Artificial intelligence occasionally hallucinates. It produces responses that are coherent, persuasive, and entirely detached from reality. A more consequential hallucination is now taking hold beyond machines. Governments, markets, investors and increasingly entire societies are beginning to construct narratives about AI that display a similar confidence while remaining equally detached from the institutional realities that determine technological leadership.Every technological revolution produces its own mythology. The steam engine promised limitless prosperity. The internet promised the end of geography. Social media promised stronger democracies. AI has now produced its own mythology about intelligence and what it means to create it.AdvertisementIf frontier AI could be bought with a US$10 billion cheque, several Indian conglomerates could have reached the frontier years ago. The inconvenient reality is that money can finance an ecosystem, but it cannot compress decades of scientific inquiry, institutional evolution and industrial capability into a single budget cycle. In AI, it borders delusion.There is another uncomfortable consequence of India’s startup success. The country’s most celebrated founders have become ambassadors for the ecosystem itself. In doing so, some public narratives risk confusing exceptional entrepreneurship with institutional excellence. The founders most capable of identifying what remains missing in the ecosystem are often those least likely to say so publicly. Having become symbols of national technological success, they are expected to reinforce confidence in the very ecosystem they once struggled to navigate. Few entrepreneurs would willingly invite friction with policymakers, regulators or the wider political establishment while building businesses that depend upon all three.The global race for AI is therefore being misunderstood. It is routinely described as a contest between algorithms, language models and technology companies. It is, in reality, a competition between national innovation systems built on societal capabilities.AdvertisementOpenAI did not emerge because the United States decided to build a chatbot. China’s advances did not materialise because of a single industrial policy. Europe’s continuing influence does not derive solely from regulation. Each reflects decades of investment in research, talent, industrial capability, capital formation and institutional confidence. AI is merely the most visible expression of those deeper foundations.Also Read | RBI has a playbook for a stronger rupeeDigital sovereignty in the age of AI will not be built through mutual-admiration groups between policy, capital and entrepreneurship. It will be built by the collective willingness to confront uncomfortable institutional truths before markets eventually do it for us.The greatest strategic error in the age of AI may therefore not be technological. It may be cognitive. Nations rarely fall behind because they underestimate their competitors. They more often fall behind because they overestimate themselves. The most dangerous hallucination is not produced by AI. It is produced by the stories societies choose to tell themselves.The greater difficulty is that the global conversation around AI has become excessively preoccupied with visible outcomes while neglecting the invisible conditions that make those outcomes possible. Large language models, billion-dollar valuations, sovereign AI missions and headline-grabbing investments dominate public discourse because they are measurable and politically attractive. The harder questions seldom receive comparable attention because they expose structural realities that cannot be altered within electoral cycles or quarterly earnings.The frontier of AI is not merely built upon algorithms. It rests upon a far wider architecture of scientific capability, advanced mathematics, semiconductor design and fabrication, high-performance computing, energy infrastructure, research universities, open scientific collaboration, entrepreneurial experimentation and institutions capable of sustaining all of them over decades. Every breakthrough model is merely the visible tip of a remarkably deep iceberg.This explains why the global AI race cannot be reduced to a competition between technology companies. OpenAI is not America’s competitive advantage. It is one manifestation of an ecosystem that has spent generations investing in scientific research, defence innovation, venture capital, immigration, university excellence and entrepreneurial risk-taking. DeepSeek is not China’s story. It is an outcome of industrial policy, manufacturing depth, engineering talent, patient state capacity and sustained investments that long predate today’s excitement around generative AI. Even Europe, often criticised for regulatory caution, is attempting to construct institutional legitimacy around artificial intelligence because it recognises that technology without governance eventually weakens public trust.India therefore confronts a far more demanding challenge than financing artificial intelligence. It must ask whether the surrounding ecosystem is evolving with equal seriousness. How many globally respected AI laboratories are emerging from Indian universities? Are our doctoral programmes attracting and retaining the finest scientific talent? Does private capital possess the patience to fund frontier research rather than only application-layer innovation? How many universities consistently produce frontier research rather than competent graduates? How many laboratories are advancing foundational science instead of adapting discoveries made elsewhere? How many Indian companies are building the next generation of chips, computing infrastructure or foundation models rather than applications layered upon them or becoming new-age realty-landlords for global digital giants? How many entrepreneurs can move from laboratory to market without spending years negotiating regulatory complexity? These are the very substance of technological capability.Scientific cultures mature slowly. Universities establish reputations over generations. Industrial ecosystems evolve through accumulated experience. Entrepreneurial confidence grows through repeated cycles of failure and reinvention. None of these can be accelerated simply because artificial intelligence has become geopolitically fashionable.you may likeThat is precisely why the current enthusiasm around AI demands greater intellectual discipline rather than greater rhetorical confidence. There is no doubt that India should invest aggressively in artificial intelligence. Importantly, it should encourage more entrepreneurship, expand computing infrastructure, strengthen research funding and build globally competitive capabilities.Nations seldom become technological powers because they dream boldly. They become global technological powers because they spend decades doing the unglamorous work that dreams rarely acknowledge.Sridharan is a corporate advisor and writer; Venkatanarayanan is a strategic security and digital policy researcher. Their recent book, Power of Uncertainty, is about the paradox of AI, influence, and sovereignty in the 21st century