The landscape of global artificial intelligence has undergone a seismic shift this week as prominent technology firms in China and Japan unveiled advanced AI models designed to fill the vacuum left by tightening United States export controls. On Wednesday, the Chinese cybersecurity titan 360 officially introduced Tulongfeng, a sophisticated AI tool specifically engineered to compete with Anthropic’s Mythos, a high-performance model currently restricted by the Trump administration. Simultaneously, Tokyo-based startup Sakana AI announced the release of Fugu, a frontier model that the company claims matches the capabilities of Anthropic’s Fable 5 and Mythos Preview. These developments signal a rapid regional response to the U.S. government’s decision two weeks ago to ban non-Americans from accessing Anthropic’s most powerful cybersecurity-focused AI tools, sparking a race for AI sovereignty across Asia.
The Rise of Tulongfeng and China’s Strategic AI Defensive
The unveiling of Tulongfeng by 360 marks a significant milestone in China’s efforts to achieve parity with American AI capabilities, particularly in the critical domain of cybersecurity. According to reports from the Chinese firm, Tulongfeng is not merely a general-purpose large language model but a specialized tool designed to automatically identify and analyze software vulnerabilities at a scale and speed previously unattainable. This launch was accompanied by the introduction of Yitianzhen, a secondary tool built to automate cyber defense mechanisms and streamline incident response for enterprise and government networks.
Zhou Hongyi, the founder of 360 and a prominent figure in the Chinese tech industry, framed the release of these tools as a matter of national strategic importance. Speaking at the launch event, Zhou emphasized that vulnerability-detection AI has become a "national strategic asset." He specifically warned against the dangers of "one-way transparency," a geopolitical scenario where certain nations possess advanced AI-driven offensive and defensive capabilities while denying them to others. By launching Tulongfeng, 360 aims to ensure that Chinese infrastructure remains resilient in an era where AI-driven cyber warfare is becoming a tangible reality.
The timing of the announcement is seen by industry analysts as a direct rebuttal to the U.S. export ban. While Anthropic’s Mythos is touted as a revolutionary tool for securing code, its unavailability in the Chinese market has provided a clear opening for 360 to position its own suite of products as the primary alternative for domestic industries.
Sakana AI and the Fugu Model: Seeking Resilience Through Orchestration
In Japan, the response to the shifting AI landscape has taken a different but equally strategic form. Sakana AI, a startup that has rapidly gained international attention since its founding in 2023, launched Fugu—named after the Japanese blowfish. Fugu is marketed as a frontier AI model that "stands shoulder-to-shoulder" with the now-restricted Fable 5 and Mythos models from Anthropic. However, unlike traditional monolithic models, Fugu is designed with a specific focus on "orchestration."
David Ha, co-founder and CEO of Sakana AI and a former Google researcher, described Fugu as a tool for the "next frontier" of AI development. The model is built to act as an agent that can coordinate and manage access to various other AI models through their respective APIs. This architectural choice is a deliberate hedge against the volatility of international tech policy. Ha argued on social media that relying on a single, foreign provider for national AI infrastructure is a systemic risk that the recent U.S. export controls have made impossible to ignore. "Access to top models can disappear overnight," Ha noted, advocating for "collective intelligence" as a practical defense against the concentration of technological power.
Sakana AI’s spokesperson clarified that while the timing of Fugu’s release coincided with the Anthropic ban, the research behind the model had been in development since last year and was presented at the International Conference on Learning Representations (ICLR) this spring. Despite this, the company’s marketing has leaned into the current geopolitical climate, advertising Fugu as a way to deliver "frontier capability without the risk of export controls."
A Chronology of the Anthropic Export Ban and its Global Fallout
The current tensions in the AI market can be traced back to a series of escalating policy decisions by the U.S. executive branch. To understand the significance of the Tulongfeng and Fugu launches, it is necessary to examine the timeline of events leading to the current market fragmentation:
- May 2026: Anthropic, a leading U.S. AI lab, reports a historic growth trajectory, with its run-rate revenue crossing $47 billion. The company nears a $1 trillion valuation ahead of a highly anticipated initial public offering (IPO).
- Early June 2026: The Trump administration issues a series of executive orders aimed at protecting American technological intellectual property. Citing national security concerns regarding the potential for AI to be used in autonomous cyberattacks, the administration places strict export controls on "frontier cybersecurity models."
- Two Weeks Ago: The U.S. government specifically orders Anthropic to cease providing global access to its Mythos and Fable 5 models. These models are restricted to U.S. citizens and domestic entities, effectively cutting off enterprise customers in Europe and Asia.
- Last Week: Leaders gather at the G7 summit in Evian, France. AI access and export controls emerge as a central topic of debate. Ren Ito, co-founder of Sakana AI, attends the summit and advocates for a collaborative approach to AI development, warning that hoarding technology could alienate key allies.
- This Week: Sakana AI launches Fugu in Tokyo, followed by 360’s launch of Tulongfeng in Beijing, both positioning their tools as direct competitors or replacements for Anthropic’s restricted software.
The Economic Impact on Anthropic and the U.S. AI Sector
The export ban on Mythos and Fable 5 represents a significant challenge for Anthropic, which had been positioning itself as the primary provider of high-security AI for the global enterprise market. While the company’s $47 billion run-rate revenue reflects its dominance in the U.S., the loss of access to Asian markets—particularly in tech-heavy hubs like Tokyo, Seoul, and Singapore—could dampen its long-term growth prospects.
Market analysts suggest that the "America First" approach to AI regulation may have unintended consequences. By restricting access to its most advanced tools, the U.S. is inadvertently subsidizing the growth of competitors in Asia. Companies like Sakana AI, which was founded by alumni from Google and Stability AI, are already highly capitalized, recently raising $135 million in a Series B round at a $2.65 billion valuation. With the added incentive of providing "sovereign AI" solutions, these firms are now seeing an influx of interest from government agencies and large corporations looking to de-risk their operations from U.S. policy shifts.
Analysis of Implications: The Shift Toward Sovereign AI
The emergence of Tulongfeng and Fugu highlights a broader trend toward "AI Sovereignty." This concept refers to the ability of a nation or region to develop, maintain, and control its own AI infrastructure without being dependent on foreign entities that may change their terms of service or access based on political whims.
For Japan, Sakana AI represents a middle ground. The company is optimized for Japanese language and culture, filling a niche that U.S. models often overlook. However, the company remains vocal about the importance of U.S. models to the Asian ecosystem. Ren Ito, in an op-ed for Project Syndicate, urged the U.S. federal government to prioritize maintaining access for its closest allies. He argued that AI should be a shared technology rather than a hoarded resource, suggesting that the current moment should be viewed as a temporary friction rather than a permanent realignment.
In contrast, the Chinese approach, exemplified by 360, appears more focused on total independence. By framing AI as a "national strategic asset," Beijing is signaling that it views the technology through the lens of traditional defense and security. The risk of "one-way transparency" mentioned by Zhou Hongyi suggests that China will continue to pour resources into domestic alternatives to ensure that its cybersecurity capabilities remain on par with the West, regardless of U.S. trade policies.
Looking Ahead: A Fragmented Global AI Market
The long-term impact of these launches will depend on whether Tulongfeng and Fugu can truly match the performance of Anthropic’s offerings in real-world applications. If these Asian models prove to be effective, it may become difficult for U.S. firms to regain their market share even if export controls are eventually relaxed. Enterprise customers value stability and reliability; once a company integrates a local AI solution like Fugu into its infrastructure to avoid political risk, the "switching costs" to return to a U.S. provider may be prohibitively high.
As the AI race continues, the focus is shifting from who has the largest model to who can provide the most reliable access. For now, the vacuum created by the U.S. ban on Mythos and Fable 5 has been filled by a new generation of Asian technology, signaling that the era of undisputed U.S. AI hegemony may be drawing to a close. The global community now watches to see if other regions, such as the European Union, will follow suit in developing their own "frontier" alternatives to safeguard their technological futures.
