Amid Claude Mythos and Fable 5 pause, Infosys co-founder backs argument on why India hasn't built a ChatGPT

Infosys co-founder Kris Gopalakrishnan backed an X post arguing that Indian IT firms are not built to compete in developing frontier AI models like OpenAI’s ChatGPT. The discussion gained traction after Anthropic restricted access to its advanced models, Fable 5 and Mythos 5, for foreign users due to U.S. export controls, raising questions about India’s AI ambitions. The post, by an X user named Piramal, claimed Indian IT giants—TCS, Infosys, Wipro, and HCL—have not failed in generative AI but are fundamentally different from Silicon Valley AI startups. Gopalakrishnan shared the post, writing, ‘Thanks for the right perspective.’ Piramal outlined four reasons why Indian IT firms cannot replicate OpenAI’s approach. First, frontier AI requires massive infrastructure investments, often backed by governments or tech giants like Microsoft and Amazon. Indian firms, constrained by shareholder expectations, cannot risk large, uncertain expenditures—such as buying Nvidia H100 chips—without triggering stock crashes. Second, Indian IT plays a critical economic role, generating over $200 billion annually in foreign currency through service exports. These inflows stabilize the rupee, bolster foreign exchange reserves, and support India’s economic resilience. Piramal argued that prioritizing speculative AI over proven revenue streams could destabilize this model. The debate reflects broader tensions: whether India should focus on homegrown AI models or leverage its strengths in enterprise technology and services. While Indian firms excel in steady cash flow and employment, their corporate structures limit high-risk AI investments, leaving frontier AI development to better-funded global players.