End-to-end AI solutions need of hour: Netweb's Swastik Chakraborty

AI customers are shifting from buying infrastructure to seeking business outcomes, boosting demand for end-to-end AI solutions in India

The number of customers looking for business outcomes and overall success from the deployment of artificial intelligence (AI) and related infrastructure has increased manifold compared to those looking only for the deployment of AI infrastructure, based on the number of tokens they are likely to use or the number of graphics processing units (GPUs) they want, said Swastik Chakraborty, the vice-president of technology at Netweb Technologies.

 

“Earlier, these things used to be in isolation. There would be someone with a data centre de­sign, and someone else would be providing solutions for cooling, power, and other things. With the advent of AI, all of these th­i­ngs have converged, which in tu­rn calls for an intricate design and an end-to-end solution,” Chakraborty told Business Standard.

 

 

The company’s solutions, he said, now span services such as networking, providing the right AI model, and guiding companies on the right model governance approach.

 

“At the end of the day, the customer is focused on the business outcome for the investments they have made. Our primary focus remains on ensuring that we help them achieve that in the most cost- and reso­urce-optimised route,” he said.

 

The company, which recently raised ~1,200 crore through a qualified institutional placement, saw Indian and international investors such as Goldman Sachs Asset Management, Nomura, ICICI Prudential MF, HDFC MF, Kotak MF, Tata MF and others participate in the fundraising.

 

The company has seen increased order momentum from companies participating in the ~10,372 crore India AI Mission, the subsequent upturn in demand for sovereign AI infrastructure, and increasing demand for indigenous high-performance computing solutions, Chakraborty said.

 

For companies in India, the correct way to approach AI would be to start the journey towards becoming producers of AI services instead of just relying on being the consumers, he said. He added that while the investment in AI inference and infrastructure can continue, an equal focus needs to be given to creating more tools and services.

 

India also needs to adopt multiple alternative energy sources to address its lack of a sufficient and stable power supply, so that newer tools and services built on the world's leading frontier AI models can be deployed for everyone, he said.

 

One approach to solve this problem could be federated data centres, where access to data is controlled based on the needs of the developer or the consumer, Chakraborty said.

 

“In AI, every day that is lost is taking a couple of steps behind the world in the development race. If we have to be at the forefront, we need to be focusing on handling power and energy requirements and taking a different approach for data centre architecture,” he said.