When Anthropic cut off access to its Fable 5 and Mythos 5 models for non-U.S. users last Friday, it didn't just rattle its Indian customer base. It exposed a structural dependency that India's AI ecosystem has been building for years. The directive, reportedly triggered by jailbreak vulnerabilities flagged by Amazon's CEO, forced Anthropic to block foreign nationals—including its own employees—from using the models. The timing was particularly awkward, coming right after Anthropic announced a partnership with Tata Consultancy Services to expand enterprise AI adoption in India.
For Indian startups and enterprises, this isn't an abstract geopolitical debate. It's an operational risk. Companies like Atomicwork, which splits engineering between Bengaluru and the U.S., now face a competitive disadvantage if their non-U.S. team members can't access frontier models. Aakrit Vaish of Activate described waking up "shocked and confused" by the news, and is now pushing portfolio companies to diversify away from a small set of U.S.-based model providers.
The debate has quickly shifted from "should we build our own models?" to "can we afford not to?" India's current AI strategy, the IndiaAI Mission, allocates roughly $1.2 billion over five years—a fraction of what training a single frontier model costs. Lightspeed's Hemant Mohapatra points out that capital isn't the binding constraint; talent, compute access, and execution matter more. But infrastructure costs for training are astronomical, and most Indian AI work remains application-layer, not foundational.
Sridhar Vembu of Zoho argues for embracing smaller, open-source models—both Indian and Chinese. Former Infosys executive Mohandas Pai is calling for a $5 billion annual AI fund plus a $21 billion credit guarantee program for cloud and hardware. Policy expert Prasanto Roy frames this as a supply chain problem: "American AI models are bound to American geopolitics." For India, the question is whether to accelerate domestic capability building or accept that access to frontier AI will remain contingent on U.S. foreign policy.
Source: TechCrunch