Two weeks after the United States government ordered Anthropic to block foreign access to its frontier Mythos and Fable 5 AI models, Asian technology companies are already filling the gap with competing alternatives. A Chinese cybersecurity firm and a Japanese AI startup have launched models that explicitly rival Anthropic's banned offerings, signalling a potential long-term restructuring of the global AI market.

The export ban, effective from June 12, 2026, required Anthropic to suspend access to its most advanced models by foreign nationals, including the company's own foreign-born employees. Unable to selectively enforce such a broad restriction, Anthropic disabled both Mythos and Fable 5 for all customers globally. The move has sent shockwaves through Asian technology markets that had come to depend on frontier American AI capabilities.

AI technology concept with server racks and data centre

Sakana AI's Fugu: The Orchestration Approach

Tokyo-based Sakana AI, co-founded in 2023 by former Google researchers including Llion Jones — a co-author of the seminal Transformer paper — and David Ha, has released Fugu, named after the Japanese word for blowfish. Rather than training a monolithic frontier model, Fugu is a 7-billion-parameter orchestration model that routes tasks across a pool of available external models, coordinating them as a team. Sakana claims the result matches the performance of Anthropic's Fable 5 on key benchmarks while costing a fraction to train and operate.

Sakana's approach is fundamentally different from the conventional AI arms race: instead of building bigger models, Fugu functions as a "collective intelligence" layer that decides which model should handle each subtask. The company completed a $135 million Series B round in November 2025 at a valuation of approximately $3 billion. Its research behind Fugu was presented at the ICLR conference in spring 2026, well before the export ban came into effect.

"Sakana Fugu is something we have been building since last year," a company spokesperson told TechCrunch. "The timing simply happened to coincide with a moment that brought it more attention than we expected." The company's website now prominently advertises "delivering frontier capability without the risk of export controls." Co-founder Ren Ito argued in a Project Syndicate op-ed that AI should not become a technology that is hoarded, urging the US government to preserve access for its closest allies.

360 Security's Tulongfeng: A National Strategic Asset

China's 360 Security took a more assertive stance. At the ISC AI 2026 cybersecurity conference in Beijing on June 24, founder Zhou Hongyi unveiled two AI tools collectively branded as Yitian Tulong. Tulongfeng is designed for automated software vulnerability discovery — directly targeting the same capability that made Anthropic's Mythos model notable. Yitianzhen automates cyber defence and incident response.

Zhou described vulnerability-finding AI as a "national strategic asset" and warned of what he termed "one-way transparency" — a scenario where some nations possess advanced capability to probe software for security flaws while others cannot, creating a dangerous strategic imbalance. 360 claims its tools have already identified 3,432 vulnerabilities, with 105 confirmed by Chinese authorities, though Reuters could not independently verify these figures.

Zhou conceded that Chinese AI models still lag behind American frontier models by an estimated 20 to 30 percent in baseline capability. However, he argued that waiting to achieve parity before deploying defensive AI tools is not a viable strategic option, especially when vulnerability detection capabilities have direct national security implications.

Why This Matters for India

India has emerged as one of the most important battlegrounds in this unfolding geopolitical AI dynamic. Both Anthropic and OpenAI have described India as their second-largest market after the United States. The US export ban has reignited a long-running debate about whether one of the world's largest AI markets can afford to depend on technologies built and controlled elsewhere.

India's response has been multi-pronged. The IndiaAI Mission, approved in 2024 with an outlay of ₹10,372 crore ($1.2 billion), aims to expand domestic compute infrastructure and support indigenous AI development. Homegrown AI labs like Sarvam AI have released open-source foundation models, while companies like HCLTech have invested in Indian AI startups. The broader debate about US restrictions on advanced AI and what they mean for India has gained renewed urgency as Asian alternatives emerge.

The emergence of Sakana's Fugu and 360's Tulongfeng provides Indian enterprises and government agencies with additional options to hedge against US export control risks. The hedge strategy — maintaining access to multiple model providers rather than depending on a single source — aligns with growing calls within India's technology policy community for diversifying AI supply chains.

Market Impact and Long-Term Outlook

Anthropic had reached a $47 billion revenue run-rate by May 2026, with a significant portion derived from Asian enterprise customers. The company recently filed a confidential IPO prospectus with the SEC, disclosing a valuation of $965 billion. However, the export ban has introduced an unprecedented level of geopolitical risk into Anthropic's business model.

Neither Sakana nor 360 claims to fully replace American frontier AI. Sakana frames Fugu as a hedge — a way to preserve capability even if access to a single provider disappears overnight. 360 positions its tools as a national strategic necessity. But the broader signal is clear: within two weeks of the ban, concrete competitors have emerged in both a close US ally (Japan) and a strategic rival (China). Even if the ban is eventually lifted, the trust dynamic has shifted permanently. Asian enterprises now understand that frontier AI access can be revoked by foreign policy decisions made without their input or consent.

The long-term trajectory points toward a more fragmented global AI landscape where sovereign capabilities coexist with American frontier models. This fragmentation carries both risks — including reduced economies of scale and slower overall progress — and opportunities, particularly for markets like India that can leverage multiple AI ecosystems.

Frequently Asked Questions

What exactly did the US export ban on Anthropic do?

The US government ordered Anthropic to block access to its Mythos and Fable 5 models by any foreign national, including the company's own foreign-born employees. Unable to implement such broad restrictions selectively, Anthropic disabled both models globally.

How does Sakana AI's Fugu work differently from conventional AI models?

Fugu is an orchestration model — it coordinates access to multiple external models via their APIs, deciding which model handles each part of a problem. This approach allows it to match frontier model performance without training a single massive model from scratch.

What does this mean for Indian AI startups and enterprises?

Indian businesses that depend on Anthropic's frontier models face restricted access, potentially putting them at a competitive disadvantage relative to US-based teams. The emergence of Asian alternatives provides hedging options, but the incident has accelerated India's push toward AI self-reliance through initiatives like the IndiaAI Mission and homegrown foundation models.

Will the export ban be lifted?

The White House has signalled it is unlikely to extend similar restrictions to other AI companies and has privately attributed the directive to specific concerns about Anthropic's models. However, no timeline for lifting the ban has been announced, and the geopolitical dynamics surrounding frontier AI export controls continue to evolve.

Sources