June 28, 2026, (Inside AI) — Asian AI startups are rapidly filling the void created by a U.S. export ban that restricts non-Americans from accessing Anthropic's advanced models, including Mythos and Fable 5. Two new models, Tulongfeng from China's 360 and Fugu from Japan's Sakana AI, promise comparable capabilities without geopolitical restrictions.
The ban, now in its third week under the Trump administration, blocks export of the cybersecurity-focused Mythos and the agent-oriented Fable 5. In response, Chinese firm 360 unveiled Tulongfeng on Wednesday, claiming it matches Mythos in cybersecurity tasks. Tokyo-based Sakana AI launched Fugu earlier this week, asserting it rivals Fable 5 and Mythos Preview in agent orchestration.
Zhou Hongyi, founder of 360, framed the launch as a strategic necessity. He said:
"This kind of powerful weapon that can change the landscape of cyber offence and defence cannot be held only by others."
Sakana AI described Fugu's timing as coincidental, noting the research predates the ban. Its website highlights frontier capability free from export controls. The model can coordinate other AI systems via APIs, a direct challenge to Fable 5's agentic features.
Anthropic's revenue hit a $47 billion run-rate in May 2026, but its reliance on Asian enterprise customers remains undisclosed. Regional competitors now offer models tuned for local languages and cultural nuances, potentially eroding Anthropic's market share permanently.
The ban's unintended consequence is accelerated regional self-sufficiency. As trust shifts, even a future policy reversal may not restore lost ground. These developments underscore how export controls can reshape global AI competition overnight.
