Just now, a news article flooded the screen: The US government has directly imposed export controls on the two strongest models of Anthropic, Fable 5 and Mythos 5. I just tried using my Australian account and it really didn't work. In fact, I have tested the Fable 5 these days and it is quite useful. If you don't use it to run multi tasks, it is great to use it to assist in investment research. There is a reason why it is expensive. Why did it happen? Commerce Secretary Howard Lutnick sent a letter to Dario Amodei on Friday, listing Mythos 5 and Fable 5 as export controls, prohibiting access to any location outside the United States and all foreign visitors within the country. The trigger was another company claiming to have successfully jailbroken Mythos, alerting the government to national security risks.  The more crucial background is that the government had previously attempted to delay the release of these two models by Anthropic but was unsuccessful, so it directly used the export control card.  The specific concern mentioned in the official documents is that the model may bypass network security protection through specific prompts. Anthropic strongly disagrees, claiming that Fable's security barrier has undergone thousands of hours of Red Team testing and is stronger than any deployed model, and that no one has found a "universal jailbreak". A paradox worth pondering: Anthropic is currently on both the Pentagon's blacklist (considered too dangerous for government use) and the Commerce Department's licensing system (considered too dangerous for foreign use). Essentially, Washington is treating cutting-edge AI as a military grade strategic asset. What about foreign employees? This is the harshest part of this command. The ban explicitly covers "any foreigner, whether within or outside the United States, including foreign employees of Anthropic. That means Anthropic's own non American engineers cannot touch these two models. To ensure compliance, Anthropic has completely disabled Fable 5 and Mythos 5 for all customers, but all other Claude models are not affected. In terms of specific implementation, it belongs to the "one size fits all" approach - because it is impossible to accurately distinguish the nationality of users at the product level. So actually, now it's all customers, whether they're from the United States or not, don't even think about using it Americans themselves are also experiencing piss off. How long will it last? Short term, non permanent probability is high. A government official told Axios that the model needs to be locked in until the US government's national security system is "reinforced," and said this could happen in the coming weeks. Anthropic himself has set the tone as' misunderstanding 'and stated that he is working to restore access as soon as possible. So this is more like a regulatory pressure+negotiation, rather than a final judgment. However, even if restored in the future, the compliance framework may remain in place for a long time, as it is already being treated as a strategic asset What is the impact on IPO? The current timing is very delicate. Anthropic secretly submitted its S-1 on June 1st, with a valuation of approximately $965 billion, aiming to go public in October, led by Goldman Sachs and JPMorgan Chase.  My judgment is divided into two levels: Short term is a bearish noise, but the timing is critical. Anthropic just submitted its S-1 on June 1st, aiming to go public with a valuation of nearly trillions in October. This is the first sample used to test whether cutting-edge AI valuations can withstand the public market. The risk section of the prospectus now needs to add a new clause: the government can shut down your flagship product at any time. When underwriters price, regulatory discounts will only be thicker. But conversely, the stronger the ability, the more targeted the government will be - this news is to some extent endorsing Mythos' strength. If the ban is lifted within a few weeks and Anthropic packages it as a narrative of 'we are too advanced to be regulated', it may not necessarily be a net negative for long-term valuation. What we really need to focus on is whether this is a one-time event or another data point indicating the continued deterioration of the relationship between the government and Anthropic after the Pentagon blacklist. If it is the latter, 'policy risk' will become a structural valuation suppression factor relative to OpenAI. Let's see if this crisis has the potential to bring good opportunities to acquire PreIPO.