As computing transitions from the information era to the intelligence era, the role of the PC is being redefined—from a tool that analyses data to one that generates and acts on it. Intel is positioning itself at the centre of this transition. During the company’s first-quarter 2026 earnings conference call, CFO Dave Zinsner shared that the company’s AI PC revenue grew 8% sequentially and now represents greater than 60% of its client CPU mix.
In this conversation, Santhosh Viswanathan, Vice President and Managing Director – Intel APJ, outlines how AI PCs differ from traditional systems, and how India’s low PC penetration presents a unique opportunity.
What are the benefits of AI PCs as opposed to regular PCs?
PCs solve for productivity and information. While a traditional PC services the information era well, AI PCs are built for the intelligence era. Earlier, this device got us all the data, which we analyzed to build products and solutions. Now, agents can do many of those things. The PC will become a personal AI factory in many ways; it will be a source of agents and intelligence. AI PCs are more efficient because they can run models locally and have real-time back-and-forth communication. Whereas a traditional compute device requires one to go back and forth to the cloud. From a cost and frugality perspective, this is an effective model in which enterprises and consumers can run AI because it encourages smaller, local models to run natively on available data. For India, the biggest advantage is that an edge computer is the most cost-effective way of running AI at scale.
What’s the global vs India adoption rate for AI PCs?
India’s PC adoption remains below 10%. In the US and other mature markets, it is over 90%. PC adoption is around 40% in Brazil and about 20% in Indonesia. We are way below in embracing even a regular PC. In the information age, one had options like books and teacher-based models. But in the intelligence age, things will change. India can’t afford to stay at this 10% ratio. Every young person must have access to compute. AI PCs will start to hit mainstream price points. The broader adoption in established segments will now use more AI PCs. By default, every PC will be an AI PC at some stage. The question is, how fast can we get there?
How does AI/AI PC adoption differ between consumers and enterprises—and within enterprises, between large firms and MSMEs?
India’s challenge is not just AI PC adoption, but PC adoption itself. The majority of the organized enterprise sector, and some medium-sized companies, have access to compute. Where it remains underutilised is among the youth -- middle school to high school students. In the large enterprises, a lot of fixed computing exists. So while adoption in large enterprises is good, education and really small businesses primarily lag.
How do you see AI PC adoption evolving? Will users transition gradually from regular PCs, or could we see a direct leap to AI PCs?
Anybody looking for a PC will probably seek an AI PC. Separating the three compute engines, we get CPUs, GPUs, and NPUs. While one can run everything on a CPU, the battery will probably not last long enough. The performance of some AI models will be mediocre because CPUs are like specialized computing platforms, which do well in precision areas. Meanwhile, GPUs have parallel instructions. And NPU is the low-power, most efficient, nascent type of engine with low latency. It becomes an efficient architecture because it uses the best of all three capabilities. To reiterate, anybody who wants to buy a PC will look for an AI PC, because as the number of agents increases, the PC can become a personal AI factory.
What work does Intel carry out in India?
We have the second-largest R&D site in India, outside of the US. So across all business units, we have key capabilities based out of Bangalore. India has significant touch points across clients, data centers, and the network as part of the extension of our global capabilities here. It’s a key engine for our growth here.
Are you collaborating with the government on the India AI Mission—particularly in terms of providing compute or related infrastructure?
As agentic AI scales, CPUs become more critical. Earlier, they mainly acted as head nodes in data centres, but with agentic AI, their role expands. The architecture of data centres is rapidly evolving, with GPUs combined with CPUs now forming the core architecture. This is transitioning rapidly. We do partner with the government on the AI mission. The first stage was great. We are training many GPUs, and have the local model. The next stage of agentic AI requires data centers to look different. It requires the framework of agentic AI, and there could be different models in which you scale. We would closely work with government stakeholders on the same.
Published on April 26, 2026