Asian AI Startups Challenge U.S. Dominance with Sovereign Models Amid Export Restrictions

June 27, 2026
Asian AI Startups Challenge U.S. Dominance with Sovereign Models Amid Export Restrictions
  • Asian AI startups Sakana AI and 360 are rolling out advanced models in response to U.S. export restrictions on Anthropic tools, signaling a shift toward sovereign AI and reduced dependence on U.S. providers.

  • Sakana AI launches Fugu, an orchestrator model designed to coordinate multiple AI systems and APIs for complex agentic tasks, positioning itself as a regional competitor to U.S. frontier models.

  • 360 debuts Tulongfeng as a direct alternative to restricted U.S. tools, with its partner Yitian Tulong focusing on cyber defense automation, shortly after export bans on Mythos and Fable 5.

  • Investors will watch how these alternatives perform in practice—accuracy, speed, safety—and how IT services firms adapt to become model-agnostic and help clients navigate regulatory and supply-chain risks.

  • Global businesses are adopting a multi-model strategy to mitigate access risk, testing models from both U.S. and local providers to preserve operational resilience.

  • Industry voices argue for orchestration and collective intelligence as hedges against single-provider national infrastructure dependence.

  • The move reflects a strategic pivot in AI, with regional competition and policy changes driving diversification of toolchains and potential shifts in the global AI ecosystem.

  • Sovereign AI aims to run localized models that can operate independently and address vendor lock-in, regulatory risk, and local language and cultural nuances.

  • Analysts note non-U.S. players entering frontier AI to fill gaps left by export controls, potentially reshaping global AI ecosystem dynamics and spawning new regional AI hubs.

  • 360’s founder notes China’s models lag behind the U.S. by 20–30% in base capability but argues for proactive development, highlighting Tulongfeng’s reported vulnerability discoveries amid ongoing verification concerns.

  • Sakana AI, founded by former Google researchers, emphasizes a seven-billion-parameter orchestrator that routes tasks across models, aiming for frontier capability with export-control risk mitigation.

  • Sakana frames Fugu as collaboration-friendly rather than a replacement for U.S. models, underscoring that U.S. tools remain important to Asia and that any realignment is a hedge, not a permanent shift.

Summary based on 4 sources


Get a daily email with more Startups stories

More Stories