'It's Just Like Colonial War...' - How India Is Powering AI Startup Ecosystem - Parag Naik & Dr. Satya Gupta Decode

Artificial Intelligence (AI) is primarily powered by the semiconductor ecosystem and at the Republic AI Pre-Summit 2025, Parag Naik, the CEO of Saankhya Labs and Dr. Satya Gupta, the President of VLSI Society of India & Former Chairman of IESA spoke at length explaining how the ecosystem in India has evolved, while powering India's AI startups.

India's Semiconductor Manufacturing Ecosystem

According to Parag Naik, the CEO of Saankhya Labs, "There's a lot of buzz about semiconductor manufacturing. The manufacturing ecosystem in India is very nascent, whereas there's another class of semiconductors called fabless design companies. These are companies that actually build chip. For example, Nvidia and QUALCOMM etc."

He added that practically every chip that is built in the world has a Bengaluru hand in it.

"It's just like the colonial war. We fought for the British and got a few Victoria crosses. The same thing happens most of the times we build the chips, but the credit goes to, you know, MNCs. So the design ecosystem is fairly robust, very well developed, but there are lots of service companies, but in the last 15-16 years, companies such as ours that eventually got acquired built some very sophisticated, fabless companies and semiconductors," he said.

On being asked why brands like QUALCOMM did not come out of India, Naik responded saying, "Semiconductor is not in fashion right? People laughed at us. But we got funded by angels. Most of my investors came from villages on NH-4 which is the highway connecting Bangalore and Pune. We did not have leadership around that time. That is the reason why it did not happen, but it will happen."

Dominating The Semiconductor Manufacturing Space

On being questioned about the possibility of India dominating the semiconductor manufacturing space, Dr. Satya Gupta, the President of VLSI Society of India & Former Chairman of IESA discussed the government's design linked incentive scheme.

He explained that the advent of AI and the foundation tech for AI was developed 30 years ago. The difference between that and today is the power of semiconductors, he noted.  

According to Dr. Satya Gupta, "many of you might have seen a picture of IBM 5 megabyte drive being loaded on the truck and today it is a small SD card you can have a terabyte of memory."

Both the combinations of compute power and memory power has been given by semiconductors and this has empowered the AI today, he said.

"Today we can run multi-billion parameter models and that is the difference. Nvidia's market valuation is equal to the Indian GDP today. This is because semiconductor is essential to AI," he added.

"While companies today are doing semiconductor design services for most of the companies, the government has this design-linked incentive scheme which is basically targeted at creating fabless companies. When we started out the VC ecosystem was not there but the situation has been reversed now," he said.

According to him, it is important to have the mechanism to invest in the semiconductor side of things.

Transition From CPU To GPU

According to Naik, "one of the OpenAI founders said that why not try GPU, which has a lot of horse-power, and that is how the AI infrastructure evolved. From video games to AI. "

He added that there is a need to design custom hardware.

Additionally, according to him, "from a power consumption perspective GPU is very inefficient and if we continue on the same trajectory that we are going today, you will not have global warming, you will have global heating."

A lot of people in the US are already investing in nuclear fusion to even power these AI data centres. They are simply power-hoggers. There are fare more efficient ways to do this, he noted.

Republic AI Pre-Summit 2025

The Republic AI Pre-Summit 2025, hosted by Republic Media Network in Noida on July 25, brought together top minds from across India—policymakers, innovators, and industry leaders—to shape India's AI future. Topics included the National AI Mission, GPU infrastructure, and AI in healthcare, education, and governance—all aimed at building a “Viksit Bharat by 2047.”

Also Read: India’s High-Tech Innovation and UK’s Expertise Set to Revolutionize Financial Sectors: NSE CEO