ANTHROPIC JUST PRINTED ITS FIRST DOLLAR OF PROFIT. THE UNDERDOG AI LAB IS NOW A MONEY MACHINE.
Read the full story at TechCrunch
OK so here is a story that nobody saw coming four years ago. Anthropic, the company that was basically founded because a bunch of people walked out of OpenAI thinking Dario Amodei and his sister Daniela could do AI safety better, is about to post its first ever profitable quarter. Let that sink in. The safety-first, slow-and-careful, let-the-adults-talk AI lab is now the one bringing in actual money.
The numbers are ridiculous. In Q1 2026, Anthropic made $4.8 billion in revenue. That is already insane. Then they told investors Q2 is tracking toward $10.9 billion. In one quarter. That is more than double in three months. And the operating profit? $559 million. The company has been burning cash since 2021 and now it is going to post profit for the first time in its entire existence.
Now before we get too excited, there is a catch, and the catch is a big one. Anthropic itself said do not expect this to last. The reason they will be profitable in Q2 is basically a timing thing. They have not yet fully loaded up on the compute they have committed to buying. Once those bills come due, and they will come due, the losses return. Anthropic has committed to spending an absolutely terrifying amount on infrastructure over the next several years, which is why they keep raising money at ever-higher valuations approaching a trillion dollars.
But here is why this still matters. Profitability, even for one quarter, changes how people look at you. It means the business model actually works. It means companies are paying real money for Claude and getting enough value out of it to keep writing checks. It means the revenue growth is not just venture capital being shoveled out the window but real enterprise demand. That is a different category of company from the Anthropic of 18 months ago that nobody outside of serious tech people had even heard of.
The timing of the announcement is also worth noting. This news dropped the same day OpenAI announced it is barreling toward a September IPO. So on the same day, you have OpenAI getting ready to go public and Anthropic quietly dropping the bombshell that it is already profitable. Whether that was intentional or just cosmic coincidence, it lands like a statement. The IPO is the headline everyone covered. The profitable quarter is the number that actually matters more long-term.
The story of Anthropic’s rise is genuinely something. In May 2025, a little under 10 percent of businesses were paying for Anthropic products. By May 2026, that number had climbed over 25 percentage points. OpenAI, in the same period, lost market share. You have one company going from a niche safety-focused research lab to the business AI of choice, and another company that invented the modern chatbot watching companies quietly switch lanes. The question now is whether Anthropic can hold onto this momentum when the enormous compute bills start arriving in Q3. But for now, the quiet kid in the room just had a very loud moment.
CLAUDE JUST DETHRONED CHATGPT AMONG PAYING BUSINESSES. THE RECEIPTS ARE BRUTAL FOR OPENAI.
Read the full story at TechCrunch
The data comes from Ramp, which is a corporate expense management company with over 50,000 businesses running their expenses through it. Every time a company pays for an AI subscription using Ramp, it shows up in their data. It is about as close to ground truth as you can get without literally breaking into the accounting departments of every company in America.
And the numbers say this: for the first time ever, more businesses are paying for Anthropic than for OpenAI. Specifically, 34.4 percent of Ramp clients are paying for Anthropic products, versus 32.3 percent for OpenAI. It is not a massive gap on paper, but what makes it significant is the direction of travel and the speed at which it happened.
In May 2025, Anthropic had only 9 percent of businesses. One year later, it has 34.4 percent. OpenAI in the same period went from roughly 33 percent to 32.3 percent. So Anthropic grew by more than 25 percentage points and OpenAI lost ground. That is not a trend. That is a rout happening in slow motion, in public, with nobody at OpenAI apparently stopping to fix whatever is causing it.
The economist at Ramp who analyzed the data put it pretty bluntly. Anthropic started with the most demanding, technical customers: the finance teams, the lawyers, the developers who live inside their tools and can immediately tell when something is sharper or sloppier. Those people moved to Claude first. Then the rest of the market started following. This is how technology market share actually shifts in practice. It starts at the edges with the most demanding users and works its way to everyone else.
What is funny is how fast this happened. ChatGPT was the name everyone knew. OpenAI was the company that defined the category. They were the iPhone of AI. And somehow in about 12 months, a company that most normal people have never heard of overtook them in enterprise adoption based on data from 50,000 real businesses with real credit cards.
OpenRouter, which tracks AI model usage across developers, also shows Anthropic above OpenAI. Multiple independent signals point the same direction. Part of what happened is that Claude became the coding model of choice for serious developers. Once your engineers are using Claude every day and telling their managers it is better, the enterprise contract tends to follow. The tool wins bottom-up adoption and the business decision gets made from the top. OpenAI built the product that made the world care about AI. And now, because it spent the last year dealing with board drama, restructurings, lawsuits, and an IPO prep, a company founded by its own former employees is quietly eating its lunch one corporate credit card at a time.
ELON MUSK’S AI LAB IS RUNNING 46 ILLEGAL GAS TURBINES IN MISSISSIPPI AND CALLING THEM TRUCKS. NAACP WANTS THEM SHUT DOWN NOW.
Read the full story at TechCrunch
You genuinely cannot make this one up. Elon Musk’s AI company, xAI, the one whose Colossus supercomputer he keeps bragging about on X, is operating 46 natural gas turbines at its Mississippi data center. These are not small backup generators. These are industrial power plants, each one mounted on a flatbed trailer. And here is the beautiful part: because they are on trailers, the state of Mississippi considers them mobile, and mobile equipment is exempt from air pollution regulations for a full year.
So you have a company running what are functionally permanent power plants, each one burning natural gas and pumping emissions into an area that is already one of the more polluted regions of the South, and the entire regulatory framework is being sidestepped because technically, the machines have wheels. It is the kind of scheme you would see in a movie about a company that really needs electricity and really does not want to fill out environmental permits.
The NAACP, which filed a lawsuit on behalf of local residents, asked a court for an emergency injunction to shut the operation down. The legal challenge is being handled by the Southern Environmental Law Center, which says the turbines are operating in violation of federal law. Under the Clean Air Act, equipment mounted on a trailer can still be classified as stationary if it is being operated in a fixed location for an extended period, which these obviously are.
The Greater Memphis Chamber of Commerce put out a press release not long ago noting that about half of the 35 turbines then in operation would remain on-site permanently. The rest were supposed to be temporary. That was then. Now there are 46. The count went up, not down. The temporary turbines just sort of became permanent. Funny how that works.
The region around the data center is not some abstract location on a map. It has a history of industrial pollution and a population that has been fighting against exactly this kind of environmental abuse for decades. The NAACP did not choose this fight randomly. They are representing people who live and breathe near these machines every single day, people for whom this is not a news story but a public health problem happening in their own backyards.
And the company on the other side is not some anonymous industrial operator. It is the AI company of one of the most famous, powerful, and politically connected men on the planet. The same man who tweets constantly about humanity’s future and his commitment to making life better for everyone. Running 46 machines that are, according to the lawsuit, violating federal air quality law in one of the most historically overburdened communities in the country. The case is in court. Whatever the legal outcome, the situation speaks for itself.
NINE CALIFORNIA JURORS NOW HOLD OPENAI’S FATE. HERE IS EXACTLY WHAT THEY HAVE BEEN ASKED TO DECIDE.
Read the full story at TechCrunch
If you have been following the Elon Musk versus Sam Altman trial and feeling vaguely confused about what anyone is actually arguing about, you are not alone. It has produced a staggering amount of testimony, leaked documents, personal drama, and courtroom theater. But at the end of all of it, there are nine regular California residents in a jury room who have to answer a fairly narrow set of legal questions. And those questions could determine whether OpenAI is forced to unwind its for-profit structure entirely.
Here is what is actually happening. Musk donated millions to OpenAI back in 2015 when it was founded as a nonprofit dedicated to ensuring AI benefited humanity rather than being controlled by any single company or person. Years later, OpenAI created a capped-profit arm that took billions in investment from Microsoft and has been growing into one of the most valuable companies in the world. Musk says his money was donated for the nonprofit mission and was instead used to enrich the founders and Microsoft. He wants the structure unwound.
The jury is being asked three questions. First: did OpenAI violate the specific terms under which Musk made his donations? Second: did anyone get unjustly enriched as a result? Third: did Microsoft, knowing all of this, help cause the problem? These sound dry on paper but the stakes attached to them are anything but. A verdict for Musk could potentially force OpenAI to restructure, or put the company in a situation where nobody is quite sure what happens next.
OpenAI’s defense is that Musk knew exactly what he was signing up for, that private investment was always going to be necessary to reach AGI, that his donations were fully spent by 2020, and that the for-profit structure still serves the nonprofit mission because ChatGPT is free for billions of people. Their lawyers also pointed out that Musk himself tried to take over OpenAI’s for-profit structure at one point, and was hiring away OpenAI researchers for his own AI projects while still serving on the board. The legal doctrine of unclean hands was invoked, which is the lawyer’s way of saying you do not get to sue someone for doing what you were also doing.
The wildest detail to surface from the testimony is that Altman said Musk once floated the idea of giving OpenAI to his children. Not as a joke, apparently. That was a real option someone considered at some point. Which tells you something about the state of everyone’s thinking in the early years of this project, and also maybe explains a few things about how we got here.
OpenAI’s lawyers argue that forcing a restructuring eight years after the original arrangements were made is both legally unreasonable and potentially catastrophic for a company that now employs thousands of people and is approaching a public listing worth hundreds of billions of dollars. Musk’s lawyers say the harm started when Microsoft put in $10 billion in 2023 and that the timing is fine. Whatever the jury decides, the AI industry is watching very closely. The answer to who owns the mission of the most important AI company in history turns out to be exactly as messy and human and weird as you would expect it to be.