How Japan's Sakana AI and China's 360 Respond to US AI Export Restrictions: A Turning Point for Asian Technological Autonomy

US Export Controls Trigger Asian AI Alternative Wave

In June 2025, the US government, citing national security concerns, banned the export of Anthropic's most powerful AI models, Mythos and Fable 5. This decision quickly sparked responses from Asian tech giants. Chinese cybersecurity company 360 launched an AI tool called "Tulongfeng," claiming its capabilities rival Mythos. Meanwhile, Japanese AI startup Sakana AI released a new model, "Fugu," explicitly marketing it with the selling point of "advanced capabilities free from export controls."

Sakana AI: A Japanese Native Model Built by Researchers Who Left Google

Sakana AI was founded in 2023 by former Google researchers Ren Ito, Llion Jones, and David Ha. The company focuses on developing models that efficiently utilize small datasets and are deeply adapted to the Japanese language and culture. The Fugu model not only claims to compete with Fable 5 but also can coordinate other models through various API interfaces, suitable for complex agent systems. Although Sakana AI states that the release timing coinciding with the US ban is purely "coincidental," its website has made "advanced capabilities not subject to export controls" a core promotional message, directly targeting Asian enterprises and government clients sensitive to the politicization of technology.

Quick Follow-up by China's 360

360, a leading Chinese cybersecurity company, has launched the Tulongfeng model to directly compete with Anthropic's Mythos. Although specific technical details have not been disclosed, this move signals China's strategic intent to accelerate "de-Americanization" in the AI field. 360's entry further strengthens the trend of Asia forming an independent technological pole.

The Complex Game of Technological Autonomy and Geopolitics

Sakana AI co-founder Ren Ito stated at the G7 summit that US models remain important for Asia, but Washington must maintain a technology pathway for its allies. This statement reveals the real dilemma for Asia: on one hand, the US remains a leader in fundamental AI research; on the other hand, export controls are forcing regional companies to seek alternatives, inadvertently accelerating localized innovation.

Experts point out that export restrictions may weaken the global market share of US AI companies in the long run. Although alternative models from Japan and China started later, they are developing differentiated competitive advantages by optimizing for local needs (such as Japanese language support, cultural adaptation) and avoiding political risks.

Implications for the Global AI Ecosystem

The US ban on the export of its most powerful AI models has effectively created a "protected" market space for the Asian AI ecosystem. The rapid responses from Sakana AI and 360 indicate that technological sovereignty has become central to the digital strategies of various countries. In the future, global AI may fragment into multiple competing and partially decoupled ecosystems: a US-led one, a China-led one, and a middle ground sought by allies like Japan.For Japan, the success of Sakana AI is not just a breakthrough for a startup, but also represents the possibility of Japan's transformation from a "follower" to a "differentiated innovator" in the AI field. The advantages of its models in efficiency, language adaptation, and compliance may earn Japan a unique position in the Asian and even global markets.

--- *This article is based on TechCrunch reports and public information analysis, and does not represent an objective evaluation of any model.*