The Arm A-Z in 2025: Powering AI everywhere

2025 was the year Arm compute went everywhere AI needed to be, from cloud to car, data center to drone. With standout product launches, high-profile collaborations, and bold new use cases, Arm and our industry-leading ecosystem helped to shape the future of intelligent computing. Here’s a look back at the year’s milestones, from A to Z.

A is for the Aston Martin Aramco Formula One™ Team, with the Arm partnership showcasing how data and compute, powered by the Arm compute platform, drive performance, from wind tunnel simulations to split-second pit-lane decisions, while also offering extensive support to the career of Jessica Hawkins, who’s head of Aston Martin Aramco’s F1® Academy, as documented in DRIVEN.

B is for Branding, with Arm introducing new product naming architecture to better communicate our platforms externally, including new Arm Compute Subsystems (CSS) brand names across every key market and simplifying IP numbering by aligning it with platform generations and performance tiers. 

C is for Chiplets, with a series of Arm-led announcements, initiatives and partnerships designed to scale chiplet innovation, including the first specification of the Arm Chiplet System Architecture made publicly available, and the Arm-imec partnership to accelerate chiplet adoption across the automotive industry.  

D is for Developer innovation, with Arm supporting a global ecosystem of over 22 million developers building on the Arm compute platform across cloud and edge environments. 2025 initiatives from Arm included:  

• Support for ExecuTorch 1.0 GA release to redefine the edge AI developer experience, and;  

• Launches of a new Cloud Migration Assistant Custom Agent and broader cloud developer initiative to help developers automate, optimize and accelerate the shift of workloads to Arm.    

E is for Edge AI, with Arm introducing the world’s first Armv9 Edge AI platform – featuring the Arm Cortex‑A320 CPU and Ethos‑U85 NPU – to enable AI models over one billion parameters and deliver cloud‑scale edge intelligence for next-gen IoT, smart cameras, and industrial automation for partners including Advantech, AWS, and Siemens.  

F is for Fast Company Recognition, with Arm ranked #7 on Fast Company’s 2025 Most Innovative Companies in AI list.  

G is for (next-generation) Graphics, with Arm delivering an industry-first through Arm neural technology, which will add dedicated neural accelerators to 2026 Arm GPUs, bringing PC-quality, AI-powered graphics to mobile for the first time.  

H is for Hyperscalers, with half of the compute shipped to top hyperscale cloud providers in 2025 being Arm‑based, as leading cloud hyperscalers like AWS (Graviton5) and Microsoft Azure (Cobalt 200) announced new generations of cloud processors built on Arm, while new Google Cloud’s new Axion-based N4A VMs and C4A metal, powered by Arm Neoverse, propelled cloud-to-car development.   

I is for Integration, with Arm Accuracy Super Resolution (ASR) integrated into Fortnite Mobile to bring console‑level graphics and enhanced performance to the mobile game – this followed Arm ASR, an open‑source upscaling technology, being made publicly available to game developers at the 2025 Game Developer Conference (GDC). 

J is for Jetson Thor, with NVIDIA and Arm enabling intelligent solutions for next-generation robotics, autonomous systems and vehicles through NVIDIA Jetson Thor and DRIVE AGX Thor that are both built on Arm Neoverse V3AE CPUs.  

K is for Kompanio Ultra, MediaTek’s system-on-chip (SoC), featuring Armv9 CPUs and Arm Immortalis-G925 GPU, for the latest Chromebook Plus devices, including the Lenovo Chromebook Plus 14″ and Acer Chromebook Plus Spin 514, for industry-leading performance, efficiency, multimedia and AI capabilities. 

L is for the Arm Lumex CSS Platform, a purpose-built platform to accelerate AI experiences on flagship smartphones and next-gen PCs, with Lumex uniting Arm’s highest performing CPUs with Scalable Matrix Extension version 2 (SME2), GPUs and system IP to deliver performance, efficiency, and time-to-market benefits for the mobile ecosystem, including the latest flagship smartphones, like the vivo X300 and Oppo Find X9 series, and leading apps, like Alibaba Taobao, Alipay, and Google apps, that are benefiting from SME2 AI acceleration.  

M is for Meta, with the Arm-Meta strategic partnership scaling AI efficiency across every layer of compute, from megawatt-scale data centers training the world’s most advanced AI models to milliwatt-scale edge devices powering on-device intelligence, like Meta’s Arm-based Ray-Ban smart glasses.

N is for NVLink Fusion, with NVIDIA NVLink Fusion now integrated into Arm Neoverse, delivering a high-bandwidth, coherent CPU-GPU interconnect that lets partners build and scale unified, rack-scale systems.

O is for Open Compute Project (OCP), with Arm appointed to organization’s the Board of Directors, underscoring our leadership role in defining open standards in converged AI data centers, while also contributing vendor-neutral Foundation Chiplet System Architecture specification to the OCP to drive and accelerate chiplet innovation across automotive and data center markets.  

P is for Policy, with a joint paper from Arm and Special Competitive Studies Project (SCSP) outlining how policy and innovation can accelerate AI energy efficiency, and power ongoing U.S. competitiveness in AI, while Arm also joined over 100 organizations in signing the White House AI Pledge that commits to significant investments in AI education. 

Q is for Quarterly Earnings, with Arm surpassing $1 billion in revenue for three consecutive quarters in Q4 FYE25, Q1 FYE26 and Q2 FYE26 – a first in the company’s history – with the milestone reflecting significant demand for Arm‑based compute across cloud, edge and physical AI.  

R is for Reports, with Arm publishing landmark reports that provide a range of thought leadership and technical insights into the new eras of AI and silicon design including:  

• The Arm AI Readiness Index, a comprehensive analysis of global AI readiness and implementation from enterprises worldwide; and  

• Silicon Reimagined: New Foundations for the Age of AI, which examines how silicon technologies are evolving to meet the unprecedented computational demands of AI. 

S is for Software-driven intelligence, with Arm powering the Geely EX5 to bring real-time safety, seamless screen integration, natural‑language voice interaction, and intelligent park assist to the vehicle, with Geely entering an international market (the UK) for the first time. 

T is for TIME 100 Recognition with CEO Rene Haas and Arm both honored by TIME for shaping the modern compute landscape and future of AI.

U is for Universal Robots, a company delivering intelligent automation in factories and warehouses worldwide through its Arm-powered robotics, with Arm also enabling a variety of partners to build advanced robotics products, including Deep Robotics, R2C2, and Under Control Robotics.

V is for Ved Chirayath’s Ocean Mapping, which uses Arm-powered compute to miniaturize and accelerate real-time, AI-enhanced underwater imaging tools to support ocean conservation, with the explorer’s work part of an Arm National Geographic campaign exploring unique individuals using Arm-based technology to do amazing work, like filmmaker and photographer Frederique Olivier.

W is for (AI) Workstations, with the launch of NVIDIA DGX Spark, a compact Arm-based AI supercomputer, providing developers, researchers and enterprises up to a petaflop of AI performance and support for models up to 200 billion parameters entirely on-device – this led to a range of DGX Spark Workstations being launched by leading OEMs towards the end of 2025. 

X is for X improvements, with the new SME2-enabled Arm C1 CPU cluster achieving up to 5x AI speed-up and 3x improved efficiency compared to the previous generation CPU cluster, leading to more seamless, responsive on-device user experiences across leading mobile apps and features.  

Y is for Years of innovation, with Arm technologies standing the test of time, from the 40-year anniversary of the Arm architecture to the company’s 35-year birthday that recognized the 35 Arm-powered technology products that have redefined modern computing.

Z is for Zena CSS, Arm’s standardized, pre-integrated compute platform for AI-defined vehicles, which accelerates software and silicon development and deployment, helping to deliver AI capabilities faster and more efficiently. 

Looking to 2026

As we look ahead to 2026, one thing is clear: Arm will continue to be the driving force behind building, optimizing, and scaling the world’s most advanced, efficient AI experiences everywhere. We look forward to what the next year brings and the ongoing evolution of AI and compute on the Arm platform.