In this era of dynamic markets and rapid innovation, competition drives Arm to become a better company.

Ian Smith, Vice President at Arm, expressed confidence as he observed technological advancements and market changes in the rapidly changing AI era.

The Arm Tech Symposia 2023 was held on the 16th at the InterContinental Seoul COEX in Samseong-dong, Seoul.

Held under the theme 'The Future of Computing in the AI Era,' this event took place in 7 cities across 4 countries: Korea, Japan, China, and Taiwan. In Korea, over 1,900 people registered, and the venue was bustling with attendees.


Hwang Sun-wook, Country Manager of Arm Korea, emphasized Arm's ecosystem that contributes to the expansion of the developer community and partner ecosystem, stating, “For domestic fabless companies and startups "We support the Arm Flexible Access Program every year, and 18 companies benefited this year, with a total of 33 companies currently producing samples," he said.

General Manager Hwang stated, “I look forward to sharing Arm’s latest trends in various applications to explore future opportunities and provide valuable insights through this event,” adding that “Ian Smith, Vice President of Product Marketing at Arm, Baek Jun-ho, CEO of Furiosa AI, and Baek Jun-hyun, CEO of Jaram Technology, delivered keynote speeches to share insights on the semiconductor industry as it enters the AI era.”

In his keynote speech, Vice President Ian Smith said, “Arm has been transitioning into a computing platform for years,” adding, “As a foundational technology, Arm’s solutions contribute to the overall spread of artificial intelligence from the edge to the server through mobile, IoT, automotive and SDV, and infrastructure.”

■ AI on Edge Devices Expected to Be the Mainstream

Vice President Ian Smith noted that “configuring AI at the edge is a trend,” predicting that “as device inference expands, there will be more instances where AI inference devices are connected to data centers,” and emphasized the importance of enhanced security and software integration in this context.

While generative AI and LLMs are receiving attention today, Arm has been leading the way in delivering AI at the edge for years. It has arrived. In the smartphone sector, 70% of third-party AI applications are running on Arm CPUs.

However, as the industry seeks ways to deliver AI sustainably and move data efficiently, it must evolve to run AI and machine learning (ML) models at the edge. Meanwhile, this is a challenging task as developers are working with increasingly limited computing resources.

Arm is collaborating with NVIDIA to implement NVIDIA TAO, a low-code open-source AI toolkit for Ethos-U NPUs, which enables the creation of performance-optimized vision AI models for deployment on processors.

NVIDIA TAO provides an easy-to-use interface for building based on TensorFlow and PyTorch, representative free open-source AI and ML frameworks. For developers, this means they can easily and seamlessly develop and deploy models, while bringing more complex AI workloads to edge devices to deliver enhanced AI-based experiences.

■ Edge AI, Neural Network Development

A key aspect of the continued growth of AI is advancing the deployment of neural networks at the edge.

Arm and Meta are working to bring PyTorch to Arm-based mobile and embedded platforms at the edge through ExecuTorch. Developers can use ExecuTorch to deploy state-of-the-art neural networks required for advanced AI and ML workloads across mobile and edge devices much more easily.

Moving forward, through the collaboration between Arm and Meta, AI and ML models can be easily developed and deployed using PyTorch and ExecuTorch.

The collaboration with Meta is based on the significant achievements Arm has already made in investing in the Tensor Operator Set Architecture (TOSA), which provides a common framework for AI and ML accelerators and supports a wide range of workloads used in deep neural networks.

TOSA is expected to become the cornerstone of AI and ML for various processors and billions of devices built on Arm architecture.

■ “Am’s Commitment to Building a Broad Ecosystem in the Dynamic AI Market”

Arm is expected to become a company that shapes the future of computing in the AI era. The current data center ecosystem faces challenges with high costs and limited options for AI semiconductors, suggesting that Arm will build the foundation and make efforts to resolve these issues.

In addition to its own technology platform already under development, Arm is collaborating with leading technology companies such as AMD, Intel, Meta, Microsoft, Nvidia, and Qualcomm Technologies on initiatives focused on implementing advanced AI capabilities for low latency and secure user experiences.

During a media briefing, Vice President Ian Smith stated, “Arm’s goal is to provide more platforms, and it is important to focus on what our partners expect,” adding, “A broad ecosystem including software stacks, platforms, and partners is driving computing innovation.”

Arm has grown as a company that sells IP and focuses on mobile and edge devices, but it is currently transforming into a provider of computing platforms. Since current CEO Rene Haas took the helm, the company has been pursuing growth by focusing on IoT, automotive and SDV, infrastructure, and consumer sectors.

Just as CEO Rene Haas predicts that everything from mobile devices to coffee machines is becoming computerized and that even more will be intelligently computerized, Arm is preparing early for the future of the AI era.

Arm is confident that the future of AI will be built on Arm, as the foundation for global AI deployment is already based on Arm, and it extends from sensors, smartphones, and software-defined vehicles (SDVs) to servers and supercomputers.