AI services a real opportunity for IT to adapt, move fast: Vinod Khosla
The AI revolution will lead the world to an era of abundance where people will not have to work and will have the freedom to pursue their passion, Vinod Khosla, founder of Khosla Labs told ET’s Swathi Moorthy and Surabhi Agarwal in an interview. About his doomsday prediction for the IT services industry, the Indian-American technocrat and venture capitalist said Indian services companies have a “real opportunity” to offer AI services all over the world, but they have to move fast and adapt quickly to save themselves from extinction. He also called the launch of the latest model by homegrown Sarvam as ahead of where DeepSeek was a year ago. Edited excerpts:
You made a strong statement yesterday that perhaps by 2030, there will be nothing like the IT services or BPO industry anymore. What does it mean for India's economy?
The thing with technology is that either you change with it and leverage it, or it leaves you behind. The notion that there will be IT services over some period, whether it happens by 2030 or 2035, is hard to predict. The idea of providing software services to western enterprises will go away. AI agents can do any BPO work, probably more accurately and faster than almost any human. One of the things Indian companies can do is to say, ‘We won't hold on to that business but we will hold on to those customers and provide them new services.’
More importantly, the world isn’t as well equipped to adapt to AI as India is, especially the developing world. They have either very high costs in the West or very little knowledge, like in Africa or Southeast Asia. Indian services companies can do AI services all over the world, and that is the real opportunity. But only those who move rapidly will take advantage of it, and it will probably be a bigger opportunity than what they’re leaving behind.
India is a money-poor but people-rich and time-rich economy. With artificial general intelligence (AGI) pegged to arrive in the next 2–5 years, with job displacement in tow, what does it mean for India?
I think the next five years will look very great to economists. They measure productivity and GDP growth, and that looks great. At some point…and my bet is from 2030 to 2040, we will see strong political forces and politics influencing whether AI gets adopted in each country. You will see chaotic policies, politicians pushing certain benefits, and saying there can't be job losses, which will slow down AI adoption
But I have no question that GDP growth will accelerate because of it. But because productivity is growing, you might see job losses or disruption in jobs. People will have to change what they do dramatically. When you get into the next 15 years, from 2040 to 2050, I think there will be such a large production of goods and services that things like food will be cheap, and entertainment will be cheap. I know it sort of seems odd to believe, but by 2050, it’s going to be very exciting. It will truly be an era of abundance if politics allows it to happen.
You had a meeting with the prime minister yesterday. What can you tell us about the discussions?
The main thing is the idea that AI can make it possible to give every Indian a personal doctor who is available to them 24x7 and provide a level of healthcare that has not been possible and is affordable, at a cost much lower than what healthcare costs in India are today, and probably implemented through a system like Aadhaar. The other thing is that all 250 million children, and every adult learner, can use AI tutors for education. There are many Indian farmers, hundreds of millions of them, working on small plots. They can have a PhD-level agronomist available to them for free. These are not things we have to wait for in the future. I think they're available as soon as we adapt them from what is being done in the West for our country.
You have been talking about AI for a long time. Now that it has come into play, did you in any way anticipate the scale that we are seeing today?
It’s very hard to predict the future. But the one thing that wasn’t hard to predict is that when AI technology gets to a certain level, it would make a huge impact. So, the level of impact is not a surprise at all. The fact is that this societal impact will keep increasing both in India and globally for the next 10–15 years. It’ll be huge in every domain you can examine.
If you look at how the world order on AI right now is, a lot of investment as well as technology creation is currently dominated in the US and China. What does this mean for the Global South and India, which has ambitions to be a developed nation by 2047?
2047 coincides with the independence anniversary. I think that’s a good time to shoot for a future where nobody in India needs a job. If you’re a farm worker labouring in 110-degree heat, it’s not really a job. It’s servitude to survival. If those things go away, if there’s lots of production of goods and services, that’s what people need.
So, can the bottom half of the Indian population have a 10x better standard of living than we have today by 2047 without having a job? Yes. I do think by 2047, which would be an awesome goal for India, we can declare minimum standards of living in the country that are so high that nobody needs to worry about survival.
There are some hard areas, too. Housing won’t become free unless we dramatically change how we do housing. Technology breakthroughs in housing are very much needed to make it much more affordable for the 500 million people who don’t have great housing conditions.
Most of the investment in AI is going into the US and China right now. What should be India's playbook if it must really lead in the AI race? As a developing nation, we cannot afford to pump billions of dollars into models or infrastructure...
People like Sarvam are developing state-of-the-art models in a lot of areas, but not in every area. I don’t think we should rely on being totally independent, which is very hard, at least now. Maybe, by 2035, when there’s enough growth in the economy and we have far more than doubled our growth rates, we will have the money to invest.
I just want to give you an example of DeepSeek. It completely changed the paradigm of how much money it takes to develop a state-of-the-art model. Sarvam’s models today are ahead of where DeepSeek was a year ago. Money is an important tool, but creativity is an even more important tool. Talent and knowledge are important tools. And India clearly has the talent, knowledge, and expertise to do these things.
Coming to your bets as a venture capitalist — you have made some futuristic bets into OpenAI, Emergent and Sarvam AI. Tell us about the kind of ecosystem you are seeing in the US, primarily in AI epicentre San Francisco, and in India.
I think AI is much more dispersed than people think. What happens is the best people move to California because others like them are there, and they get a lot of support. These support ecosystems are very important. In California, you are unemployable if you work for a large company for 10 or 15 years. You must have startup experience. I know it sounds silly.
In this context of a world where AI will take care of most of the jobs, what does wealth and power really mean, considering that they are concentrated among a small percentage of the population?
I think it is the wrong focus. I think what is the right focus is what happens to the bottom half of the population. How well off are they? And do they have freedom? Wealth concentration is only a problem if it imposes a lack of freedom on the other 90% of the population. I think we can have both. Now, it's very hard to predict what happens to the notion of wealth. I think starting from 2035 to 2050, we'll have a hugely deflationary time. If onions cost 10 paisa a kilo, do you care about wealth? No. You care about relationships, following your passion, whether it's competing in the Olympics or doing your own photography. I do think people will have a lot more freedom.
You are talking about a deflationary time. But if you look at the past year, AI has also widened the economic divide between the rich and poor…
Why do you think that’s the case? I challenge that in terms of the facts. If a farmer who has a one-acre plot can get a PhD agronomist, why does wealth matter? I don't think the disparity has increased based on wealth. I think there are other factors that are more important. In fact, it has enabled people who are curious, which tend to be young people — that’s why so many of them are swarming to this summit. If curiosity becomes important because AI can help you learn, it helps only if you want to learn. If you don’t want to learn new things, then AI can’t help you. So, no amount of wealth will make up for what you can learn on the internet with good AI.
You made a strong statement yesterday that perhaps by 2030, there will be nothing like the IT services or BPO industry anymore. What does it mean for India's economy?
The thing with technology is that either you change with it and leverage it, or it leaves you behind. The notion that there will be IT services over some period, whether it happens by 2030 or 2035, is hard to predict. The idea of providing software services to western enterprises will go away. AI agents can do any BPO work, probably more accurately and faster than almost any human. One of the things Indian companies can do is to say, ‘We won't hold on to that business but we will hold on to those customers and provide them new services.’
More importantly, the world isn’t as well equipped to adapt to AI as India is, especially the developing world. They have either very high costs in the West or very little knowledge, like in Africa or Southeast Asia. Indian services companies can do AI services all over the world, and that is the real opportunity. But only those who move rapidly will take advantage of it, and it will probably be a bigger opportunity than what they’re leaving behind.
India is a money-poor but people-rich and time-rich economy. With artificial general intelligence (AGI) pegged to arrive in the next 2–5 years, with job displacement in tow, what does it mean for India?
I think the next five years will look very great to economists. They measure productivity and GDP growth, and that looks great. At some point…and my bet is from 2030 to 2040, we will see strong political forces and politics influencing whether AI gets adopted in each country. You will see chaotic policies, politicians pushing certain benefits, and saying there can't be job losses, which will slow down AI adoption
But I have no question that GDP growth will accelerate because of it. But because productivity is growing, you might see job losses or disruption in jobs. People will have to change what they do dramatically. When you get into the next 15 years, from 2040 to 2050, I think there will be such a large production of goods and services that things like food will be cheap, and entertainment will be cheap. I know it sort of seems odd to believe, but by 2050, it’s going to be very exciting. It will truly be an era of abundance if politics allows it to happen.
You had a meeting with the prime minister yesterday. What can you tell us about the discussions?
The main thing is the idea that AI can make it possible to give every Indian a personal doctor who is available to them 24x7 and provide a level of healthcare that has not been possible and is affordable, at a cost much lower than what healthcare costs in India are today, and probably implemented through a system like Aadhaar. The other thing is that all 250 million children, and every adult learner, can use AI tutors for education. There are many Indian farmers, hundreds of millions of them, working on small plots. They can have a PhD-level agronomist available to them for free. These are not things we have to wait for in the future. I think they're available as soon as we adapt them from what is being done in the West for our country.
You have been talking about AI for a long time. Now that it has come into play, did you in any way anticipate the scale that we are seeing today?
It’s very hard to predict the future. But the one thing that wasn’t hard to predict is that when AI technology gets to a certain level, it would make a huge impact. So, the level of impact is not a surprise at all. The fact is that this societal impact will keep increasing both in India and globally for the next 10–15 years. It’ll be huge in every domain you can examine.
If you look at how the world order on AI right now is, a lot of investment as well as technology creation is currently dominated in the US and China. What does this mean for the Global South and India, which has ambitions to be a developed nation by 2047?
2047 coincides with the independence anniversary. I think that’s a good time to shoot for a future where nobody in India needs a job. If you’re a farm worker labouring in 110-degree heat, it’s not really a job. It’s servitude to survival. If those things go away, if there’s lots of production of goods and services, that’s what people need.
So, can the bottom half of the Indian population have a 10x better standard of living than we have today by 2047 without having a job? Yes. I do think by 2047, which would be an awesome goal for India, we can declare minimum standards of living in the country that are so high that nobody needs to worry about survival.
There are some hard areas, too. Housing won’t become free unless we dramatically change how we do housing. Technology breakthroughs in housing are very much needed to make it much more affordable for the 500 million people who don’t have great housing conditions.
Most of the investment in AI is going into the US and China right now. What should be India's playbook if it must really lead in the AI race? As a developing nation, we cannot afford to pump billions of dollars into models or infrastructure...
People like Sarvam are developing state-of-the-art models in a lot of areas, but not in every area. I don’t think we should rely on being totally independent, which is very hard, at least now. Maybe, by 2035, when there’s enough growth in the economy and we have far more than doubled our growth rates, we will have the money to invest.
I just want to give you an example of DeepSeek. It completely changed the paradigm of how much money it takes to develop a state-of-the-art model. Sarvam’s models today are ahead of where DeepSeek was a year ago. Money is an important tool, but creativity is an even more important tool. Talent and knowledge are important tools. And India clearly has the talent, knowledge, and expertise to do these things.
Coming to your bets as a venture capitalist — you have made some futuristic bets into OpenAI, Emergent and Sarvam AI. Tell us about the kind of ecosystem you are seeing in the US, primarily in AI epicentre San Francisco, and in India.
I think AI is much more dispersed than people think. What happens is the best people move to California because others like them are there, and they get a lot of support. These support ecosystems are very important. In California, you are unemployable if you work for a large company for 10 or 15 years. You must have startup experience. I know it sounds silly.
In this context of a world where AI will take care of most of the jobs, what does wealth and power really mean, considering that they are concentrated among a small percentage of the population?
I think it is the wrong focus. I think what is the right focus is what happens to the bottom half of the population. How well off are they? And do they have freedom? Wealth concentration is only a problem if it imposes a lack of freedom on the other 90% of the population. I think we can have both. Now, it's very hard to predict what happens to the notion of wealth. I think starting from 2035 to 2050, we'll have a hugely deflationary time. If onions cost 10 paisa a kilo, do you care about wealth? No. You care about relationships, following your passion, whether it's competing in the Olympics or doing your own photography. I do think people will have a lot more freedom.
You are talking about a deflationary time. But if you look at the past year, AI has also widened the economic divide between the rich and poor…
Why do you think that’s the case? I challenge that in terms of the facts. If a farmer who has a one-acre plot can get a PhD agronomist, why does wealth matter? I don't think the disparity has increased based on wealth. I think there are other factors that are more important. In fact, it has enabled people who are curious, which tend to be young people — that’s why so many of them are swarming to this summit. If curiosity becomes important because AI can help you learn, it helps only if you want to learn. If you don’t want to learn new things, then AI can’t help you. So, no amount of wealth will make up for what you can learn on the internet with good AI.
Next Story