Indian AI startups confront new risks after Anthropic pulls Fable 5

Anthropic, a leading US-based AI developer, abruptly withdrew access to its advanced models, Fable 5 and Mythos 5, for foreign nationals on June 12, 2026, following a direct order from the US government citing national security concerns. This unprecedented move, the first known instance of AI models being subjected to export controls, has sent shockwaves through India's burgeoning AI ecosystem, exposing the deep vulnerabilities of Indian startups heavily reliant on foreign foundational models. This incident highlights a critical inflection point where global AI development is colliding with geopolitical realities, transforming AI models into strategic national assets. The US government's directive was reportedly instigated by concerns raised by Amazon CEO Andy Jassy to the White House regarding Fable 5 potential for misuse in cyberattacks, despite Anthropic disputing the severity of these risks. This follows earlier policy shifts by Anthropic in April 2026, which restricted third-party applications like OpenClaw from accessing its Claude models via consumer subscriptions, underscoring a broader industry trend of tightening control over powerful AI capabilities and API access. The immediate fallout has left Indian enterprises, including giants like Tata Consultancy Services (TCS) and Infosys, scrambling after losing access to systems they were integrating or building upon. The crisis has intensified urgent calls within India for accelerated development of 'Sovereign AI' capabilities, with industry leaders pushing for substantial funding, indigenous foundational models, and a pivot towards open-source alternatives to mitigate future geopolitical risks and ensure technological self-reliance. The abrupt withdrawal serves as a stark warning, forcing India to confront the imperative of building its own robust AI infrastructure rather than relying on the precarious goodwill of foreign providers.