NVIDIA plans to pump more millions into Cambridge
US technology giant NVIDIA believes its investment in the Cambridge-1 supercomputer initiative could top $100 million – and that’s just the starting point.
The company, which is bidding to acquire Cambridge chip architect Arm for $40 billion, is plotting many more initiatives that will fire the local and national economies if the UK and other governments sanction the takeover deal.
NVIDIA president and CEO Jensen Huang joined Arm CEO Simon Segars in an online summit to unveil a bold and lucrative agenda that would ramp Cambridge’s technology capability several notches – bringing untold benefits to the Cluster.
Quizzed about the potential level of investment by NVIDIA in Cambridge, Huang confided: “I think we have something really, really special as a starting point. I recently announced a supercomputer – the largest supercomputer in the UK – called Cambridge-1. And the excitement is fantastic.
“In terms of the Cambridge-1 supercomputing centre, call it $100 million – just as a starting point. It's a big investment. It is the most powerful supercomputer in the UK and researchers are super excited about it.
“I am so enamoured with Cambridge and the engineering team there, and we have every intention to invest more.
“The way people want to use computers is different; the way people want to design computers is different. And the thing that I really love – what Arm has enabled us to do – is design specialised types of computers.
“I happen to believe that this open way of allowing the industry and enabling the industry to create bespoke and different types of computers in this diverse world is exactly the way you want to go about it. Now what we could do is just give it a whole bunch more speed and scale.
“In the UK, in Cambridge, we have one of the world's premier microprocessor and IP development centres. There are no equals. And it is one of the reasons why the CPU cores that have been made by Arm over the years have become established in the way they have.
“We have every intention to not only continue to do this there but to invest in doing it more there. There's something that is really special in Cambridge; it is the birthplace of computers and genomics.
“And we have a lot of partners in the Cambridge area which is at the intersection of computational biology, computational genomics and computer science.
“Our opportunity here in Cambridge is to not only grow that HQ but also to bring NVIDIA’s ability to invest in AI research and robotics research and computational biology and all of the capabilities we have, to help the UK become a centre of world-class AI development.”
Segars said the alliance with NVIDIA is poised to revolutionise future iterations of technology. He said: “When I think about the way the end market competes, it's about creating products, it’s doing new things. It's innovating. And really when I think about our business over the last 30 years, what we've done is create these building blocks that are really useful in all these chips, which saves people time.
“It saves people from having to reinvent the wheel; it means that they can apply their R & D resources on things which are really differentiating. And I think the more we can do that, the more we can put these tools in people's hands where they can then innovate around it, build on top, the better.
“So through this combination, we're going to create more stuff; we're going to be able to look at more markets at the same time, go deeper, go further up the stack, really understand the end applications to the world that NVIDIA has done over the years, and create a richer portfolio of IP to companies who are then going to be in a better place to innovate on top of a bigger platform and create even more competitive products in the market. This is only good for competition.”
Discussing the potential Arm acquisition specifically, Huang said he could only see positives. “For the vast majority of people I've spoken to, the transaction is super welcome. And they're super excited about it.
“One of the things that that they love is the idea that NVIDIA’s leadership in artificial intelligence and the software and the platform technologies could be brought out to the far edges of the markets where Arm is today.
“Arm has great presence. That's where the IoT is going to happen. That's where the edge AI is going to happen. There's a lot of wonderful things that could happen out there. But the fundamental technologies of artificial intelligence aren't out there yet.
“Simon and I both recognise that going into the future, independence doesn't equate to strength or to vibrancy of the ecosystem. To bring computing out to the edge and to the new places that we want to go, having a platform stack, having software capability, having an understanding and reach into the ecosystem, working with the developers in the ecosystem – that capability is richly valued. We have an opportunity to bring that together.
“I think there's a fallacy in equating independence with the strength and vibrancy of Arm, which ultimately is what the customers of Arm want. They want not just an open Arm, an independent Arm, but a strong Arm and a strong Arm that allows them to go into some wonderful new markets.”
Segars said Arm and NVIDIA were taking a long view of market potential to map out potential progress.
He said: “I think it really is important to look at the future when thinking about what we're going to be able to do – because the complexity of the systems that our licensees are building is going up and up and up.
“You know, once upon a time, our roadmap was 7 and 9 and 11. And they were fundamentally designed around mobile. And if you had an application that kind of looked like that, we had a product for you, we were able to create a microcontroller family – relatively low cost investment for us, didn't need a big software stack around it – it's been massively successful.
“There's about a billion of them a month that get sold. But they're relatively simple devices in terms of the software and the complexity – everything that sits around them.
“It's a different world in the future than it has been in the past. And we need a different approach to go and address that and keep providing all these technologies that our licensees are asking for. And that's what we're uniquely going to be able to do.”
NVIDIA is basing the Cambridge-1 supercomputer at Kao Data Centre in Harlow and Jensen believes this combination of resources and expertise is powerful.
He said: “The concept of a data centre is a computer infrastructure that is secure, powerful and that has the ability to support multiple tenants running multiple things remotely.
“This data centre could be a very large thing – it could be the size of a football field; it could be a supercomputing centre the size of a basketball court; it could be an enterprise data centre twice the size of that.
“It could be a 5G base station; it could be a retail store, it could be a factory, it could be a broom closet. It could be all the way out on the edge, on four wheels, because it's inside a car.
“All of these computers in the future are going to be very data centre-like; they're going to be secure, they're gonna be cloud native; they're gonna be multi-tenant, they're going to be able to run AI. You will manage these data centres from afar and data centres will come in all kinds of different sizes and shapes.
“The one thing for sure is we're gonna try to make a contribution in servicing and enabling data centres all over the world. What's going to be novel is that the software stack they run on top of it is going to be completely different.
“Some of it will run VMware, some will run Linux, some of it will run Kubernetes on Linux, and there are all kinds of different types of software stacks that you want to run.”
So will there be more Arm technology in the data centres of the future?
Segars said: “That's the plan. What Jensen has described is a massive diversity of what a data centre is – requiring different solutions. Our business model is all about enabling people to build different solutions around common standards. So you get a lot of software reuse. And so it's easy to build optimised solutions for the kind of problem that you're looking to solve.
“But it's an enormous space. And it's going to take a long, long journey and a lot of investment to really fulfil the potential that there is in the datacenter.”
Jensen said Arm was so popular because it gave customers the ability to build their own computer and, by extension, their own computer industry.
“However, building a vibrant computer industry takes more than Arm,” he added. “It is the enabler, but it takes more than that. So in the combination of our companies we’ll be able to provide an open CPU ISA, for license with all kinds of different sizes and shapes and scales. And together, we're going to build an ecosystem around the ARM CPU so that people all over the world could decide what kind of shapes and sizes of computers they would like to make.
“There are supercomputers in France being built with Arm; supercomputers in Japan being built with Arm; supercomputers in China being built with Arm. There are supercomputers in the United States – we're building one with Arm – and we're building one in the UK with Arm.
“So there's all kinds of different systems and sizes and architectures that you could create because of the open licensing model; because of the openness of it.
“And we should remember: you want open but you want strong open. Nobody wants to build a weak computer. Nobody wants to build an inadequate computer.
“We've been working with Arm for a long time and have a great working relationship with Arm already. We license the architecture.
“Nvidia is doing well; Nvidia has a great strategy ahead. We don't have to buy Arm. We want to buy Arm.
“You know, this combination is going to be great for the ecosystem. It's going to expand the reach of Nvidia's ecosystem, although these markets are rather new to us – mobile and embedded and others. These markets are going to benefit from AI.”
Segars concluded: “There are people that say, ‘Oh, we should stay independent.’ But I think this is a much better outcome. The definition of “everything” is getting bigger and bigger and bigger. And I want to address everything.”