Can Armās Mobile Lead Translate to AI? Chip Designer Bets on Efficiency
BY
PYMNTS | JANUARY 13, 2025
|
British chip designer
Arm Holdings is eyeing robust opportunities in AI as it strives to maintain its near-total monopoly in mobile devices.
According to Arm Chief Commercial Officer Will Abbey, the company is going full steam ahead into powering artificial intelligence devices ā and sees its power-efficient design as a competitive advantage especially since AI is an infamous power guzzler.
In an interview with PYMNTS, Abbey said Arm chip designs focus on power-efficient, high-performance central processing units (CPUs). This has been Armās hallmark since its inception, leading to its dominance in the mobile device market. Armās chip designs are used by Apple, Nvidia, Google, Microsoft, Amazon, Samsung, Intel, Qualcomm and others.
However, the companyās reach extends far beyond smartphones. Abbey outlined Armās four main business lines: client devices (anything with a display), infrastructure (data centers), IoT (Internet of Things) and autonomy (including electric vehicles and robotics). Across these sectors, Armās designs have shipped in nearly 300 billion devices, with the company delivering about 35 billion devices each quarter.
In 2020, Nvidia offered to
acquire Arm for $40 billion from Japanese tech giant SoftBank in a cash and stock deal to create a chip powerhouse. But British regulators scuttled the deal due to antitrust concerns and the transaction was
terminated two years later. In September 2023, Arm
went public, with SoftBank retaining a 90% stake.
As the AI revolution heated up, Arm executives said in a 2024 earnings call that they see āstrong momentum and tailwinds from all things AIā in complex devices like Nvidia Superchips that combine a GPU and CPU to edge devices like Samsung smartphones.
Armās planned expansion into AI comes at a time when more consumers and businesses are more comfortable using the technology. In fact, PYMNTS
data showsmost business leaders believe generative AI will have positive impacts on workplaces.
Unique Business Model
Armās business model is unique in the semiconductor industry. Rather than manufacturing chips, the company licenses its designs to partners who then produce the actual silicon. Clients can further customize chips with Arm designs. This flexibility lets Arm broadly influence the semiconductor ecosystem, with partnerships spanning over 1,000 companies including major foundries like TSMC, Samsung and Intel.
This approach has allowed Arm to achieve remarkable market penetration. In the mobile sector, Arm boasts a staggering 99% market share. The company is also making significant inroads in other areas, including PCs, automotive applications and data centers.
The companyās market position is growing in other areas as well. Beyond mobile devices, Arm is gaining ground in the PC market, particularly with AI-enabled Windows PCs. In the data center space, where Nvidia leads in AI training, Armās CPUs also play a critical role alongside GPUs. The Grace Hopper super chip exemplifies this partnership, combining Nvidiaās GPU technology with Armās CPU designs.
The automotive sector presents another growth opportunity as vehicles become more electronically sophisticated, with Armās designs found in applications ranging from body sensors to advanced driver assistance systems, according to Abbey.
Arm to Become a Chip Manufacturer?
Asked about rumors that Arm is interested in becoming a chip manufacturer, Abbey declined to comment. Last May, Reuters reported that the SoftBank-controlled company plans to
develop its own AI chips in 2025, with a prototype ready by spring. If Arm does indeed become a chipmaker, it would directly compete with many of its licensees.
When it comes to AI, Arm is leveraging its strengths in CPU design to address the growing demand for AI-capable devices. Abbey said that AI fundamentally relies on matrix multiplications, which CPUs have always been adept at handling. With the latest version of its architecture, Arm has introduced special instructions to make these operations even more efficient, delivering better performance while using less power.
This focus on power efficiency is crucial as AI workloads become more prevalent and energy-intensive. Abbey emphasized the importance of balancing performance with power consumption: āWe as a society are going to have to make informed choices of, ādo we want to keep our lights on, or do we want to keep compute taking place for AI.āā He said Armās approach, which combines high performance with power efficiency, positions the company well to address these challenges.
Armās strategy also relies heavily on its robust software ecosystem. With a community of 20 million developers creating content for Arm-based devices, the company has built a āflywheel effectā that drives adoption across various markets, Abbey said. This ecosystem is particularly important as Arm expands its presence in areas like data centers, where itās competing with established players like Intel and AMD.
When asked about the recent formation of an x86 advisory group by Intel, AMD and other tech giants, which would compete with Armās architecture in AI, Abbey said he viewed the move as an endorsement of Armās long-standing approach to providing choice and flexibility in the market. āWeāre a big believer in standards. Weāre a big believer in choice,ā Abbey said, adding that ācompetition is healthy for the whole of the ecosystem.ā
The x86 chip architecture is the foundation of modern computing in its over four decades of use. Last October, Intel, AMD, Dell, Meta, Lenovo, Google, HP, Microsoft, Oracle, Red Hat, Broadcom and others
came together to form this advisory group to ensure interoperability across hardware and software. Broadcom CEO Hock Tan said the computing industry is at a ācrossroadsā and x86 architectural decisions made today will affect systems for decades.
Looking ahead, Abbey identified the shift of AI processing from cloud to edge devices as a key trend, emphasizing the need to balance performance with power efficiency and security. This transition presents new challenges in protecting personal data while delivering AI capabilities directly on consumer devices.
Arm is focused on continuing to improve the power efficiency and performance of its designs while expanding its software development community. Abbey sees these three elements ā software development, power efficiency and performance ā as critical to ābringing AI to the masses.ā
British chip designer Arm Holdings is eyeing robust opportunities in AI as it strives to maintain its near-total monopoly in mobile devices. According to
www.pymnts.com
