Arm and the future of AI at the edge

Arm Holdings has positioned itself at the center of the AI ​​transformation. In a wide-ranging podcast interview, Vince Gesaitis, Arm’s head of global government affairs, offered the organization’s decision-makers a look at the company’s international strategy, the evolution of AI as the company sees it, and what lies ahead for the industry.

From cloud to edge

Arm believes that the AI ​​market is about to enter a new phase, moving from cloud processing to edge computing. While much of the media attention so far has focused on massive data centers, with models trained in the cloud and accessed through it, Jesaitis said that most AI computing, especially inference tasks, is likely to become increasingly decentralized.

“The next ‘aha’ moment in AI is when native AI is processed on hardware you could never have imagined before,” Jesaitis said. These devices range from smartphones and earphones to automobiles and industrial sensors. Arm’s IP is already embedded, literally, in these devices — it’s a company that just last year was the IP behind more than 30 billion chips, placed in devices of every conceivable description, around the world.

Deploying AI in edge environments has many benefits, with the Arm team citing three key “wins.” First, the inherent efficiency of low-power Arm chips means that the energy bills for running computing and cooling are lower. This keeps the environmental footprint of the technology as small as possible.

Second, placing the AI ​​locally means that latency is much lower (with latency determined by the distance between local processes and the location of the AI ​​model). ARM refers to uses such as simultaneous interpretation, dynamic scheduling of control systems, and features such as near-instantaneous operation of safety functions – for example in industrial IoT settings.

Third, “kept local” means that there is no potentially sensitive data sent off-premises. The benefits are clear for any organization in highly regulated industries, but the increasing number of data breaches means that even companies working with relatively benign data sets are looking to reduce their attack surface.

The silicon arm, optimized for devices with limited power, makes it well-suited for computing where it’s needed on Earth, the company says. The future may be one in which AI is integrated into all environments, rather than centralized in a data center run by one of the large providers.

arm and global governments

Arm actively engages with global policymakers, seeing this level of engagement as an important part of its role. Governments continue to compete to attract investment in semiconductors, and supply chain issues and concentrated dependencies remain fresh in the memories of many policymakers from the time of the Covid pandemic.

Arming lobbyists for workforce development, and currently working with policymakers in the White House on an education coalition to build an “AI-ready workforce.” Local independence in technology depends as much on the capabilities of the workforce as on the availability of hardware.

Gesaitis pointed out that there is a difference between regulatory environments: the United States prioritizes what the government there calls acceleration and innovation, while the European Union takes the lead in terms of safety, privacy, security and legally enforced standards of practice. Arm aims to find a middle ground between these approaches, building products that meet stringent global compliance needs, while promoting advances in the AI ​​industry.

The enterprise case for Edge AI

The case for incorporating Arm’s edge-focused AI architecture into enterprise transformation strategies can be compelling. The company emphasizes its ability to deliver scalable AI without having to focus on the cloud, and is also pushing its investments in hardware-level security. This means that issues such as memory exploitation (beyond the control of users connected to centralized AI models) can be avoided.

Of course, sectors that are already heavily regulated regarding data practices are unlikely to see relaxed governance in the future – and the opposite is largely inevitable. All industries will see greater regulation and greater penalties for non-compliance in the coming years. However, to balance, there are significant competitive advantages available to those who can demonstrate the safety and security inherent in their systems. It is in this regulatory landscape that Arm sees itself as a fit for homegrown AI.

In addition, in Europe and Scandinavia, environmental, social and governance (ESG) objectives will be increasingly important. Here, the power-hungry nature of Arm chips offers significant advantages. This is a trend that even US hyperscalers are responding to: AWS’s latest SHALAR suite of low-cost, low-power Arm-based platforms is there to meet this specific demand.

Arm’s collaborations with cloud scalers like AWS and Microsoft produce chips that combine the efficiency and power needed for AI applications, the company says.

What’s next from the arm and the industry

Jesaitis pointed to several trends that companies may see in the next 12 to 18 months. Global AI exports, especially from the United States and the Middle East, ensure that domestic demand for AI can be met by major service providers. Arm is a company that can provide leading providers in these contexts (as part of its suite of offerings) and meet the growing demand for edge-based AI.

Jesaitis also sees advanced AI as a champion of sustainability in an industry increasingly criticized for its environmental impact. Since the largest market for Arm’s technology has been in low-power mobile computing, it is inherently “greener.” As organizations hope to achieve energy goals without sacrificing compute, Arm offers a way to combine performance and accountability.

Redefining “smart”

Arm’s vision for AI at the edge means that computers and the software running on them can be context-aware, cheap to run, secure by design, and — thanks to near-zero network latency — responsive. “We used to call things ‘smart’ because they were connected to the Internet,” Gesaitis said. “Now, they’re going to be really smart.”

(Image source: “Factory Floor” by danielfoster437 licensed under CC BY-NC-SA 2.0.)

Want to learn more about AI and Big Data from industry leaders? Check out the Artificial Intelligence and Big Data Expo taking place in Amsterdam, California and London. This comprehensive event is part of TechEx and is co-located with other leading technology events. Click here for more information.

AI News is powered by TechForge Media. Explore other enterprise technology events and webinars here.