India showcases homegrown AI models at Modi summit

Images

×

Indian Prime Minister Narendra Modi takes a group photo with leaders of artificial intelligence companies at the AI Impact Summit in New Delhi on Thursday. japantimes.co.jp

A solar farm in Nakai, Kanagawa Prefecture, in March 2016. Japan gets about a tenth of its electricity from solar panels despite having nearly no domestic production of photovoltaics (PVs). japantimes.co.jp

Sonic the Hedgehog, Castlevania's Alucard and the weak yet lovable Slime from Dragon Quest are just some of Japan's iconic gaming franchises celebrating midlife anniversaries in 2026. japantimes.co.jp

Japan agency aims for both decarbonization, growth japantimes.co.jp

US ‘totally’ rejects global AI governance, White House adviser tells India summit france24.com

‹ ›

India is trying to manufacture its own “DeepSeek moment”—a sudden leap into globally competitive, low-cost AI models—by putting startups on stage at a government-backed summit and framing “AI sovereignty” as a national project.

The Japan Times reports that Indian artificial intelligence companies presented domestically built tools this week in New Delhi, amid official ambition to become a major AI power. The political theatre includes photo ops with Prime Minister Narendra Modi, grand language about global leadership, and a promise that India can build models tailored to local languages and domestic needs.

But analysts quoted by the Japan Times are blunt about the gap between aspiration and capability. Replicating China’s recent surge—where a high-performance chatbot became a symbol of cost-efficient scaling—would require not just clever engineering but deep access to compute, high-quality training data, and a mature ecosystem for model deployment. India has talent, but it also has bottlenecks: hardware supply chains, capital intensity, and the chronic temptation to substitute policy branding for competitive pressure.

France 24, previewing the summit’s closing, highlights the event’s international framing: “world leaders” presenting a joint approach to AI. That language is doing a lot of work. A “joint approach” often means harmonized rules, shared compliance frameworks, and safety regimes negotiated by governments and large incumbents—exactly the kind of architecture that entrenches the biggest players while advertising itself as protection.

India’s push for “homegrown” models could, in principle, be a pro-market story: entrepreneurs building products for a massive user base in dozens of languages, competing with Silicon Valley and Shenzhen. Yet the risk is that “sovereignty” becomes code for licensing, grants, and procurement favoritism—where committees decide which models count as national champions and which startups get access to subsidized compute.

India’s best comparative advantage—its chaotic, competitive private sector—does not naturally align with state-directed AI roadmaps. If New Delhi wants a genuine breakthrough, it will need to do less selecting and more removing friction: cheaper and more reliable power, easier imports of advanced GPUs and networking gear, predictable rules for data and liability, and procurement that rewards performance rather than connections.

Otherwise, India may still get its “DeepSeek moment”—just not the one it wants: a demonstration that industrial policy can produce impressive stage demos while the real innovation happens elsewhere.