Asian AI Startups Launch Mythos-Like Models Amid U.S. Export Ban Impact
Asian AI startups are launching advanced models to rival Anthropic's restricted systems, capitalizing on ongoing U.S. export bans.
As U.S. export restrictions enter their third week, Asian artificial intelligence startups are moving quickly to deploy advanced systems that compete directly with restricted American models. According to TechCrunch, Chinese cybersecurity firm 360 unveiled an AI tool named Tulongfeng on Wednesday, positioning it to go head-to-head against Anthropic’s Mythos model. The Trump administration currently bans the powerful cybersecurity-focused AI model, along with its more restricted counterpart Fable 5, from non-American users.
Earlier in the week, Tokyo-based Sakana AI introduced Fugu, a frontier model named after the Japanese word for blowfish. According to the company, the new product stands shoulder-to-shoulder with leading systems like Anthropic’s Fable 5 and Mythos Preview. Designed specifically for agents, Fugu features the ability to orchestrate access to other models through their APIs, offering frontier capabilities intended to sidestep the risks associated with tightening export controls.
Sakana AI’s leadership noted that the release timing coincided with the export ban, though the underlying research was presented earlier this spring at ICLR and built upon work initiated last year. Co-founded in 2023 by David Ha, Llion Jones, and Ren Ito, the Tokyo startup focuses on affordable generative models optimized for small datasets, Japanese language, and local culture. While targeting Japanese businesses and government agencies seeking to reduce exposure to export limits, the company emphasizes that U.S. models remain important to the region rather than signaling a permanent market realignment.
Beyond regional adaptation, Sakana co-founder David Ha framed Fugu as a strategic hedge against centralized power. Writing on X, Ha argued that relying on a single provider for national infrastructure introduces severe risks when access can disappear overnight, positioning orchestration models as the next frontier beyond larger standalone systems. Meanwhile, China’s 360 took a different approach with its product rollout, introducing tools aimed at automating software vulnerability discovery and incident response.
Founder Zhou Hongyi described vulnerability-finding artificial intelligence as a national strategic asset, raising concerns over what he termed one-way transparency, where select actors retain advanced capabilities while others do not. While Anthropic has experienced substantial growth, crossing a run-rate revenue of $47 billion in May 2026, the full impact of the export restrictions on its Asian enterprise customer base remains unclear. As the export order drags on, local alternatives are rapidly stepping into the market space left behind by American labs.
Sources
- 01 Asian AI startups launch Mythos-like models as Anthropic’s export ban drags on — TechCrunch — Startups