As India hosts global leaders including France’s Emmanuel Macron and Brazil’s Luiz Inácio Lula da Silva at the
AI Impact Summit that opened on Monday, New Delhi is positioning the conclave as a decisive moment in shaping a more inclusive, development-driven global AI order. This is the first time that the AI Impact Summit is being hosted by a developing nation.
“People, Planet and Progress” are the pillars for the AI Impact Summit 2026, being coordinated by the Ministry of Electronics and Information Technology (MeitY), the official statement released by the Indian government said. Firstpost’s Rajat Mishra spoke with MeitY Secretary S Krishnan on India’s AI policy and plans for technology advancement in the view of the AI summit.
Krishnan laid out India’s blueprint for democratic access to AI infrastructure, semiconductor self-reliance under ISM 2.0, data sovereignty safeguards, and strategic tech resilience in a fractured US–China landscape.
Edited excerpts:
India is hosting the AI Impact Summit, expecting global leaders including France’s Emmanuel Macron and Brazil’s Lula da Silva. What does India expect to gain from the mega AI conclave?
Fundamentally, India wants to emphasize that access to AI needs to be democratized and inclusive. We should not create a situation where technology and access to it are dominated by a limited number of companies in a few parts of the world, depriving the rest of the world of the benefits of development through its application.
Our approach is to ensure that — which is why it is called the Impact Summit — AI truly delivers development outcomes across the world, particularly in the Global South, including India.
We believe India is well suited to take on this role for several reasons. Through the India AI Mission, we have adopted a model that ensures AI infrastructure — compute, models, and data — is more widely accessible. It is a unique model that we are implementing, and it can work for the rest of the world as well.
That is one key element: we have adopted a fairly frugal yet democratic approach that allows access to a large number of stakeholders.
The second key element is that while infrastructure is important, the real impact of AI will be felt through applications and solutions built using that infrastructure. These applications must serve real sectors of the economy — agriculture, manufacturing, healthcare, education, governance.
If we create the right infrastructure, it can benefit people across the country. So these are the two core objectives of the summit.
We also want AI to remain human-centric. Governance frameworks must ensure that people remain at the center.
Another outcome is the opportunity for Indian businesses, startups, IT firms, and technology companies to showcase their AI applications. This allows them not only to demonstrate their capabilities domestically but also to present their work to the rest of the world.
So the summit serves as both a policy platform and a global showcase.
Since India’s AI focus has been on making the technology accessible to the masses, and the summit emphasises people, planet, and progress, is India risking lagging behind tech innovators?
I do not think we should look at it as lagging behind tech innovators.
Many of the leading AI models have had India-educated or Indian-origin scientists and technologists behind them. So where are we lacking?
We are working on our own models, and they are being developed and will be launched shortly. The prime minister [Narendra Modi] has also met the teams involved. From that perspective, we are not lagging behind tech innovators.
In the infrastructure space, yes, in terms of GPUs and related hardware, we do not yet produce them. Eventually, we should develop that capacity. Our semiconductor journey has just begun and will accelerate over time.
Until then, we do not want innovators and researchers in India to be denied access to up-to-date AI infrastructure. We are enabling that access.
AI servers are already being assembled in India. Many components are manufactured domestically. While we currently import GPUs, the capability to assemble and integrate systems exists.
So in many areas, we are catching up. There are areas where we still have some distance to cover, especially in GPUs. But we will get there.
India’s strength lies in its talent and research capacity. That is why we are recognised as a vibrant AI ecosystem and rank third globally after the US and China.
So I do not think we are lagging in innovation or technological capability.
Is India at risk of becoming a data provider to foreign AI platforms instead of owning foundational AI technologies? What do you have to say on that?
The point here is that data is something that can be used. We are building models of our own, and those models are going to use Indian data.
As far as personal data is concerned, it can be shared only with consent. Consent must be provided, and there is a framework within which that consent is structured.
I personally believe that we are not heading toward that scenario.
Once our own models are available and we use them, there are also approaches by which we can use the power of overseas or open models without necessarily sharing data.
For instance, if models are hosted within India — which is already happening — the data remains within the country and is used within the country.
Further, there are mechanisms to create an orchestration layer within India, where databases can be queried locally. You generate the query, access the AI model’s reasoning capability through that orchestration layer, and do so without sharing the underlying data.
There are multiple ways this can be designed, and we are working on them.
Therefore, I do not think the fear that India will become a data colony is valid.
You had said that AI poses the highest risk to white-collar jobs and that the focus must be on upskilling. Given the layoffs happening across corporate India, how do you see this?
You need to understand that statement in context.
In previous episodes of automation and mechanisation — the earlier industrial revolutions — what got replaced were jobs done with manual effort.
Right now, for the first time, knowledge-oriented jobs, or jobs performed based on specialised knowledge, are at greater risk. That is different from previous industrial revolutions. That was the context of my statement.
I also said that in the Indian context, the total number of white-collar jobs we have is limited compared to the West. So the risks to India are lower.
Second, our white-collar jobs can be realigned. Instead of writing code or programs, professionals can move toward creating AI solutions and applications and then helping deploy them. That is where upskilling comes in.
The Niti Aayog report points out that while we may lose about 2 million jobs in one area, the potential exists to create about 4 million jobs in another area within the IT industry.
That is what we need to prepare for.
We know that the budget has proposed ISM 2.0. I just wanted to understand how ISM 1.0 is different from ISM 2.0.
You will see the outlines of ISM 2.0, but the focus now is to build on whatever base we have created under ISM 1.0.
In ISM 1.0, we focused on two or three aspects. First was the establishment of a fab and packaging units for manufacturing. Second was limited assistance for startups and MSMEs operating in the semiconductor design space. Third was the modernisation of the semiconductor laboratory. Fourth was creating human resources for semiconductors.
We have made progress on all these aspects under ISM 1.0. Now we need to take it to the next level under ISM 2.0.
There will still be focus on establishing fabs and advanced packaging units because we do not yet have an adequate number and that needs to move forward.
Second, we need to look at the overall ecosystem — manufacturers of equipment, materials, gases, and chemicals. All of them need to be supported so that the entire ecosystem develops.
Third, we need to focus more strongly on the design space. Earlier, our scheme supported only MSMEs and startups. We did not support larger design companies in India. However, design requires resources and involves risk. We need to enable that to happen.
For that, we need to scale up the assistance provided and also crowd in venture capitalists and private equity to invest in companies working in this space so that intellectual property can be created.
These are some of the issues we need to address in ISM 2.0 in addition to what has already been done under ISM 1.0. We will be moving to the next stage of our semiconductor journey.
Are there any concrete timelines for the first commercial production of 28nm or below chips made in India? Do you have a timeline for 28 nanometer chip production?
The unit coming up in Dholera is technically expected to go up to 28 nanometers. They are finalizing the technology and will move forward with it.
As per the contract, there is an implementation period of about three to four years. It is closer to four years for implementation of the program. They will have to complete it within that period.
Let’s discuss geopolitics. In a world fractured by US–China tech rivalry, is India aligning with a bloc or positioning itself as a neutral manufacturing bridge?
It is important to understand that every large country and geography needs resilience and some degree of strategic autonomy in key technology areas. Recent geopolitics, and even the Covid situation, have shown that you cannot become over-dependent on one geography or one particular part of the world. You need inbuilt resilience.
For that resilience to be created, there is considerable work that India has to do. We are in this effort to create strategic autonomy for ourselves.
At the same time, we recognise that electronics and semiconductors operate within a global value chain. Nothing is done in only one country.
But you cannot become over-dependent on one particular country. You need two or three alternatives. That is the kind of resilience we are aiming to create.
We are also contributing to a global objective by helping build resilience in the supply chain of these key materials and strategic technologies. That is the second aim.
Third, by building this capacity, India can become an indispensable part of at least a segment of this global chain.
This is the direction in which we are moving. It is not just in India’s interest, but also in the interest of global stability and resilience
India is now being talked about in forums like Pax Silica. Is India actively positioning itself as a third pole in the global semiconductor and AI supply chain?
I would not describe it as first pole, second pole, or third pole. Different players in the semiconductor supply chain have indispensable roles.
For example, a company like ASML, based in the Netherlands, is indispensable to the semiconductor ecosystem. It is not in the United States or China, but it plays a critical role.
Similarly, many Japanese firms manufacture high-purity materials that are essential to semiconductor production. They are also a vital part of the global value chain.
So there are multiple locations in this ecosystem, and each has specialisation and importance in a particular area.
What we want to build is specialisation, skill, and capability within India so that we form part of that overall global value chain.
In doing so, we offer alternatives and contribute to greater resilience by diversifying the supply chain. That is the space we are aiming for, while also ensuring some degree of strategic autonomy within the country.
What is the biggest execution risk in India’s semiconductor and AI missions that policymakers publicly don’t talk enough about?
It is not that we are hiding anything. The point is that semiconductor manufacturing is a form of advanced manufacturing technology.
So far, we have strong capabilities in design. Almost 20 per cent of semiconductor design manpower is in India. But we have very little presence in the actual manufacturing of semiconductors.
Manufacturing requires a different set of skills. We are just starting to build that capability, so we have to ensure that the necessary skills are developed.
That will take a few years. We are working actively to create a skilled workforce to operate in this space. It will take effort and time, but we will do it.
India is eyeing advanced 5 to 7 nanometer chips as the semiconductor mission enters the next phase. What is the progress on 5 to 7 nanometer chips?
The point is that we have to look at a development path that is both affordable and manageable given our resource constraints.
We identified legacy chips as a sweet spot where the cost trade-off was appropriate and where they are still extensively used.
In the next generation, we are examining whether 5 to 7 nanometers is a space we can enter and build capacity in, starting from the research stage and developing what is required.
So at what stage are we right now as far as 5 to 7 nanometers?
We are at the starting stage. We are trying to figure out how to structure it. That is something we will have to work through. It is part of ISM 2.0.
India is entering semiconductors decades after Taiwan, Korea, and China. Are we trying to catch up or building a different model? What is the honest timeline to competitiveness?
Let’s be clear — it’s not that India didn’t try before; it just didn’t work. We have tried a number of times. There have been sporadic efforts. A semiconductor lab was set up in the 1960s. It was also attempted in the early 2000s. Some overseas companies tried to set up in India, but we could not enable it at that time.
It is only now that we are in a position where there is serious interest and serious capability to ensure that this will actually get set up.
It is a long journey and an evolving journey. As the Prime Minister himself said at the last Semicon event, it will require at least a decade or two before India gets established in this space.
But we have to do it. If you do not start, you will not be able to do it. We have started, and the start is very promising.
The India Semiconductor Mission was launched in 2021 and revised in 2022. The first agreements were signed in 2023. In the course of the calendar year 2026, about four units should come on stream.
The progress has been significant. We recognise that there is still some way to go, both in terms of technology nodes and capacity. But that does not mean we should not start now or continue to work hard to create a niche for ourselves in this space.
What stands out for you in the Union Budget?
In some ways, you can call it a tech budget. A number of aspects of technology have received attention in the budget. Both the electronics and IT sectors have been very enthused by the treatment given to them in the budget.
And the recent budget changes aim to clarify taxation for foreign data centers. What is the intended impact on investment and local jobs that you see?
I think there are two parts to this. One part is that some people have described it as a tax holiday. I don’t think it’s a tax holiday. What it basically does is clarify what the taxation impact would be.
No country seeks to tax exports, and this is entirely meant for foreign operators — cloud or international operators — who use data centres in India to store overseas data and provide services overseas. It is not meant to store Indian data or provide services to Indian customers. If that is being done, then it has to be through a separate vertical. It should be treated on a level-playing field with other Indian companies servicing the Indian market.
This is primarily meant for the overseas market, and we do not export taxes to the overseas market. So we are just clarifying the tax treatment for that segment. That’s all. Let’s be clear on that.
What it will do is encourage the establishment of data centers in India. Many announcements have already come in regarding data center expansion. In doing so, it supports a number of jobs because an ecosystem builds around it.
It is not only construction, maintenance, and related jobs, but also an ecosystem that develops around it, including development centers and associated services. So we expect a fair amount of job creation, particularly at the higher end of the business.
If the first Indian fabs go live with older node technology, how do we prevent them from becoming obsolete before scale is achieved?
I don’t think they necessarily become obsolete. There are many applications, particularly in automotive technology and consumer electronics, which only need semiconductors of that technology. The very advanced, very small semiconductors are basically needed for mobile phones and similar devices. Even now, for missiles and space applications, 180 nanometers is good enough. So you don’t need it to be miniaturized beyond that. I think the legacy nodes continue to have, and will continue to have for quite a while, a substantial market. So I don’t see obsolescence as an immediate risk now.
How will success of India’s AI mission be measured? Will it be measured in the startups created, global patents, GDP contribution, or something else? What is your assessment?
Ultimately, what it does for human prosperity is what matters. How it enhances human capability, how it enables better quality of services to be delivered, and how it contributes to an increase in real output in the real sectors of the economy, thereby increasing prosperity. I think that is how it will be counted.
And how do you allay fears of young students, especially if educational institutions are not upgraded?
I would allay the fears of students by saying that the nature of jobs in many of these sectors may vary, but there will be jobs.
What is most important in many of the new jobs that will come up in areas where AI will have a big influence is that AI is a horizontal technology implemented across all sectors. So you do not have to worry about whether you can code or program.
What you have to focus on is whether you know your domain or your sector well enough to use AI to enhance your productivity. You should become really good in your respective domains and know how to use AI.
That means all students will need to understand how to use this technology so that their productivity and capability increase.
AI could deliver things we previously thought were only in the realm of imagination. So to that extent, there is a very bright future. The key is to be solid in your domain and learn how to use the technology.
End of Article
Leave a Reply