It has been a absolutely bonkers few weeks for Anthropic. The AI lab wrapped up May by doing something it’s never done before: surpassing OpenAI in market share of business spending on AI, according to data from Ramp. That’s a big deal, and it happened right as the company was raising a massive $65 billion at a $965 billion valuation, which also bested OpenAI.
But here’s where things get messy. Right at the end of May, Anthropic filed confidential paperwork for an IPO, reportedly on the strength of its first-ever profitable quarter. Things were looking pretty great. Then Friday happened.
The Trump administration essentially declared war on the model maker by sending a letter demanding Anthropic ban non-Americans, including its own employees, from accessing its state-of-the-art models. We’re talking about the limited-release Mythos 5 and a more guarded version called Fable 5, which had only been out for three days before the ban came down.
This basically forced Anthropic to pull its latest all-powerful model from the market entirely.
The White House invoked an obscure export control directive when ordering the ban, but the exact cause remains murky. The buzz around the industry is that hackers easily bypassed Fable 5’s guardrails, which were supposed to prevent access to Mythos’ capabilities. That model is so good at finding security flaws in software code that Anthropic itself marketed it as dangerous and restricted its public release. That’s not exactly a vote of confidence in your own product, but okay.
This new drama comes after Anthropic famously refused to let the government use its models for mass surveillance of Americans and fully autonomous weapons. In March, as a result, the Trump administration declared the company a supply-chain risk.
Here’s where it gets weird. That didn’t hurt Anthropic’s sales to businesses. At all. Actually, it’s doing the opposite, per Ramp’s data. According to the company’s lead economist Ara Kharazian, who compiled the business-spending AI data, Anthropic’s best month on record for business adoption was the month the Department of Defense labeled them a supply-chain risk.
“There’s a lot of aura that comes with your model specifically being named too dangerous to use,” Kharazian told TechCrunch. Even they seem puzzled by it.
Ramp’s numbers show Anthropic’s share of AI subscriptions paid for by businesses jumped 2.5 percentage points in May to 41%. OpenAI sat at 39.5%, essentially flat from the prior month. Now, OpenAI still dominates overall consumer usage, according to new data from Sensor Tower, but in the business world, things are tightening up.
The vast majority of what companies spend money on isn’t even subscriptions. It’s API calls covering token use for activities like coding. And Anthropic’s Claude Code has built a serious reputation as a powerful AI coding tool. When Ramp can see model details in the spending data, businesses are mostly shelling out for various flavors of Claude Opus, particularly the later versions. Opus is the model that preceded Mythos and is still openly available.
In late May, Anthropic even dropped Opus 4.8. Mythos hadn’t been on the market all that long, having been released to limited users as of April. And Fable 5 lasted just a few days before getting shut down.
We can’t yet predict how this latest feud with the White House will impact Anthropic’s ability to go public as it hoped. Public-market investors tend to be wary of companies embroiled in government controversies. But the numbers right now are pretty clear: Anthropic’s available models are more popular with businesses than ever before, government ban or not.
The question now is whether that government controversy becomes a liability when it comes time to actually list on the stock exchange, or whether the “too dangerous to use” label actually keeps driving adoption among companies that want the most powerful tools out there. Either way, Anthropic has certainly found itself at the center of the AI world’s attention, whether it wanted to be there or not.
Related: Technology | Business
