2022 Arm Ecosystem Report

Survey of Arm ecosystem covering artificial intelligence, automotive, 5G, mobile, IoT

2022 Arm Ecosystem Priorities and Perspectives

Foreword

Rene Haas, CEO, Arm

Rene Haas
CEO, Arm

Introduction

Last year, we released the first Arm Ecosystem Predictions and Perspectives report. As you might expect, the report was dominated by the uncertainty created by the COVID-19 pandemic and its effect on how we live, work, and communicate.

Yet as the world begins to open up again, we’re looking ahead at what 2022 and beyond might bring and the trends shaping the industry.

One thing became clear to me last year; it's very difficult to predict the future, and that's coming from a relative position of strength, given our conversations with technology innovators range out 5-10 years. With this mind, we've taken a slightly different approach to the report for 2022: Focusing on our ecosystem's priorities, since we can interpret those for you with real clarity.

We surveyed more than 900 technologists in the Arm ecosystem, and respondents are innovating on countless fronts, from smart phones and watches to edge servers and HPC systems to robotics and autonomous vehicles. And when we look at responses from those who say Arm is a fundamental part of their work, we are able to see the full breadth of focus across the ecosystem.

Clearly, Arm is everywhere. Combine that with the proliferation of artificial intelligence (AI) across many applications, and Arm technology clearly is the future of AI.

^{Of particular interest to me is how AI and machine learning (ML), more robust security, and the drive to push more compute to the edge and endpoints factor into the future as envisioned by the Arm ecosystem."}

Arm believes that 100 percent of digitally shared data will be securely processed at the endpoint, in the networks or cloud, and that is happening because of the pervasive and accelerating deployment of our technology, as well as our focus on specialized processing solutions: The flexibility to deploy CPUs, GPUs and NPUs, Arm Cortex-A-class, R- or M-class devices in a mix-and-match fashion to help developers customize their products.

That freedom of design doesn’t really exist outside the Arm ecosystem.

• One-quarter of respondents are using Arm technology to build edge server / gateway solutions
• 1 in 5 software developers are using Arm to develop ADAS car safety technology
• 25% of hardware developers are using Arm to design health wearables

As part of this report, we’ve identified several trends. Of particular interest to me is how AI and machine learning (ML), more robust security, and the drive to push more compute to the edge and endpoints factor into the future as envisioned by the Arm ecosystem. At Arm, we’ve relentlessly innovated IP, tools, and ecosystem support to help drive some of these trends. This year, we introduced new features in the Armv9 architecture that laid out a path toward more robust device-level security with confidential compute and demonstrated how specialized processing can help teams customize their innovations, quickly and flexibly. Part of that relentless focus to make developers’ lives easier includes things such as PSA Certified, the standards-based security certification framework and organization, and SystemReady, the compliance certification program that ensures Arm-based servers, infrastructure edge, and embedded IoT systems are designed to specific requirements, enabling generic off-the-shelf operating systems to ‘just work.’

In this report, you’ll also find insight and analysis in these areas from our senior Arm leaders: Dipti Vachani, SVP and GM of the Automotive group, Mohamed Awad, SVP and GM of our Infrastructure group, Chris Bergey, whose Client business works with partners across the mobile computing realm, and Paul Williamson, SVP and GM of our IoT business.

I hope you enjoy reading the following pages.

The Edge: Speed of Insight Is Key

Overview

The cloud compute model is being transformed as vast new sources of data are being captured from the analog world. This requires deploying the right kind of efficient compute solutions along the compute continuum from cloud to edge to endpoint.

Here, speed is key. A third of respondents said the increased speed of insight gained from reduced latency was the primary benefit of moving compute from the cloud to edge servers and gateways. And that speed is what distributed compute, enabled by heterogeneous processing, will deliver.

In the survey, this notion of distributed computing is seen as a gateway to a more sustainable world. When isolated by job role, one in five C-suite respondents thought the primary benefit of moving from cloud to edge was a reduced carbon footprint, an opinion shared by only 6% of hardware and software developers.

Fortunately, specialized compute solutions are delivering 30% and 40% performance improvements generation after generation in the infrastructure, which helps reduce the carbon footprint of computing across the computing spectrum.

Mohamed Awad, SVP and General Manager, Infrastructure Line of Business

At the Edge, Insights are key
drivers for new business models

The edge is coming into its own.

A growing number of customers are beginning to understand the opportunities created by moving computing assets closer to sensors and end devices. A year ago, respondents to the forecast said cost savings achieved through digital infrastructure investments would have a greater impact than insights it would help them obtain into their operations. This year, insights loom far larger with over 1/3 selecting speed of insight as the most important driver of the edge.

Retailers are looking at ways of analyzing camera data to better understand consumer preferences. Logistics companies are launching pilots to capture data from ships and rail cars to see if it can help them reduce fuel consumption or repair costs. When you consider the vast amount of data involved in some of these applications and the speed in which decisions need to be made, performing inference and other tasks as close as possible to the end user becomes more compelling.

However, the road to edge nirvana isn’t straightforward. To ensure compatibility and accelerate adoption, we need robust frameworks and standards initiatives like Project Cassini. Security is also a major concern: the edge will dramatically expand the surface area for attacks. Similarly, 5G infrastructure will be needed to link these new edge assets to consumers. But this year’s results show that the momentum is going in the right direction and growing rapidly.

And while there’s always increasing pressure to shrink development cycles, the complexity of today’s emerging technologies can affect these cycles. Of those surveyed working in IoT, fewer than 1 in 10 software developers see achieving the right level of processing power as a challenge.

IoT: Security & Development are Top Challenges

Overview

While executives worry about lengthy and complex software and product development cycles, IoT software developers are more concerned about ensuring security, according to our survey. Security (38%) is the second-highest challenge facing respondents who target IoT industries.

Fortunately, the industry has linked arms in recent years to embrace holistic approaches to end-to-end security. This ranges from rapidly expanding design-for security frameworks and certification efforts such as PSA Certified, to confidential computing innovations at the device level that take the concept of hardware compartmentalization to new, more secure levels.

Paul Williamson,
SVP and GM IoT & Embedded

Simplifying development to enable scale

What jumps out here is that everyone is concerned about lengthy and complex software development cycles. Indeed, where the software rubber meets the hardware road, complexity is developers’ second-biggest concern behind security. Software developers are often required to deliver in a number of areas, so time to completion is critical for them and complexity doesn’t help.

Another major concern rooted in complexity is the porting of software. Executives in particular worry that portability issues can mean sticking with older technology they’ve deployed or accepting a delay in the project to gain access to newer technology.

While security is a concern for all, software developers worry the most. Likely, this is because many of the highest-risk attacks enter through software, rather than hardware.

Lower on the list of concerns is that today’s developers and engineers have vastly more choice in processing performance, efficiency, and heterogeneity than ever. Respondents’ previous concerns, it appears, have refocused into other potential challenges.

I hear loud and clear the 25% of c-suite respondents who name cost savings as having a greater impact than any other factor: lowering the cost of developing on Arm silicon is imperative.

Software developers, too, understand the significant benefit of CI-CD in accelerating time to revenue and cutting the costs associated with long, drawn-out development cycles.

However, it’s evident that virtual platforms have yet to demonstrate their value to those most likely to benefit. Only 1 in 10 software developers recognize the great impact they will have in the coming years.

• 1 in 5 software developers and system architects say CI-CD will have the greatest positive impact on their success in the next 5 years.

• In the IoT consumer technology market this rises to 43%.

• 1 in 4 C-suite respondents say cutting costs through simplification and automation will have the greatest impact.

Running at speeds comparable to the real hardware, Arm’s Fixed Virtual platforms are complete simulations of an Arm system, including processor, memory, and peripherals. With FVPs, software developers can work on application, firmware, and early driver development far ahead of hardware availability.

This way, the whole ecosystem, including OEMs, can start to develop long before the silicon is available – resulting in fairer distribution of economic value, faster time to revenue, and greater return on investment for everyone. For SIPs, that means stronger demand for chips before they’re even taped out.

In the next five years, I strongly believe we’ll be able to demonstrate just how powerful being able to develop on Fixed Virtual Platforms (FVPs) can be in accelerating development on Arm silicon.

5G: Tapping Massive Opportunity

Overview

There’s a difference of opinion when it comes to what’s driving 5G adoption. Over 25% of developers think mobile devices will lead the adoption of 5G, nearly double the number of executives, who see smart cities and the shift from the cloud to edge as bigger drivers than mobile. Respondents agree that mobile devices will be the first place that 5G data throughput will start to make a difference but beyond that, people are unsure where the use cases will show first.

Regardless, faster, low-latency, high-bandwidth 5G networks can connect anything, anywhere, with improved performance, efficiency, and cost. Remote surgery, autonomous vehicles tied into road infrastructures, new applications for streaming technologies—the sky’s the limit.

As 5G continues to evolve, Arm-powered solutions help more companies take advantage of its potential—from the compute infrastructure through to the network and device. The lightning-fast data speeds, super-low latencies, and more powerful and intelligent network services of 5G, unlock a new era with unbound potential for connected consumer devices and IoT.

Chris Bergey, SVP, General Manager,
Client Line of Business

Mobile set to be lead 5G driver

Mobile devices are the strongest driver for 5G technology, especially among software and hardware developers.

The smartphone is the crucible of 5G innovation, with developers benefiting from more reliable, higher performance network connectivity, alongside a scalable software development platform. This presents phenomenal opportunities for the ecosystem, with more immersive video, mobile gaming and xR experiences, and new innovative features and tech coming to the billions of 5G-ready smartphones

Thanks to 5G, new user experiences, such as augmented reality (AR), will significantly develop and advance, defining 5G innovation in the mobile space. Developers will then be able to translate their innovations to new markets with a common platform. Not just driving cool consumer trends and experiences, but also creating brand new business and economic opportunities.

Mohamed Awad, SVP and General Manager, Infrastructure Line of Business

5G: It's happening and it's going to be everywhere

The debate on 5G has changed. People are no longer asking “do I need 5G?” They are asking “where am I going to use it first?” Worldwide 5G connections are expected to pass 600 million this year with services available in over 84,000 locations.

Most of us will likely first experience 5G through mobile devices. Over 25% of hardware and software developers thought mobile devices would lead 5G adoption of in our survey. But take a look at the view from the executive suite. C-level execs and VPs said that smart cities and the shift from the cloud to the edge will be more prominent drivers. They also saw a strong opportunity in digital health. If you think of hardware and software developers mostly concerned with projects a year or less out, and VPs focusing on three to five years out, you can begin to see how 5G is an expanding opportunity.

We are also seeing growing interest in private 5G networks at factories and industrial sites, which are often populated with systems that have never been connected to networks. Private 5G equipment, software and services will be a $75 billion market by 2026, according to research firm, ABI Research. But, like the edge, it will take work. Cost, power consumption, compatibility, and security are all formidable challenges we need to overcome.

Automotive: The era of the software-defined vehicle dawns

Overview

When we asked those who said their primary market was automotive which trends and technologies they thought would most impact the industry in the next five years, two clear themes emerged.

Interestingly, 61% said the evolving functional safety standards and requirements would have the greatest impact. C-suite and developers were largely agreed on this point.

Two-thirds of developers said the move toward the software defined vehicle would have the biggest impact on the industry – though only 44% of respondents in executive roles shared this view.

Dipti Vachani, Senior VP & General Manager, Automotive Line of Business

Safety, software defined vehicles lead auto trends

While executive and developer opinions seem generally aligned across the automotive technology industry, there’s a significant difference in how the management of design and verification challenges through digital twins is perceived to impact the automotive industry.

Fully self-driving cars may be a little way off but increased levels of assistance coupled with applications that have restricted operational design domains see us making real progress. To enable this new range of capabilities, vehicle electric architecture will undergo challenging evolutions as disparate ECUs are merged into domain controllers.

A digital twin is a “live” digital recreation of these domain controllers within an individual car on the road. This enables developers to test out ADAS, autonomous, and other future software workloads.

The importance of this to the industry really seems to be a question of differing mindsets. To a hardware or a software developer right now it’s a tool to help them get their job done and perhaps identify flaws and bugs.

Most common use of digital twin?

Predictive maintenance
Faster design
Performance analysis

To an executive, though, it’s the pathway to new revenue streams: real-time and continuous diagnostics, predictive maintenance, and so on.

Two-thirds of software developers recognize the profound impact that the software defined vehicle (SDV) will have on their work and the industry as a whole. What’s perhaps less evident at this stage is how important digital twins will be in enabling developers to deploy solutions faster.

Mobile & Infrastructure: Efficient sustainable development

Overview

In a recent blog on Arm Blueprint, Arm CEO Simon Segars wrote:

^{Progress cannot come at the expense of the environment.

To avoid catastrophic climate change, the technology industry must focus on decarbonizing compute – reducing energy consumption and lowering emissions wherever compute happens."}

Simon Segars, CEO, Arm

The majority of respondents agreed: the tech sector should find way to continue to innovate while simultaneously reducing its carbon footprint.
However, while few believed in innovation whatever the cost, CxOs and VPs were most likely to put innovation before reducing their carbon footprint.

Meanwhile, 1 in 5 developers felt that the industry should reduce its carbon footprint, even if that means imposing limits on technological progress.

Chris Bergey,
SVP & GM, Arm Client Line of Business

Improving efficiency to help scale mobile

It’s encouraging that over two-thirds of hardware developers focused on mobile computing believe that the industry must find ways to reduce its carbon footprint while innovating for the future. We need to confront the realities of environmental impact of our industry.

Compute performance in smartphones has grown exponentially in recent years, as new iterations of Arm technology enable ever-greater performance within a limited thermal envelope and needs for increased on-time without recharging.

It means that smartphones designed and built three or four years ago remain highly capable devices today. The average lifespan of a smartphone will increase from 2.58 years in 2014 to 3.77 years in 2024, according to market research platform Statista.

By building the very best devices with the highest possible performance, we can ensure that each device offers the latest apps and services for longer, as more apps and services are added year-on-year.

We will still see premium devices launched each year and selling prices remaining strong and even strengthening, and we are already seeing devices being used for longer, passed among families, and resold. But we also need to do more to bridge the digital divide in low-income economies, which have until now relied on feature phones rather than smartphones.

Mohamed Awad, SVP and General Manager, Infrastructure Line of Business

Innovation should not come at the planet’s cost

Achieving more computing performance for every watt of power has long been a key driver for Arm. Datacenter workloads and internet traffic are nearly doubling every two years. We need to improve performance per watt at the same rate or better to keep computing from increasing its carbon footprint.

But the benefits of continually improving energy efficiency go even further. Greater performance per watt leads to lower costs for customers and a better total cost of ownership (TCO) for carriers and datacenter owners. Cloudflare, a cloud service firm whose technology sits behind 1/6th of the world’s websites, recently reported that Arm Neoverse allows them to process 57% more requests per watt.

It’s encouraging to see that 51% of respondents working on datacenter deployments said the tech sector should find ways to innovate while reducing its carbon footprint. At the same time, 29% of datacenter respondents say they won’t sacrifice innovation for sustainability.

My message to them is simple:

You can have both."

Neoverse N2, which is based on the Armv9 architecture, for instance, will provide a 40% uplift over Neoverse N1 while new features such as Scalable Vector Extension version 2 and Memory Tagging Extension. And we predict that by 2030, hyperscalers could increase their compute capacity by 25% at the same power levels by switching to Arm.

AI: Powerful promise for healthcare

Overview

A world made better by artificial intelligence (AI) has been one of humanity’s longest-held technological visions, and in 2021 that vision is being realized.

AI is in our pockets, our homes, our cars, our workplaces, and it’s already making a real difference in how we experience the world around us.

A quarter of all respondents stated that the greatest impact of AI would be in healthcare. C-suite respondents were even more likely to view healthcare (35%) as having the greatest impact.

Healthcare and many other industries taking advantage of AI are no doubt yielding the benefits of powerful and power-efficient new hardware systems that are enabled by specialized compute.

No longer are AI developers forced into a one-size-fits-all hardware approach. They’re increasingly mixing and matching CPUs, GPUs and NPUs that best deliver their AI-based innovation visions. The ability to run AI from endpoint to cloud, on the billions of Arm devices in the market, creates a tremendous opportunity for new and innovative applications.

Paul Williamson, SVP and GM IoT & Embedded

AI’s Promise Seen in Healthcare

I’m not surprised that after the events of the past 18 months, healthcare is seen as a key area in which AI will have the most impact. AI technology is already enabling healthcare professionals to diagnose and treat patients far more quickly thanks to its ability to spot patterns in data.

During the pandemic, it enabled companies like DarwinAI to rapidly deploy initiatives to help in the fight against COVID-19. DarwinAI built COVID-Net, a deep neural network (DNN) using the Arm NN inference engine to identify critical COVID-19 symptoms in chest X-rays.

It’s also very interesting to see how much work is going into data center and storage technology within healthcare. Healthcare is going to be a big driver in data: Genome data alone is doubling every eight months. Before healthcare introduces AI into therapies, they will have to figure out the back end.

The healthcare market is still dominated by health-wearables development, but there’s emerging focus in other application areas that can be enabled through AI. For example, of the respondents developing healthcare technology, almost 1 in 5 will be working on AI-powered smart camera/computer-vision products in the next three years. These products might detect anomalies in live scans, monitor and analyze patient behavior or use facial recognition to ensure that everyone in a hospital is supposed to be there.

With such applications comes an intensifying focus on data consumption, analysis and security. For instance, as applications leverage more AI to make sense of these data, developers need to think more broadly about how to properly leverage compute from the cloud to the edge and endpoints. This means understanding how to put the right compute capabilities along that spectrum for optimal performance and time-to-results. It also means understanding how to think about end-to-end security that ensures patient privacy and trust as well as regulatory compliance.

Survey Methodology

This report is based on the views of more than 900 respondents in varying roles, regions and focuses across the Arm ecosystem and beyond—from CEOs of small and medium-sized enterprises to hardware developers in multinationals, and from OEMs to chipmakers and foundries.

Insight was gathered via a survey conducted by Pulse, a Gartner company. Respondents were split between Arm email subscribers (52%) and Pulse’s own community of technology executives (48%).

Job role

• 59% of respondents were CEO/VP level
• 63% of hardware and software developers stated that Arm technology was fundamental to their work
• 68% of respondents are based in North America
• 29% work for large companies of 10,000 employees or more