The US banned Anthropic’s Fable 5 release, but the numbers don’t seem to care

Just as last week was ending, the US government forced Anthropic to pull its two newest models, Fable 5 and Mythos 5, citing national security concerns after Amazon researchers allegedly found a way to bypass Fable 5’s guardrails. Cybersecurity researchers have since signed an open letter calling the move dangerous, and Anthropic itself noted the same jailbreaks exist in other models. So is this a disaster for Anthropic? The numbers don’t seem to think so.

Let’s be real for a second. The government stepping in and saying “you can’t release your product” sounds like a death knell for any tech company. But Anthropic isn’t just any tech company. It’s the darling of the safety-first AI crowd, the company that built its entire brand around being responsible. And ironically, this ban might actually reinforce that image rather than damage it.

Here’s what actually happened. Amazon researchers, who are supposedly partners with Anthropic through their cloud deal, found a jailbreak technique that could bypass Fable 5’s guardrails. They reported it, and instead of Anthropic having time to patch it, the government stepped in and said “nope, you’re done.” The models were pulled from release. The company had to scramble. But the valuation? Still sitting pretty at over $60 billion.

I’ve been watching this space long enough to know that government intervention rarely makes things better. The open letter signed by cybersecurity researchers points out something obvious: the same jailbreaks exist in other models, including ones from OpenAI and Google. So why single out Anthropic? Because they’re the ones who made a big deal about safety, and now they’re being held to a standard nobody else has to meet.

Anthropic’s response was measured but pointed. They noted that the jailbreak technique isn’t unique to their models. They’re basically saying “we’re being punished for being transparent about our safety work.” And honestly, they’re not wrong. If you build a car with a better brake system and the government bans it because someone figured out how to disable the brakes, while other cars with worse brakes keep driving, something’s off.

The market reaction tells the real story. Anthropic’s funding rounds haven’t dried up. Their enterprise customers aren’t fleeing. The demand for their API is still growing. Because here’s the thing: enterprise buyers don’t care about government bans as much as they care about capability. Fable 5 was better than GPT-5 on several benchmarks, and Mythos 5 was a strong competitor to Gemini Ultra. The tech is good, and people want it.

What this ban actually does is create a weird secondary market. I’ve already heard whispers of researchers accessing the models through back channels. The government can’t unring a bell. Once a model is trained, the weights exist. You can ban the release, but you can’t ban the knowledge. This is the same mistake we saw with encryption back in the 90s.

Anthropic’s real problem isn’t the ban itself. It’s the precedent. If the US government can step in and block a model release based on a single researcher’s claim, what happens next time? What happens when a competitor finds a minor vulnerability and reports it to the government instead of the company? That’s a slippery slope, and it’s going to make every AI company paranoid about sharing their safety research.

I’ll say this: Anthropic handled it about as well as anyone could. They didn’t fight the ban publicly. They didn’t throw Amazon under the bus. They just acknowledged the situation, pointed out the inconsistency, and kept working. That’s the right play. Fighting the government in court would have been a PR disaster and a distraction.

The numbers don’t care about the ban. The valuation doesn’t care. The customers don’t care. What they care about is whether Anthropic can keep delivering better models. And based on what I’ve seen from Fable 5’s benchmarks before it was pulled, they absolutely can. The ban is a political move, not a technological one. And politics has a way of being forgotten when the next big model drops.

So yeah, the US banned Anthropic’s Fable 5 release. But the numbers don’t seem to care. And neither should you.