Anthropic Fable 5 is being watched by Trump, who is laughing behind the scenes? The media reveals the three major beneficiaries

Anthropic's latest two models, Fable 5 and Mythos 5, have been shut down by the Trump administration. Tech reporters analyze that behind this is not just a single event, but also the "bad boy effect" involving Amazon CEO, internet defenders, and Claude.
(Background recap: Why is US banning top Anthropic AI model exports impossible? Foreign media reveal through the "cryptocurrency war": history proves regulation always fails.)
(Additional background: Anthropic's resistance to government pressure—"a phone call blocking Fable 5"—was ineffective, and Claude is about to introduce real-name verification.)

Table of Contents

Toggle

• The White House's discovery
• The three main beneficiaries
• Also impacting Taiwan
• The "invisible war" of AI models

Recently, Anthropic was shut down by the Trump administration for its latest models, Fable 5 and Mythos 5, triggering a chain reaction. But this seemingly simple "export control" action hides a more complex game behind it.

The White House's discovery

Last Friday, the US government sent a letter to Anthropic listing "national security concerns" and demanded that the models not be used by foreigners. Anthropic simply shut down all models, as many of its employees are also foreign nationals.

What truly triggered this was a report from Amazon CEO Andy Jassy. He informed the White House that Amazon researchers discovered that the safeguards on the Fable 5 model had been bypassed. A week earlier, Amazon researchers successfully bypassed Fable 5’s protections and demonstrated to the White House the "imperfections" of the Claude model.

The three main beneficiaries

TechCrunch Equity hosts Anthony Ha, Rebecca Bellan, and Sean O’Kane analyze the three beneficiaries of this incident.

First: Amazon, with Amazon CEO Andy Jassy reporting Anthropic’s models to the White House.

Sean O’Kane pointed out that Anthropic differs from other AI research labs and doesn’t have a good relationship with the White House. He described it as "not even having made a normal phone call."

Second: Internet defenders. Several cybersecurity experts jointly wrote to Trump, demanding the withdrawal of the regulation. They argued that this would cause US defenders to lose advanced AI detection capabilities.

Third: Claude’s "bad boy effect."

Bellan said that every time Anthropic shuts down, Claude’s download numbers spike. She described it as:

Everyone loves the bad boy. They say, "This is the most powerful model, and if even Trump thinks so, it must be great."

Data analysis from Ramp also confirmed that after Anthropic clashed with the White House last time, Claude’s downloads significantly increased. Many initially thought ChatGPT was the AI assistant, but suddenly, Claude became the more responsible, more "resistant" choice.

Also impacting Taiwan

This AI game actually has an indirect connection to Taiwan. Local Taiwanese AI startups are also using Claude models for foundational training. When Anthropic shut down, Taiwanese developers could no longer access the latest model APIs. The export ban on Anthropic models also indirectly affects Taiwanese AI startups.

What’s more noteworthy is the simultaneity of these two events. Sean O’Kane pointed out that the White House shut down the models on Friday afternoon, almost simultaneously with the US-Iran nuclear negotiations. Bellan said, "They like to distract attention this way."

The "invisible war" of AI models

Anthony Ha warned that the risk to Anthropic isn’t very high. He believes, "No matter what risks exist, they tend to be exaggerated because they seem unable to communicate effectively."

He also mentioned that the "bad boy effect" of Claude is not uncommon among AI models. Both Sam Altman and Huang Renxun have said, "AI is actually very simple." But Anthropic also claims, "AI will replace all jobs," and then says, "We’ve just begun."

Behind this AI war, it’s really a contest over "who is more human-like."