Who Benefits from the Trump Administration’s Ban on Anthropic’s Fable 5 and Mythos 5?

Export‑control ban

On 13 June 2026 the U.S. Treasury issued an export‑control order that bars foreign governments, companies, individuals and U.S. persons abroad from accessing Anthropic’s Fable 5 and Mythos 5 models. Anthropic received a phone call that it had 90 minutes to remove the models from service; the order cited “national‑security threats” but gave no technical justification.

Amazon’s report and alleged jailbreak

According to Axios, Amazon submitted a research report claiming its engineers used a sequence of prompt‑engineering instructions to “jail‑break” Fable 5, bypassing safety filters and extracting information that could be used to facilitate cyber‑attacks. The Wall Street Journal reported that Amazon CEO Andy Jassy discussed the matter with Treasury Secretary Scott Bessent and other senior officials. Axios also noted that at least five additional companies contacted government officials with similar concerns on the same night.

Anthropic’s response

Anthropic’s blog states that the reported bypass only exposed minor vulnerabilities that are also present in other publicly available models, and that the security controls of Fable 5 were not fundamentally broken.

Legal status of Andrej Karpathy

Karpathy holds an EB‑1 “extraordinary ability” green card; under U.S. export‑control law he is classified as a U.S. person, so the ban likely does not restrict his access to the models despite his Slovenian origin.

Comments from industry figures

Yann LeCun is quoted as blaming Dario Amodei for creating panic that led to the ban. White House adviser David Sacks posted that the export control was a “last‑resort” measure after Anthropic’s CEO Dario Amodei refused to fix the jailbreak or withdraw the models, and that the ban will be lifted once the issue is resolved.

Speculated motives

Analysts suggest Amazon may benefit competitively because restricting Anthropic’s models reduces pressure on Amazon’s own lagging AI offerings.

Another view is that Amazon, as a cloud provider, could face liability if a model it hosts is deemed a national‑security risk, so early reporting to the government serves as a self‑protection strategy.

Potential market impact

Some observers argue that OpenAI could be the biggest beneficiary because the restriction eases competitive pressure on its models ahead of the planned IPOs of both Anthropic and OpenAI.

References

Axios: https://www.axios.com/2026/06/13/anthropic-amazon-white-house

Wall Street Journal: https://www.wsj.com/tech/ai/amazon-ceos-talks-with-u-s-officials-triggered-crackdown-on-anthropic-models-dcc90578

Reuters: https://www.reuters.com/business/retail-consumer/amazon-voiced-concerns-about-anthropic-ai-models-before-us-governments-crackdown-2026-06-13/

Anthropic blog: https://www.anthropic.com/news/fable-mythos-access