AI Access Shifts Towards Geopolitics and Infrastructure

The selective distribution of exclusive access to advanced AI models like Anthropic's Claude Mythos Preview, shown by the exclusion of Indian entities from its Project Glasswing partnership, signals a significant change in how AI access is determined. This event is more than just a missed opportunity; it indicates that national control over core technological infrastructure—such as compute power, data centers, energy, and secure testing environments—is rapidly becoming the main factor in influence and access, overshadowing market size alone.

Anthropic's Cyber Power and India's Omission

Anthropic's Claude Mythos Preview has shown an unsettling ability to find and exploit zero-day vulnerabilities across major operating systems and browsers. Its capacity to autonomously generate functional exploits with minimal guidance marks a serious rise in cybersecurity threats. Project Glasswing was designed to use this capability for defensive cybersecurity and initially secured partnerships with major global infrastructure players, including AWS, Apple, Google, Microsoft, and NVIDIA. The obvious lack of Indian companies or public institutions from this foundational group, despite India being a substantial market for advanced AI, clearly shows the geopolitical realities affecting access. Jurisdiction, institutional trust, and national-security alignment are now paramount, making market entry a secondary consideration.

The Global Push for Sovereign AI

This incident reflects a growing global trend of 'Sovereign AI,' where countries prioritize controlling their own AI infrastructure and data. This allows them to rely less on foreign providers and gain geopolitical power. The United States, for example, is rapidly building national AI infrastructure, focusing on data centers, chip manufacturing, and grid modernization. Canada has launched a significant AI Compute Strategy, and European nations are working toward digital sovereignty through initiatives like Gaia-X. China continues to invest heavily in a complete AI system. In contrast, India's new efforts, though ambitious, are still taking shape. The IndiaAI Mission aims to make computing access more widespread and develop local capabilities, with a goal of over 10,000 GPUs. However, global demand and the concentration of computing power within a few major tech giants present a major challenge. While India has a large AI talent pool and leads in AI skill penetration, it also sees the world's largest net outflow of AI research talent, creating a contradiction of having capability but not retaining it. The new IndiaAI Safety Institute is tasked with encouraging local research and development, but faces the challenge of fitting with global security-first methods, which might make its job harder.

Strategic Dependency and Cyber Threats

India's current path risks cementing its position as just a user of AI, rather than a key player in its development. The exclusion from Project Glasswing is a sign of a deeper structural reliance. Nations that do not assert control over critical AI infrastructure—computing power, advanced chip design, secure data management, and a stable energy grid—will lose influence. Global AI computing power being held by a few tech giants leaves smaller nations exposed to geopolitical changes and supply chain issues. Furthermore, the growing threat posed by AI models capable of finding and exploiting vulnerabilities, as demonstrated by Mythos, worsens this risk. CERT-In's high-severity advisories to Indian MSMEs on AI-driven cyber threats, citing the potential for automated reconnaissance and multi-stage attacks, highlight the immediate national security risks of this infrastructure gap. The significant net outflow of AI talent, despite strong domestic skill penetration, suggests a failure to create attractive environments for advanced research and development in India, which further hinders the country's ability to build truly independent AI capabilities. Relying on imported, unexamined technologies reduces transparency and limits the ability to respond to crises, leaving India unprepared for future AI-driven security challenges.

Pathways for India's AI Future

To close this gap, India must accelerate its 'Sovereignty Moonshot Plan,' focusing on key technology areas—compute, energy, chips, cloud, data, and evaluation capacity—to develop significant local capability or ensure reliable control. This requires long-term, patient investment and strong coordination across government, industry, and academia. The IndiaAI Mission's commitment to scaling computing power, potentially to 50,000 GPUs through public-private partnerships, and fostering local AI innovation centers are important steps. However, addressing the talent outflow and establishing a truly secure and independent AI ecosystem will require tackling significant infrastructure and geopolitical challenges. The ongoing development of the IndiaAI Safety Institute, focused on local R&D and context-specific AI governance, is important. India's strategic future ultimately depends on shifting from being just a user of advanced AI to a nation that can shape its development, deployment, and governance through proven infrastructure sovereignty.