Synopsis
US restrictions on foreign access to Anthropic's advanced AI models, Mythos and Fable 5, are creating a significant competitive disadvantage for Indian enterprises and IT service providers. This exclusion, particularly from Fable 5's coding capabilities, raises concerns about digital equity and strategic dependency, as Indian data contributed to the models' value.Listen to this article in summarized format
Mythos had a three-month head start before a handful of Indian companies were admitted to Project Glasswing, a cybersecurity initiative under which the company provided vetted organisations access to the model. Now, exclusion from Fable 5’s advanced coding capabilities could place India’s tech service providers at a disadvantage to American rivals.
“India’s IT services industry was built on the assumption of open access to the best global technology. That assumption has now developed a serious crack. This creates a competitive disadvantage, but the deeper risk is strategic dependency,” said Nikhil Narendran, partner at law firm Trilegal and a tech policy thought leader.
It poses a serious question of digital equity, according to him. “Frontier models are trained on the world’s data including India’s. A significant part of that data comes from countries such as India, including our text, code, languages, users and markets. Indian users and enterprises have helped make these systems commercially valuable. There is something deeply unfair about one country benefiting from such capabilities,” he said.
From a cyber perspective, too, the suspension could delay patch-fixing.
“This sudden suspension of Fable 5 and Mythos 5 changes the playbook for the cyber preparedness of Indian enterprises,” said Neehar Pathare, chief executive at cybersecurity firm 63SATS Cybertech. “It forces an immediate shift from a strategy of technological consumption to one of operational self-reliance. Real cyber preparedness means owning the stack.”
ET reported in May that essential services providers such as banks, telecom operators and power utilities had started stress-testing their codebases to prepare their defences against advanced cyber threats. Companies are testing their public-facing assets using existing AI models such as Opus 4.7 and GPT 5.5 and have also asked their suppliers to identify possible gaps.
Significantly, the US government has suspended access to any foreign national, whether inside or outside the country, including Anthropic employees. The clause has sparked confusion over talent mobility at firms such as Tata Consultancy Services, Infosys and even multinationals like Deloitte and Goldman Sachs, which recruit both American and foreign talent.
“Nation-states will soon start needing citizenship and/or security clearances to work on the next SOTA (state-of-the-art) models the way they do for defence, space, nuclear tech,” Hemant Mohapatra, partner, Lightspeed, wrote in a post on X.
In fact, netizens noted that even Andrej Karpathy, the founding member of OpenAI and currently a member of Anthropic’s pretraining team may not have access to the models. The Slovak-Canadian immigrant currently holds an EB-1 visa status.
“This is probably the first time that ethnicities have become part of geopolitical strategy, in a way that determines who can access critical technologies, and that could have far-reaching implications for R&D, hiring, funding and global collaboration,” said Abishur Prakash, an author and geopolitical strategist at Canada-based advisory firm The Geopolitical Business Inc. “To me, the bigger question is how this trickles into more division within the global economy and society.”
However, the advantage may be short-lived. Open-source models are expected to reach Mythos-level capabilities within three to five months and hence, enterprises can still continue ongoing experiments, experts said.
A recent joint research by the Massachusetts Institute of Technology and Hugging Face revealed that Chinese open-source AI models such as DeepSeek R1, Alibaba’s Qwen and Moonshot’s Kimi AI accounted for 17.1% of global downloads in 2025, surpassing the US share of 15.8%.
“The performance gap between frontier closed models and open models has been compressing and the cost factor makes Chinese models a practical choice for startups,” said Abhishek Srivastava, general partner at venture capital firm Kae Capital.
But whether firms can trust Chinese open source models remains an open question.
“From a technical trust perspective, these models are reasonably secure. As far as geopolitical or commercial tests are concerned, I don’t think it’s US versus China. So, the question is: Can we become more independent?” Srivastava said.
If India merely responds to American dependency by rushing into Chinese dependency, it won’t achieve sovereignty, according to Trilegal’s Narendran. “We have merely changed our landlord. We need enough domestic capability that trust becomes a choice, not a compulsion,” he said.
Some experts believe such constraints can become catalysts for innovation.
“When restrictions were imposed on IBM mainframes decades ago, India did not fall behind,” said Vineet Nayyar, former CEO of HCL Technologies. “We pivoted to open systems and built capabilities that ultimately helped create a world-class IT industry,” he said while recalling the formative years of HCL and the broader Indian IT industry's emergence.
“We celebrate diversity in energy sourcing today thanks to the Iran war because dependence on a single source created vulnerability. The same principle applies to AI. A more diverse AI ecosystem will be healthier for innovation, competition and resilience in the long run,” Nayyar said.
