AI agents could take over a computer and complete programming tasks in minutes or hours that used to take humans days or weeks. In many cases, the final product needed few, if any, human changes. OpenAI’s Codex and Anysphere’s Cursor have since pushed out updates to their own coding tools, but Anthropic appears to have drawn the sharpest line yet between AI that talks and AI that works.
Ethan Mollick called it “a step change.” He said, “For years now, we've been in an era of chatbots that mostly just say things. Now we've officially crossed into the era of agents that can actually do things.” That is why the revenue numbers matter now, not just as a sign of adoption but as evidence that the money is finally following the machine behavior investors have been promised for years.
The pressure is showing up in workplaces that are still trying to understand what these tools mean for hiring and output. Jordan Nanos said his small team now produces four times as much software as it did last year even though the staff count has not changed. Tim Fist, meanwhile, said “it feels sort of ridiculous” to be working on his computer-science Ph.D. because “Claude can basically do 90 percent of it.”
The tension in the market is that the old fear has not vanished; it has changed shape. Investors once worried there would be too many data centers. Now they are also confronting the possibility that there may not be enough to satisfy demand. The result is a market that still looks speculative on the capital side but suddenly looks very real on the product side.
Meta’s recent plan to lay off 10 percent of its workforce drives that point home. Mark Zuckerberg told investors that AI means projects that used to require big teams can now be accomplished by a single very talented person. That is not a bubble story so much as a restructuring story, and the people building and using these tools are already living inside it.
Sam Altman saw the earlier mood clearly when he asked last year whether investors as a whole were overexcited about AI and answered, “My opinion is yes.” Six months ago, that sounded like a warning about excess. Today it reads more like a description of how quickly the industry has moved from hype to something that is changing how work gets done.