From chips to chatbots: How India is chasing AI billions

I am writing this from New Delhi, on the second day of the AI Summit, 17th February. The air here is electric. Not because of the weather. But because what is being discussed inside these halls could change the lives of every single Indian—from a farmer in Punjab to a teenager in a small town in Tamil Nadu.

Let me start with what Prime Minister Narendra Modi said just a short while ago. He gave an interview to ANI, the Indian news agency. And his words were bold. He said he wants India to be among the top three AI superpowers in the world. Not just as a country that uses AI tools made by others, but as a country that actually builds them. He said Indian AI models should serve billions of people in their own native languages. He said Indian AI startups should one day be worth hundreds of billions of dollars and create millions of good jobs. That is the dream on the table.

Now, big dreams need big money. And that is exactly what Day Two of this summit was all about.

The headline number is $200 billion. Yes, two hundred billion dollars. That is how much investment India is expecting to attract in AI over the next two years. This was announced by Ashwini Vaishnaw, India’s Minister for Electronics and IT. He also added that another $17 billion is expected separately—mainly for advanced AI research and real products that people actually use, like health apps, chatbots, and translation tools.

So where exactly will this money go? To understand that, you need to understand something called the five layers of AI. Think of it like a five-floor building. Every floor matters. Remove one, and the whole structure collapses.

The ground floor is energy. AI systems run on enormous amounts of electricity. Without reliable and affordable power, nothing else is possible. That is why energy comes first.

The second floor is data centres and networks. Think of data centres as giant warehouses where all digital information is stored and processed. And networks—your internet, broadband, 5G—are the highways that carry that information from one place to another. Without these highways and warehouses, AI has nowhere to live and no way to move.

The third floor is compute—meaning chips and semiconductors. These are the muscle of AI. They do the actual heavy lifting, processing millions of calculations every second. Without powerful chips, AI cannot function at speed, especially when millions of people are using it at the same time.

The fourth floor is the AI model itself. This is the brain. It is trained on vast amounts of data—the way a person learns by reading thousands of books over years. It finds patterns in that data and uses those patterns to answer questions, make predictions, or take decisions. This is what powers tools like ChatGPT.

The fifth and top floor is applications. This is the part you and I actually touch and use—a chatbot answering your question, an app diagnosing a health condition, a platform helping a farmer decide when to water his crops. These are the real products that bring AI from laboratories into everyday life.

Most of the $200 billion will go into the lower floors—energy, infrastructure, data centres, and chips. The additional $17 billion will go into the upper floors—the models and the apps.

Now, global giants are already moving. OpenAI, which created ChatGPT, has been in India for a while. On 16th February, Anthropic—another major AI company—announced it is opening an office in Bengaluru. This is not just about renting a desk somewhere. These companies are building real partnerships with Indian businesses and even the government. Air India is using Anthropic’s AI tool, Claude, to build custom software. Cognizant is rolling it out to its employees worldwide. Infosys will work with Anthropic to build AI agents—think of them as smart digital helpers—for industries like telecom and banking.

Google and Meta are also expanding their data centres in India. The message is clear: India is no longer just a market where foreign companies sell their products. India is becoming a partner—a place where AI is built together.

But here is the crucial part. India cannot depend entirely on foreign AI. Western models are powerful, but they often do not understand India deeply. India has 22 official languages and hundreds of dialects. Most Indians do not use English in their daily lives. And most global AI tools still struggle with this reality.

This is why two Indian-made AI projects matter. BharatGen, backed by the government, has unveiled Param-2, a model trained on all 22 scheduled Indian languages. Sarvam AI, an Indian startup, has launched an on-device AI that works even without internet—perfect for villages with weak connectivity.

The summit’s biggest days are still ahead—19th and 20th February. French President Macron will speak. PM Modi will address the gathering. More announcements are expected.

India’s AI moment has arrived. The only question is how well India grabs it.

(Disclaimer: The views of the writer do not represent the views of WION or ZMCL. Nor does WION or ZMCL endorse the views of the writer.)