In his first major media interaction nearly 10 months after taking over as Intel India President, Gokul Subramaniam who also is the VP, Client Computing Group of the company, talks to businessline about democratizing AI, India contributions, and accelerating the Indian semiconductor ecosystem.
Can you give us some perspective on global developments within Intel?
Intel previously had product business with its own integrated design and manufacturing. Today, we are transforming Intel through two core businesses: foundry and product. One big shift is regaining leadership in manufacturing and opening that up for our products and the overall industry.
On the product front, we drive technology across diverse sectors including data center, cloud, network, edge computing, and personal computing (PC). Intel’s purpose of enriching people’s lives can’t happen with just one product or aspect.
AI has brought about a huge shift comparable to how the Internet took the world by storm. This also means the opportunity for semiconductors has increased tremendously. At Intel, we aim to democratize AI. We are probably one of the end-to-end semiconductor companies to play in the AI era.
Why now the re-emphasis on foundry?
To be successful in semiconductors, you need a manufacturing technology leadership. Having a resilient supply chain to manufacture and delivering to the demand that AI will bring—close to a trillion dollars worth of semiconductor total addressable market (TAM)—needs to have a more distributed, resilient supply chain.
Our reliance on a specific geography, or one or two big players, will not be feasible. Part of doing more of the foundry in the West is to bring back that resiliency.
On the product side, where will most of your growth come from?
AI is revamping the semiconductor industry because the silicon needed to support it is a lot. Intel’s product portfolio is focused on this continuum. Data centers, Xeons, GPUs, and dedicated accelerators for deep learning. Countries like India are deploying 5G and 6G with ORAN and there is a boom of industrial 4.0, the manufacturing edge, and IoT.
Personal computing in PCs does not personalize to the point of specificity, opening up possibilities. A year ago, people were only talking about LLMs (Large Language Models). Now, there is a discussion of SLMs and TLMs. Compute is being diversified. We will be playing in this space across the board, both in the US and India.
Enterprise AI suite does not need to be driven only by LLMs and GPUs but also by CPU-level performance. We are witnessing a boom of designing LLMs specifically for India, and are engaging with these companies.
Can you talk about Intel’s AI PCs?
We launched Meteor Lake, our Core Ultra processor, in December. Currently, we are on track to get 100 million AI PCs by 2025. Since then, we already have more than 100 Independent Software Vendors (ISVs) building applications.
With our AI PC acceleration program, people are writing applications to make it more AI-centric and AI-experience-driven. We will also get more than 300 AI-accelerated features and 100 ISVs.
Soon, we will launch Lunar Lake along with about 80 unique designs from 20 OEMs. This has roughly four times extra performance on AI capabilities over Meteor Lake and includes CPU, GPU, and NPU.
We have started conversations about mainstream AI PCs. In India, AI PCs likely will pick up more adoption, which means working on better affordability. This also means we need more relevant population-scale AI applications.
Are you working with any local partners?
Intel has been working with the ODM ecosystem and will continue to. We might be the first company to incubate more than 8 to 10 ODMs in India. The second ecosystem is the larger engineering population, startup ecosystem, and ISVs to build more applications around it.
We are also engaging very actively with academia to set up AI PC experience centers in universities where we take the multidisciplinary capability of a university campus to build applications relevant to population scale. We will start doing it this year.
What contribution is India making to Intel globally?
We have a substantial presence in India, with the largest design center outside the US and 13,000 engineers. We are probably the only end-to-end semiconductor company handling everything from instruction set architecture and various IPs (CPU, GPU, and NPU), to chip design, system hardware design, system software, and open ecosystem.
Additionally, we have manufacturing capabilities and foundry-related capabilities. Our products span multiple domains, including Data Center, Cloud, Network, Edge, and PC. In Intel India, we mirror this end-to-end capability.
Around 90 percent of our overall headcount investment here is in engineering. Our teams contribute to the foundry, and product design - from instruction set architecture and microarchitecture to SoC design, and open hardware & software systems.
As a design center, Intel India has significantly contributed to the 12th and 13th-generation Intel Core processor SoC on the client side in the past 2-3 years. We have also played a key role in developing leading Intel Xeon product lines like Sapphire Rapids and the more recent Emerald Rapids.
Moreover, we are supporting Gaudi, the deep learning AI accelerator, with both Gaudi 2 and Gaudi 3-related work happening here. We contribute to foundry services-related engineering work and have been actively involved in network and edge – Intel Tiber – making AI easier to deploy on the edge.
Can you mention Intel’s contribution to the startup space?
We incubate and accelerate startups, leveraging our technology expertise to mentor them from the early stages. Once these startups have developed solutions, we assist them in optimizing these solutions on suitable platforms.
Thirdly, we facilitate go-to-market strategies to create opportunities for them. Our strong engineering presence with sales and marketing, particularly in the Indian market, has significantly expanded our opportunities.
Even for AI PCs, when we want ISVs, we look to both startups and major players. We put our technology experts to mentor and work with small players who will get a good market opportunity through our sales and marketing with a bigger player.
A few deep-tech and AI startups that the Intel Startup program has enabled with resources, mentorship, and access to cutting-edge technology are A5G Networks, and some Bengaluru-based companies like Calligo Technologies, Niramai Health Analytix, Minionlabs India Pvt Ltd. and Pantherun Technologies Pvt Ltd. Intel Capital invests in startups shaping the future of compute across four key areas - Cloud, Devices, Frontier, and Silicon.
With more than $20 billion invested globally over the past three decades, we are always looking for companies that push the bounds of innovation.