Arm

Arm is a leading compute platform company that designs and licenses processor architectures for a wide range of products, including smartphones, tablets, embedded systems, automotive applications, and data centers.

Executive Vice President and Chief Commercial Officer Will Abbey spoke to us about how the company’s roots in developing solutions for battery-powered mobile devices position them perfectly to be a leader in the evolving world of AI computing.

Can you begin by explaining what Arm does?

Arm designs blueprints for energy-efficient processors that are used in a wide range of electronic devices. We don’t manufacture chips. We create the architecture— the underlying design of how processors work— and we license it to other companies who build their own chips based on our designs. They can also add specific features to suit their needs because our architectures are highly customizable.

What is the connection between Arm and the chips used in mobile devices and more recently with AI?

The connection is simple. It’s about providing the most powerful computing performance with the greatest energy efficiency. And that’s something that is deep within our company DNA.

Back in the early days of mobile, one of the greatest challenges was limited battery life. The engineering choices made during that time were shaped by that low-power source. That's how Arm became so well known for its efficient compute processing and that thread runs through everything that we have done since and everything we do today.

Then, when you look at the world of AI and the huge processing power that’s required, you quickly realize that the ability to deliver high performance with power efficiency is going to give you a great advantage.

A lot of the more interesting AI applications and form factors are defined as being part of edge computing. Does Arm see itself as a pioneer of edge AI?

First we need to define what edge means. The traditional computing paradigm is built on centralized data centers; anything outside the data center is considered an edge device. For a host of reasons— including bandwidth limitations and latency issues— it makes sense to process data at the periphery of the network, close to the originating source.

When you start to look at what's taking place in factories, offices, and homes, you see that more and more devices are becoming smarter and enabled. This need for more intelligence and greater compute in edge devices is only going to increase.

Many of the bigger dreams that people have for AI will depend on small, purpose-built devices that can be carried, worn, or even implanted and will enable us to interact with our environments and each other with greater immediacy and intimacy.

Our designs will be required to facilitate a massive range of new activities, computational needs and application workloads running on a broad range of edge devices.

Are your architectures primarily deployed at the edge?

We started in mobile, but today, we are present everywhere, from edge devices through cloud computing to data centers— some of the biggest global data center providers depend on Arm technology.

Data centers are central to the internet— edge devices still need to send and receive data from the centers— and the more efficiently you can do that, the better. The need for greater power efficiency within data centers is a huge challenge and one that, with our heritage, we are uniquely placed to solve.

So, you have data center compute and edge compute both running AI workloads in our common power-efficient architecture. Software developers can move easily from coding for applications on the edge one day to coding for applications in the data center on the next without needing to change how they work.

Tell us about the relationships Arm has with the developer community.

What sets us apart is our low-power, high-performance DNA. But I also think the community of software developers we work with is a superpower.

Hardware is incomplete by itself; you need the software that people create to run on it. And the more people you have developing for you, the more innovation takes place on your platform, and it becomes a continuous, virtuous circle.

Today, we have 20 million developers creating code within the Arm ecosystem for a diverse range of applications and use cases, giving us direct feedback about what they are doing and what they need, constantly helping us to push forward at pace so we can develop more innovative products to meet their needs and the needs of our end partners. And that combination and relationship with our developers is unique.

What is your vision for AI?

Arm is an enabler in the ecosystem. We enable innovation and progress— ground-breaking technology that makes a difference in the world. We are developing capabilities and platforms that will bring the far-sighted AI future closer, and I am really pleased to be in that space.

I can give you one small but significant example of how I think the future might look in terms of individuals and how they can use technology in a personal way. This is not something that's years away, it’s something that we’re working on right now. Malaria kills millions of children, and one of the reasons why it's so difficult to prevent is that the vaccines must be taken in a specific sequence within a specific time frame. In the developing world, where there are not the same records and systems that we rely on in our health services, and people are often moving around or displaced, it’s incredibly difficult to fully immunize children. Our solution—working with different actors in this space— is a low-cost, low-power biometric sensor that can keep track of the vaccination record. This gives the ability for someone through a simple scan of their finger to access their immunization record as they move from place to place.

That’s just one example of how we can use purpose-built edge devices with AI applications to positively drive health outcomes and make meaningful differences to marginalized people in the developing world.

We make a bold claim that ‘the future of AI is built on Arm.’ And when you think about the great variety of AI-enabled devices that will need to be smarter, faster, and more connected, you can see how our powerful and efficient architecture positions us to be a leader in enabling this next phase of digital.