Asian AI startups launch Mythos-like models
Sakana AI and 360 launched new advanced AI models as access to some US frontier systems remains restricted.
Model access is becoming a regional strategy issue. Asian AI companies are launching new systems as customers look for alternatives to restricted US frontier models.
What happened
Sakana AI in Tokyo launched Fugu, while Chinese cybersecurity firm 360 unveiled Tulongfeng and Yitianzhen.
The launches are being positioned as Asian alternatives to advanced US AI models while access restrictions around some frontier systems remain unresolved.
Why it matters
This is a strong AI sovereignty and model-access signal.
If access to US frontier models becomes less predictable, regional AI companies may gain momentum by offering local alternatives tuned for domestic customers, languages and policy environments.
The bigger picture
The AI model race is becoming more geopolitically fragmented. Startups and incumbents outside the US may benefit if enterprises and governments want capable models that are not dependent on restricted foreign access.