Samsung and Arm join forces to enter the edge AI market! Custom 2nm chips launched, OpenAI may become a key customer.

Samsung Electronics is deepening its collaboration with chip architecture giant Arm to jointly develop custom chips for edge AI. If the project is ultimately implemented, it is expected to bring new design orders to Samsung's system LSI business and also introduce advanced process customers to its wafer foundry business, becoming an important attempt for Samsung to enter the AI chip market and compete for custom chip orders.

On September 4th, it was reported that Arm had approved the initial development costs (NRE) for its next-generation edge AI system-on-a-chip (SoC) at the end of August and officially launched the project with Samsung Electronics. According to the division of labor, Arm will be responsible for providing the AI accelerator (AIC) architecture, RTL and other design technologies, communicating end-customer needs, and coordinating project development; Samsung Electronics' System LSI division will be responsible for SoC design, while its foundry division plans to use an advanced 2nm process for mass production.

The report, citing industry insiders, revealed that OpenAI is the candidate for the project's final customer. As AI capabilities rapidly extend from the cloud to edge devices, OpenAI's demand for dedicated AI chips is expected to rise if it further expands its edge AI offerings. Arm has previously been working with OpenAI on related mass production collaborations.

With the increasing demand for edge AI, customized chips are emerging as a new direction.

Behind this collaboration is the increased demand for chips driven by the accelerated deployment of edge AI applications. Unlike traditional AI models that rely on cloud data centers, edge AI can perform some AI inference directly on terminal devices such as smartphones, thereby reducing the computing power pressure on the cloud and improving response speed and data privacy.

As AI manufacturers extend their reach to end-user hardware, general-purpose chips may not be able to fully meet the computing power, power consumption, and cost requirements of different devices and application scenarios. Customized AI chips have therefore become a new focus for the industry chain. The project launched by Samsung and Arm this time is precisely in line with this trend.

In terms of the cooperation model, both parties adopt a Limited Use License (LUL) framework. The relevant technologies provided by Arm are limited to the development of specific customer products, and their scope of use is clearly restricted. Arm mainly plays the role of technology provider and development partner in the project, rather than directly selling chips to the market.

After completing the design and manufacturing, Samsung Electronics will directly supply the chips to end customers and realize commercialization. This means that Samsung will not only be able to obtain revenue from the chip design and manufacturing process, but also further enhance its overall competitiveness in the field of customized AI chips through the synergy of its system LSI and wafer foundry businesses.

With 2nm mass production underway, Samsung's foundry business will also benefit.

For Samsung, the more significant aspect of this project lies in its potential to create a complete AI chip solution encompassing both design and manufacturing.

System LSI handles SoC design, while wafer foundry handles 2nm mass production, enabling the two businesses to collaborate within the same project. Compared to simply securing a chip design order, this model allows Samsung to directly embed its advanced process capabilities into the AI chip development process and compete for future customers with complete solutions.

Especially in the wafer foundry business, the actual development and mass production experience of 2nm edge AI chips has strong demonstrative significance. If the project progresses smoothly, Samsung can use this to accumulate mass production cases of advanced process services for AI chip customers, providing a reference for winning more orders for AI accelerators, custom SoCs, etc. in the future.

With the demand for AI chips continuing to spread to end devices, whether Samsung can leverage Arm and the potential OpenAI project to open up the market for custom AI chips will become a new point of interest in whether its system LSI and wafer foundry businesses can achieve synergistic growth.

Risk warning and disclaimerInvesting involves risk; please exercise caution. This article does not constitute personal investment advice and does not take into account the specific investment objectives, financial situation, or needs of individual users. Users should consider whether any opinions, views, or conclusions in this article are suitable for their specific circumstances. Any investment decisions made based on this information are at your own risk.