Generative AI is evolving every week, but the adoption is happening at a slower rate: SymphonyAI's Romesh Wadhwani

The Silicon Valley tech billionaire on the current hype around GenAI, partnering with Nvidia, threats from artificial intelligence, and the need for regulating it

By Naandika Tripathi Forbes India Staff

Published: Sep 5, 2024 01:04:47 PM IST

Updated: Sep 5, 2024 08:01:22 PM IST

Full Bio

Romesh Wadhwani, chairman and founder of SAIGroup. Image: Timothy Archibald

It's an exciting time to be in AI (artificial intelligence), says Romesh Wadhwani, who has seen five generations of AI. The previous cycles were limited in capabilities. They were less flexible and agile, and the barriers to adoption at scale were high. On the contrary, generative AI is far more powerful and universal in terms of its applications, explains the 77-year-old Indian-American entrepreneur and philanthropist.

GenAI is rapidly evolving. “Our yearly disruptive technology review has turned weekly now. It is impossible to keep up with fast-moving technology. But we do the best we can,” says Wadhwani, chairman and founder of investment firm SAIGroup and its portfolio of three enterprise AI software technology companies.

Before GenAI kicked in, Wadhwani in 2017 launched SymphonyAI, which helps firms like Marks & Spencer enhance efficiency of store operations, enables retailers predict demand, and assists financial companies in detecting fraud. The Palo Alto-headquartered company is now preparing to go public in the second half of 2025, as it reached $500 million in revenue run rate last year and achieved profitability after growing revenue at a rate of about 25 percent.

SymphonyAI has a market cap of $4 billion with $310 million in annual revenue. In July, L&T Technology Services announced a partnership with SymphonyAI to bring AI operations to enterprises worldwide. The solutions are delivered to the customers through predictive and generative AI-based IT services and enterprise service management platforms to enhance productivity and simplify work.

Overall, “the group now has revenue of $800 million. We're approaching $1 billion in revenue just seven years into our startup”, Wadhwani tells Forbes India in a virtual interaction.

SAIGroup has expanded its business and launched two more AI companies—ConcertAI in 2018 focussed on biotech, and generative AI-native health company RhythmX AI in 2023 with a $50 million investment. In 2022, ConcertAI gained $150 million in venture capital at a $1.9 billion valuation and raised a total of $600 million. It recently announced a collaboration with Nvidia to develop new clinical simulations that inform future AI products for clinical trials. “We partner with the likes of Microsoft and Nvidia, but we build our own intellectual property as well, and then we build applications on top of it,” adds Wadhwani.

After obtaining a BTech degree from IIT-Bombay, Wadhwani moved to the US in August 1969 to pursue further studies and arrived in Pittsburgh, Pennsylvania, with $2.5. He got a master’s—followed by a PhD—in electrical engineering from Carnegie Mellon University. During his PhD days, Wadhwani was clear about not working for any firm and starting his own company instead. Multiple rejections and failures later, the Silicon Valley technology leader is now worth $5 billion. A decade ago, he pledged to give away 80 percent of his wealth. Wadhwani and his brother Sunil founded the Wadhwani Institute of Artificial Intelligence at the University of Mumbai in 2018, committing over $30 million in support for social good.

Also Read: What business leaders really think about Generative AI

In a candid conversation, Wadhwani speaks about why the impact of AI will be everywhere in 10 years, the risk of AI being applied for bad outcome is high, the need for regulating it, and more. Edited excerpts:

Beyond Generative AI

Generative AI is just the beginning. There is a lot of innovation that goes beyond it. For instance, agent AI, where AI-based tools and modules will perform important tasks and functions on their own autonomously, just as human beings do. These tasks will be performed by autonomous AI agents who will be able to work on their own. They will collaborate with each other and also have hierarchies, as we currently see in organisations. Humans can organise these agents into networks that connect to each other. Soon, humans and agents will be working side by side. This is the new world beyond generative AI that is unfolding daily.

OpenAI: The poster child of GenAI

When they first launched ChatGPT about two years ago, it set off a wave of excitement, and that caused a lot of AI developers to speed up their activities. It led to other people forming companies like Anthropic, Cohere, and a bunch of others who've all gone out and built different kinds of large language models (LLMs). It certainly accelerated enterprise AI because all the SymphonyAI customers have shown great interest in it. “I would credit OpenAI for opening everyone's eyes to the possibilities of AI.”

The AI hype cycle

All new technologies go through a hype cycle. A new technology emerges, everyone's interest spikes, and then they feel it can't get all their problems solved. That's the trough of disappointment. Post the overhype phase, the adoption of the technology takes off and goes through the roof over a period of time. “I think we are very close to the peak of the hype cycle. Will there be a decline in expectations, possibly?” After that, there will be absolute steady growth in the use of AI everywhere:

With consumers: New kinds of search engines.

With businesses: New kinds of business transformation.
With the government: The ability to have more effective policies and programmes.
With philanthropy: The ability to help millions of people instead of a few thousand people.

The impact of AI will be everywhere over a 10-year period. The technology is moving every week, but the adoption of the technology is not moving every week. “It is happening at a much slower rate because human beings, organisations, and particularly governments, take time to absorb technology and apply it. But I think that wave is happening.”

Threats from AI

At the current level of AI capability, the threats from AI are limited. “But when I hear things like OpenAI is planning to build GPT6, and the project is called Stargate, with an investment of $100 billion… at this level of capability, I do think that the risk of AI being applied for some bad outcome is quite high.” With that level of capability, it can automatically generate the next coronavirus, except that it would be a virus that may be a billion times more powerful, and no vaccine may be possible for 20 or 30 years because of its complexity. “This worries me. The capability of GPT6 could hack into nuclear systems or satellites in space. It can trigger wars. That’s certainly a risk.” The governments will have to be much more thoughtful and move a lot faster in terms of regulation.