On Wednesday, Chinese cybersecurity company 360 introduced Tulongfeng, an AI solution it claims can compete directly with Anthropic's Mythos. This cybersecurity-oriented model is described as highly capable, prompting the Trump Administration to bar its distribution along with the restricted Fable 5 variant to users outside the United States.

Earlier that week, Tokyo-based Sakana AI released Fugu, a frontier model named after the Japanese term for blowfish. The firm positions it as comparable to leading systems such as Anthropic's Fable 5 and Mythos Preview, with features for agent orchestration via external APIs.

These Asian releases arrive while the U.S. government restriction on Anthropic remains active. The order limiting global access to Mythos and Fable took effect two weeks earlier.

A Sakana AI spokesperson described the timing of the launch as purely coincidental, though the company has highlighted the opportunity. Its site promotes frontier performance free from export control concerns.

The spokesperson noted that development of Sakana Fugu began last year, with related research shared at ICLR this spring. They emphasized confidence in the model's independent strengths, adding that the current situation simply increased visibility beyond initial expectations.

Sakana, established in 2023 by former Google researchers Ren Ito, Llion Jones and David Ha, focuses on cost-effective generative models suited to smaller datasets and optimized for Japanese language and culture.

While targeting Fugu at Japanese firms and agencies seeking to limit exposure to export rules, the company stops short of declaring a permanent move away from U.S. AI offerings in the region.

U.S. models continue to hold value in Asia, the spokesperson stated, aligning with comments from co-founder Ren Ito at last week's G7 summit in Evian. The current situation reflects temporary conditions rather than a fixed shift toward any specific group of providers.

Ren Ito expanded on this perspective in a Project Syndicate op-ed last week, urging U.S. officials to prioritize continued access for close allies and stressing that AI should be developed collaboratively rather than restricted.

Co-founder and CEO David Ha framed Fugu as more than an opportunistic response to U.S. competitors' challenges. The model is built to manage agent interactions across multiple systems.

Orchestration models represent the next stage beyond simply scaling size, he posted on X. Dependence on one provider for critical infrastructure carries risks highlighted by recent export measures.

Access to leading models can vanish suddenly, he added, making collective intelligence a practical safeguard against concentrated control.

While Sakana presents Fugu as a way to maintain frontier AI access rather than displace it, China's 360 took a more direct stance.

The firm launched two security-focused AI products. Tulongfeng automates software vulnerability detection, and Yitianzhen handles cyber defense and incident response tasks.

The announcement carried a pointed message. Reuters reported that 360 founder Zhou Hongyi called vulnerability-detection AI a strategic national resource and warned of risks from one-way transparency in advanced capabilities.

Anthropic had been expanding rapidly, reporting a run-rate revenue exceeding $47 billion in May 2026. The share tied to Asian enterprise clients remains undisclosed.

In the weeks since the export order, companies in Tokyo and Beijing have moved to fill the resulting gap. Even if U.S. firms regain confidence later, locally trained alternatives attuned to regional language and context are already taking hold.

360 did not reply to a request for comment.

When you purchase through links in our articles, we may earn a small commission. This does not affect our editorial independence.

Kate Park is a reporter at TechCrunch, with a focus on technology, startups and venture capital in Asia. She previously was a financial journalist at Mergermarket covering M&A, private equity and venture capital.