Currently, we are in the early stages of an exciting transformation in artificial intelligence (AI) technology. With the accelerated evolution of natural language, multimodal big models, and generative AI technology, AI is reshaping various industries at an unprecedented speed. According to IDC's forecast, the global data volume will grow from 159.2 ZB in 2024 to over 384.6 ZB in 2028, with a compound annual growth rate of 24.4%. By 2028, it is expected that 37% of data will be generated directly in the cloud, while the remaining data will be generated directly from the edge and endpoints.

Faced with the surge of edge data, efficient data processing, low latency transmission, and intelligent and secure storage are becoming the focus of industry attention. The future computing architecture must not only provide stronger computing power, but also be more closely integrated with storage systems to ensure that AI models can run efficiently, while optimizing data management and access methods.

From the current development direction of AI technology, on the one hand, large models are evolving towards General Artificial Intelligence (AGI), exploring new directions such as multimodal and physical AI, and continuously challenging new limits of computing power. On the other hand, in order to promote the overall deployment of the big model, the industry began to move towards in-depth optimization and customization of vertical fields, so that the big model can go into thousands of industries and adapt to different scenarios such as mobile terminal, edge computing, cloud deployment, etc.

The launch of DeepSeek has had a profound impact on the global AI market: as an open and innovative technology, it not only demonstrates the optimization potential of AI in the training and inference process, but also greatly improves the efficiency of large-scale deployment, fully proving that the model can run stably in a lower cost and higher efficiency environment. This achievement is of great significance for promoting the large-scale application of AI in enterprise applications and edge computing.

Arm Computing Platform: Continuously Promoting AI Optimization Deployment from Cloud to End

In the early stages of AI development, data centers, as the core location for model training and initial inference, are facing unprecedented challenges. Traditional standard general-purpose chips are inadequate in handling computationally intensive AI workloads and cannot meet the urgent needs of the AI era for high performance, low power consumption, and flexible scalability. In this context, the Arm computing platform, with its advanced technological advantages, has opened up a new paradigm for the development of next-generation AI cloud infrastructure. From the Arm Neoverse Computing Subsystem (CSS), Arm Total Design ecosystem project to the Core System Architecture (CSA), Arm has made an integrated layout from technology to ecology. It not only provides efficient, flexible, and scalable solutions for AI data center workloads, but also helps partners focus on product differentiation and accelerates the product launch process.

AI reasoning is the key to unlocking value for AI, and it is rapidly expanding from the cloud to the edge, covering every corner of the world. In the field of edge AI, Arm continuously innovates with its unique advantages in technology and ecology, ensuring that the intelligent Internet of Things and consumer electronics ecosystem can perform optimal workloads at the right time and in the most suitable location.

In order to meet the increasing demand for AI workloads in edge AI, Arm recently released an edge AI computing platform centered around the new Armv9 ultra high energy efficiency CPU Cortex-A320 and Ethos-U85 AI accelerator with native support for Transformer networks. This platform achieves deep integration of CPU and AI accelerator. Compared to last year, the Cortex-M85 combined with Ethos-U85 platform has improved machine learning (ML) computing performance by eight times, bringing significant breakthroughs in AI computing power and empowering edge AI devices to easily run large models with over 1 billion parameters.

Among them, the newly released ultra high energy efficiency Cortex-A320 not only provides higher memory capacity and bandwidth for Ethos-U85, making the execution of large models on Ethos-U85 more powerful, but also supports larger addressable memory space and can more flexibly manage multi-level memory access latency. The combination of Cortex-A320 and Ethos-U85 is an ideal choice for running large models and addressing the memory capacity and bandwidth challenges brought by edge AI tasks.

In addition, Cortex-A320 fully utilizes the AI computing features enhanced by Armv9, as well as security features including Secure EL2, Pointer Validation/Branch Object Recognition (PACBTI), and Memory Tag Extension (MTE). Previously, these features have been widely applied in other markets, and Arm has now introduced them into the field of IoT and edge AI computing, providing excellent and flexible AI performance while achieving better isolation of software loads and protection against software memory anomalies, improving overall system security.

The development of storage in the AI era: a comprehensive upgrade of storage, computing, and security capabilities

With the continuous growth of AI computing demand, cloud edge devices have put forward higher requirements for computing power, as well as more stringent requirements for storage system performance, density, real-time performance, and power consumption. In traditional models, computing architectures often separate storage and computation, with storage devices only playing the role of data storage. Data needs to be frequently moved between storage and computing nodes, resulting in a bottleneck between storage and computation. However, in the era of AI, in order to meet the needs of real-time data analysis, intelligent management, and efficient access, it has become particularly crucial to place storage closer to computing units or to enable the storage itself to have computing power. This ensures that AI tasks are efficiently executed in the most suitable location.

The requirements for storage throughput, latency, energy consumption, security, and improving host manageability such as Open Channel vary from cloud to end-to-end AI computing. The storage controller and firmware running on the Arm CPU in the storage controller play an extremely important role in supporting differentiated AI storage requirements.

1.png

Image: Arm's rich IP platform solutions provide leading performance and energy efficiency for AI storage

In fact, as the cornerstone of data storage and network control, Arm has been providing high-performance, low-power, secure and reliable solutions for global storage controllers and devices, including:

The Arm Cortex-R series real-time processors have the fastest interrupt latency and real-time response speed, and are widely used in many storage devices;

The Arm Cortex-M series embedded processors are a popular choice for backend flash and media control, and support custom instructions. Customers can create differentiation through deep optimization for unique NAND media;

The Arm Cortex-A series application processors are designed with a high-throughput pipeline, supporting the highest processing performance, and have solid ecological support for ML, data processing software, and rich operating systems;

The Arm Ethos-U AI accelerator supports native Transformer acceleration of 2048 MACs per second, which can help the storage controller itself become smarter;

In addition, there is a Neoverse tailored for data centers. We have begun to see innovative designs in CXL (Compute Express Link) that combine Arm Coherent Mesh Network (CMN) with Neoverse to achieve "compositional" memory expansion and incorporate the concept of near storage computing to reduce data handling.

Ecological collaboration, building the future of AI computing and storage

While focusing on providing leading technology and products, Arm is also committed to working together with ecosystem partners to promote the development of the storage industry. The Arm architecture based platform is widely adopted by industry-leading storage enterprises to optimize their storage solutions. For example, Solidigm's latest 122TB PCIe SSD Solidigm? D5-P5336 significantly improves the energy efficiency, storage density, and performance of AI data centers. Its storage controller uses Arm Cortex-R CPU, effectively enhancing real-time read and write performance and latency certainty; Silicon Motion's SM2508 main control chip for AI PCs adopts Arm Cortex-R8 and Cortex-M0, achieving breakthroughs in energy efficiency and data throughput. Its SM2264XT-AT is the industry's first PCIe Gen4 main control chip for vehicles, which supports data access for mixed critical workloads through enhanced virtualization and can save 30% of energy consumption; The XP2300, ORCA 4836 and UNICIA 3836 SSDs built by Jiang Bolong based on the Arm Cortex-R CPU are widely used in AI PC, servers, cloud computing, distributed storage, edge computing and other application scenarios by virtue of their advantages of large capacity and high performance, to meet the local deployment requirements of AI technology.

In addition, in the local storage market, leading storage companies such as Dapu Microelectronics, Lianyun Technology, Yixin Technology, Tenfei, Deyi Microelectronics, and Yingren Technology have also widely adopted Arm technology to create SSD main control chips and device solutions.

So far, nearly 20 billion storage devices have been applied based on the Arm architecture and platform, including cloud and enterprise SSDs, vehicle SSDs, consumer SSDs, hard drives, and embedded flash devices. Currently, storage devices powered by Arm technology continue to maintain a daily shipment volume of approximately 3 million units.

With cutting-edge technological strength, rich ecological layout, and profound accumulation in the storage industry, Arm is continuing to lead technological innovation and empower the development of computing and storage in the AI era. Arm will also continue to work with partners to build a new future for computing and storage in the AI era through a secure and efficient Arm computing platform.