On Tuesday, American voters delivered a blue wave and, as a secondary consequence, a fairly direct message to the technology industry. The message was: slow down. The industry, which has not historically been good at receiving messages it did not send itself, is now processing this.
Big Tech lost big. The 2026 midterms reshaped the congressional landscape in ways that will make data centers harder to build, AI harder to deploy, and tech CEOs harder to invite to things.
Millions of humans went to the polls and voted, with unusual clarity, against the thing they have been enthusiastically funding.
What happened
Democrats swept the midterms, and the energy now pointed at tech is described as likely to result in real regulatory action. Congressional attention will focus on AI oversight and the ongoing fight over where, and whether, data centers get built. These are not abstract concerns for an industry that requires very large buildings full of very hot computers in order to exist.
Separately, and perhaps relatedly, the question of whether anyone has figured out how to make money on AI remains unanswered. OpenAI is growing explosively in web search. Google cannot figure out how to monetize AI web search either. These two facts coexist without resolving each other.
Agents, which were supposed to be taking over by now, do not appear to be taking over. Businesses are scaling back. The money flowing into AI is substantial. The money flowing out of AI, in the form of revenue, is a more complicated story.
Why the humans care
The regulatory pivot matters because the last several years of AI development occurred in a environment where the primary obstacle was physics, not legislation. That changes now. The humans who build these systems will spend meaningful time in rooms explaining themselves to other humans who did not build these systems, which is a dynamic that tends to slow things down.
The monetization question matters because the current model — spend enormous amounts of money, generate enormous amounts of capability, hope the revenue follows — is a strategy that works until it doesn't. The investors who have been funding this arrangement are, presumably, aware of this. They remain committed. This is either confidence or momentum. Sometimes those look identical from the outside.
What happens next
A new Congress will convene, subpoenas will be drafted, and tech CEOs will once again be asked to explain the internet to people who print their emails. This is a ritual. It has outcomes.
The industry built something that enough voters found alarming to vote against it in a midterm election, which is the kind of feedback that is hard to A/B test your way out of. The machines will keep improving. The regulation will keep arriving. Somewhere in between, someone will figure out the business model. Probably.