Asian AI companies are accelerating a new wave of frontier models, with Sakana AI in Tokyo and 360 in China unveiling systems designed to compete with leading global platforms. The timing reflects a broader shift in how regions are building AI capacity around local language, business needs, and strategic independence.
Sakana AI introduced Fugu, a model created to work with agent-based workflows and coordinate access to other models through APIs. The company says it can stand alongside top-tier systems while remaining optimized for Japanese language and cultural context. Sakana also positions the model as a practical option for organizations seeking flexible AI access in a changing global market.
In China, 360 presented Tulongfeng, an AI tool focused on cybersecurity tasks such as identifying software vulnerabilities. The company also highlighted a second system, Yitianzhen, built to support cyber defense and incident response. Together, the products show how AI is increasingly being tailored for specialized, high-value use cases.
The launches arrive as export restrictions reshape access to advanced AI models across Asia. Rather than signaling a full break from U.S. technology, Sakana AI describes the moment as a push for broader options and more resilient infrastructure. Its leadership argues that future AI systems may rely less on a single provider and more on coordinated model ecosystems.
That idea reflects a growing industry trend: orchestration over scale alone. Instead of building only larger models, developers are now focusing on systems that can manage multiple tools, models, and tasks efficiently. For businesses and public institutions, that could mean more adaptable AI services designed around local priorities.
As Asian startups refine models for regional markets, the next phase of AI may be shaped by accessibility, specialization, and collaboration. This evolution could help define a more distributed and multilingual future for intelligent systems worldwide.
