According to Zhitong Financial APP, Qualcomm's CEO told the jury in court that Qualcomm's business has been significantly harmed due to clients being misled by computer chip design company Arm Holdings—specifically regarding the licensing agreements for chip technologies that are critical to the chip electronic ecosystem. Cristiano Amon, who was elected as Qualcomm's CEO in 2021, testified on Wednesday during a patent licensing dispute that could impact the entire chip industry.

He testified that Arm unnecessarily created 'uncertainty' for Qualcomm's clients (including Smart Phone manufacturer Samsung Electronics)—misleading them into believing that licenses for using Arm technology would expire in 2025, when in fact Qualcomm has completed its self-developed CPU architecture Oryon and successfully launched it to market. He stated that Samsung began to worry that the chips used in its Smart Phone products might lose access to CPU infrastructure services.

"This is not good for Qualcomm, our clients, or frankly, for Arm and the chip industry," Amon told the jury at the federal court in Wilmington, Delaware. He indicated that the licensing would actually last until 2033.

Qualcomm has long been one of Arm's largest clients and a long-term partner, but the differences between the two have been growing. The core of the legal dispute between the two sides lies in Qualcomm's acquisition of chip design startup Nuvia for $1.4 billion in 2021 and the signing of a licensing agreement to use ARM architecture technology.

Although Qualcomm's self-developed Oryon CPU architecture is based on the ARM instruction set, it is a customized architectural version developed by Qualcomm based on Nuvia's innovative technology, optimized for high-performance computing and AI acceleration. This self-developed architecture will help Qualcomm maintain competitiveness in future mobile devices, Cloud Computing, and AI applications.

Arm claims that after Qualcomm acquired Nuvia, it should have renegotiated the agreement with Nuvia and required Qualcomm to completely destroy the self-developed chip architecture designs obtained in the acquisition. Headquartered in San Diego, Qualcomm is relying on Nuvia's innovative chip architecture technology to expand from Smart Phones into the PC, AI, and electric vehicle processor markets, arguing that the company holds a separate license for ARM architecture technology covering all its business areas.

This dispute is crucial for the chip industry, as many large Global Technology companies rely on chip architectures and chip products licensed from Arm and Qualcomm to drive their Consumer Electronics. Jim McGregor, partner and Analyst at Tirias Research, stated that the jury's ruling could affect future intellectual property licensing agreements and the contract law of the entire Technology industry.

Qualcomm acquired the chip startup Nuvia in 2021 with the aim of expanding its market share in computer processors through its technology in high-performance computing and AI large model processing. Qualcomm claims that the Nuvia technology obtained after the acquisition is already included in the existing Arm technology licensing agreement, allowing it to legally use Nuvia's technology to design its own CPU architecture.

Arm, on the other hand, argues that the technology licensing agreement obtained after the Nuvia acquisition should be renegotiated and demands that Qualcomm destroy all technical designs acquired during the purchase process. Arm believes that the acquisition changed the terms of the licensing agreement and a new agreement must be reached. Qualcomm insists that Arm's request for renegotiation and destruction of Nuvia designs is unfounded, which has led to differing interpretations of the existing agreement terms between the two companies, becoming the core of the dispute.

During Wednesday's inquiries, Amon stated that despite the dispute with Arm, Qualcomm would still fulfill its valid licensing commitments. When asked whether commercial friction justified violating the agreement with Arm, Amon emphasized: "I believe no company should do that."

Amon countered Arm's claim that Qualcomm knew it needed to obtain licensing authorization to deploy the technology obtained during the acquisition of Nuvia, stating that Qualcomm's existing agreement with Arm did not impose those requirements from Arm. The key issue in this case is whether the wording of the licensing agreement requires that the use of Nuvia's innovative design technology must receive reauthorization from Arm.

Qualcomm's CEO stated that after acquiring Nuvia, Arm insisted on canceling Nuvia's authorization and demanded that his company destroy all innovative results obtained in this transaction, which is simply "outrageous" and disrupts the development order of the entire chip industry. He stated that it is almost unheard of for a Technology company to demand the destruction of technology for which licensing fees have already been paid, while also requiring reauthorization.

The case is expected to conclude with closing arguments on Thursday Eastern Time. Subsequently, a jury consisting of five women and three men will begin deliberations. The case of Arm vs. Qualcomm is numbered "22-cv-01146" and is being tried in the United States District Court for the District of Delaware (Wilmington).

Who is Arm? Why does Qualcomm's Smart Phone chips rely on it so much?

Although Qualcomm has intensified efforts to advance its self-developed CPU architecture this year, such as the second generation customized Qualcomm Oryon CPU, the Snapdragon processors have been firmly tied to the ARM architecture for many years.

The UK-based company Arm, controlled by SoftBank Group led by Masayoshi Son, is a core force in the chip design field. Arm plays a unique role in the chip design industry: its architectural designs and standards are the foundation of the semiconductor products used in most smartphones worldwide.

Under the leadership of CEO Haas, the company has been striving to expand its business scope into the incredibly profitable datacenter server market, which has proven to be the most critical revenue engine for Arm currently. As part of its licensing agreement, in addition to providing the instruction set architecture and design modules for chip design, Arm also aims to provide complete chip design solutions for companies to use directly, a transformation aimed at making it more profitable and reducing heavy reliance on royalty-type sales.

Arm sells licenses for the core instruction set architecture of almost all mobile chips in smartphones globally, the Reduced Instruction Set Computing architecture—known as 'ARM architecture'—which is mainly applied in the smartphone field. However, the ARM architecture is increasingly appearing in personal computers and datacenter AI server chip sectors, especially since the AI boom swept the globe in 2023. The footprint of the Reduced Instruction Set Computing architecture—ARM architecture has fully spread from smartphones to AI datacenter server CPUs, with technology giants like NVIDIA and Microsoft choosing to design their server CPUs using the ARM architecture.

NVIDIA's self-developed Grace CPU is based on the ARM architecture, and Amazon's self-developed Graviton server processor for datacenters also uses the ARM architecture. Microsoft's custom AI chip Azure Cobalt 100, which is a server CPU, is likewise built on the ARM architecture and is specially designed to run cloud computing workloads on Microsoft Azure cloud servers.

The Reduced Instruction Set Computing architecture used by ARM provides the server CPUs based on its designs with significant advantages in energy efficiency and low power consumption when performing AI inference/training tasks compared to Intel's x86 architecture. This characteristic makes ARM architecture particularly suitable for the datacenter server sector, enabling it to meet the nearly endless demands for AI inference/training computing power with low power consumption and high efficiency attributes in conjunction with NVIDIA's high-performance AI GPUs.

Especially, ARM's latest instruction set architecture ARMv9 is designed for handling extremely complex AI workloads, significantly accelerating artificial intelligence data processing and optimizing power consumption, which is why ARMv9 can contribute up to 25% of revenue from royalty type sales. ARMv9 introduces the SVE2 instruction set, which brings a significant leap in performance in applications in AI and machine learning that require a large number of matrix operations.