Intel’s strategy to revitalise with AI innovations: Steve Long and Santhosh Viswanathan share insights

Steve Long: For Intel, it starts with semiconductors. The whole economy is built on these transistors that continue to get smaller and smaller, and we can do more with them. So, we don’t see an end to that soon. We have growth opportunities for us to play a bigger role in solving customers’ challenges. We are increasingly finding that people are looking for Intel to come in and help them turn all of these connected devices into intelligence, outputs, and data.

Then, India and I would put parts of Southeast Asia, but India is ahead of it in terms of demographic advantage, where people are increasingly connected. They are at a prime in the demographic makeup of the country, and they are starting to participate on a much bigger scale/economy. So, for us, that’s also quite exciting. On AI, we are in the first chapter. The first chapter is a very early stage, it’s been around for a long time, 40 years, and people have been talking about AI and using different parts of AI, but we think we’re at a precipice of a big, new era.

So, it’s an inflexion point, we think what Centrino was to the PC with wireless experience AI will be to both. Not only a PC experience, but AI will infuse the workload across every type of device. Now today, there’s obviously buzz around ChatGPT on Generative AI for large language models, while yes; we see it across large language models, but also Xeon and inferencing that happens, and in other spots and industries to even on your PC. Your PC can have AI capabilities, and we are about to launch a new era here in a couple of weeks at an event on December 14th. We are excited about showing what Intel can do on that.

Santhosh Viswanathan: It’s frugality, and I think from an Intel point of view, it’s just the same sense of frugality on the AI that will come through, and people will say that everybody is not developing a weather app or doing weather forecasting. So, there is a place where these big, large, super large LLM (Large Language) models will work. There are places where the smaller LLM models will work. They will be very practical for you and me to go and use, and there are places where this is just infringing. I don’t need an LLM model, it’s running somewhere, and I just get the data from their piece.

So, I think Intel’s AI is all about ensuring (utility) for every use case and not just throwing huge hardware at expensive prices. I am going to solve a problem, and what exactly are you trying to do with AI? That whole breadth of spectrum of products is what I think India will do very well on that front, very bullish about where general purpose compute CPUs will be the workhorses of AI in India and then yes, there will be pockets where you need big GPUs and accelerators. But from the client to the server to everywhere else will be general-purpose compute.

Steve Long: I would say we have always had competition. Yeah, we made some proprietary choices in our processor technology manufacturing, which have proven to be the wrong ones, and we’ve taken a couple of steps back. With the adoption of EUV in our process technology and manufacturing, we were the first folks to start with R&D, and then, we decided to go a different direction, which proved wrong.

So, Pat Gelsinger, our CEO, came in two years and change ago, and he put us back on a path of ‘we first have to regain process technology leadership’, because that’s like the foundation of everything that follows. Your products improve because you have process technology leadership and solve your customer’s problems because you have good products foundational to process technology leadership. So, we created this (roadmap) called five nodes and four years, it’s almost an external way of making it very easy for people to see.

Normally, it would take a decade for us to implement five node-by-process technology transitions, we’re doing them in four years because, again, we took two steps back. The way was paved, and there was clarity on what we needed to do so we could accelerate. We are doing that, and our execution is back on track.

Then you look at products. We’re about to launch Meteor Lake in the client space on December 14th; it’ll go public. It’ll be seven months ahead of anything the competition has on Artificial Intelligence devices. This is the third product that has a leadership product in the client space. Our Data Center products, with Sapphire Rapids, came out this year. It was late, but as it has come out, the next generation comes out on December 14th. So, we’re back on the tick-tock cadence expected from Intel. As we say, we’ll let the silicon prove itself out as it gets into the hands of the people. But we think our execution culture and our process technology leadership are well on track so that you can get back to the Intel that you knew and loved and that you should respect.

I would say the mojo is coming back. I mean, like our stock price, if you just look at an indication as measured by investors looking at what we’re doing, our stock prices are up this year as well; it’s over 60 per cent up here to date. Investors are betting with their wallets, and even our customers are. We have gained market segment share in the PC space for three quarters and are on track with that. Our Data Center products, again, people are voting with their wallets on what their choices are. Our products have a TCO advantage in that we have got to get more sampling and customers onboarded in there.

Santhosh Viswanathan: From India’s point of view, I think we are at that point where we are going to be material in the technology world. We have always been the material from a talent point of view. You have not had materiality from a market point of view. For example, Data Centers – we have 20 per cent of the world’s data, but we have 2 per cent of the world’s service. It’s not equitable when you look at it, and why aren’t we the largest data centre market in the world?

The biggest advantage we have is that, as a society, our digitization is unmatched. We take simple things like our biometrics for the answer, but not many countries have biometrics. To go and run AI at a population scale, you have to know your identities, and biometrics is a foundation. We can leapfrog in AI use cases compared to any other society that exists because we have those foundational aspects strong with us, which is digitization.

All of this sets us up to be a huge market from a technology point of view where it’s not just adoption, but it’s also the kind of usages that we go back and drive in our society. The big shifts like Make in India will be those huge propellants because they balance the supply chain. As much as we balance the chip-making supply chain, the electronic supply chain is also important. With the PLI 2.0 and some of those pieces, Intel is very active in enabling that ecosystem and those players to go build products out of India, and that’s the other big opportunity.

Santhosh Viswanathan: Intel has always been that enabler of the ecosystem. An example of India’s first channel partner ecosystem is where people were assembling desktops. Intel had the GID program that enabled it. We were from 1995 onwards that program helped build the Richie streets of the world, the national markets of the world that kind of brought the first computers to this country. So, that ecosystem building is something that we have done over decades. The thirteen thousand employees that are in India are our biggest advantage.

We have the engineering capability. When there’s a good environment that promotes the PLI and the electronics manufacturing industry, and there are enthusiastic partners, that’s the perfect combination for us to go and drive the ecosystem. We have been enabling some of these partners, and it goes into all aspects: how you design your product, how you engineer the first time and get it right, how you test your products, and what the reference products are. So, all of that is what our site engineering capabilities will support these customers.