On Wednesday, Chinese cybersecurity firm 360 announced the release of Tulongfeng, an AI‑driven tool it claims can rival Anthropic’s Mythos in vulnerability detection. The same day, Tokyo‑based startup Sakana AI unveiled Fugu, a generative model built to compete with Anthropic’s Fable 5 and Mythos Preview. Both launches arrive amid a U.S. export ban that blocks non‑American entities from accessing Anthropic’s most powerful models.
Anthropic’s ban, issued two weeks ago, was intended to prevent the diffusion of advanced AI capabilities that Washington deems a national security risk. The order halted distribution of Mythos and its restricted variant, Fable 5, outside U.S. borders. Since then, Asian vendors have moved quickly to occupy the vacuum.
Sakana’s co‑founders—David Ha, Llion Jones and Ren Ito—said the timing was coincidental, but the company’s marketing now highlights “frontier capability without the risk of export controls.” Fugu, named after the Japanese word for blowfish, is optimized for the Japanese language and culture and is pitched to local businesses and government agencies looking to reduce reliance on U.S. models. The startup stresses that its model is not a permanent replacement for American AI but a hedge against sudden loss of access.
Ren Ito, a former executive at Mercari and Stability AI, told Project Syndicate that the United States should prioritize preserving AI access for its allies. He warned against hoarding the technology, arguing that collaborative development benefits all parties. David Ha, Sakana’s CEO, framed orchestration models—systems that coordinate multiple AI agents—as the next frontier beyond sheer size, noting that reliance on a single provider poses a strategic risk.
In Beijing, 360 introduced two security‑focused tools: Tulongfeng, which automatically discovers software vulnerabilities, and Yitianzhen, designed for automated cyber‑defense and incident response. Founder Zhou Hongyi called vulnerability‑finding AI a “national strategic asset” and warned of “one‑way transparency,” where only a few actors possess advanced detection capabilities.
The moves underscore a broader shift in the Asian AI landscape. Local firms are building models that understand regional languages and regulatory environments better than many U.S. offerings. While Anthropic’s revenue run‑rate topped $47 billion in May 2026, the company’s exposure to Asian enterprise customers remains unclear. Nonetheless, the export ban has opened a window for home‑grown alternatives to gain traction.
Neither Anthropic nor 360 responded to requests for comment at the time of writing. The developments highlight how geopolitical decisions can reshape the global AI market, prompting rapid innovation and competition beyond traditional U.S. and European hubs.
Este artículo fue escrito con la asistencia de IA.
News Factory APP - noticias agénticas para impulsar tu SEO y AEO.