In September 2020, Nvidia Corporation announced a proposed acquisition of Arm Holdings, a leading semiconductor intellectual property (IP) company, in a deal valued at $40 billion. The transaction, structured as a combination of cash and stock, was intended to be the largest semiconductor acquisition in history at the time. However, after facing prolonged regulatory scrutiny and opposition from industry competitors and government authorities, the deal was terminated in February 2022.
Arm Holdings, headquartered in Cambridge, England, designs the processor architectures used in the vast majority of smartphones and a wide range of other devices, including embedded systems, networking equipment, and increasingly data center servers. The company's business model is based on licensing its IP to chip manufacturers and system-on-chip (SoC) designers, rather than manufacturing its own chips. This model has made Arm a neutral player in the semiconductor industry, with customers that include many of Nvidia's direct competitors.
Background: Nvidia and Arm
Nvidia, founded in 1993, is an American technology company known primarily for its graphics processing units (GPUs). Over the years, Nvidia expanded from gaming into high-performance computing, artificial intelligence (AI), and data center acceleration. By 2020, Nvidia had become one of the most valuable semiconductor companies in the world, with a market capitalization exceeding $300 billion. Its GPUs were widely used for training and deploying AI models, and the company had also developed the Tegra line of mobile processors, which incorporated Arm-based CPU cores.
Arm Holdings, originally a joint venture between Acorn Computers, Apple, and VLSI Technology, was spun off as an independent company in 1998. It was acquired by SoftBank Group in 2016 for approximately $32 billion. Arm's architecture is ubiquitous in mobile devices, with over 95% of smartphones using Arm-based processors. The company also licenses its designs for use in networking, automotive, and Internet of Things (IoT) applications. In the late 2010s, Arm began making inroads into the server market with its Neoverse line of data center processors, competing with Intel's x86 architecture.
The Announcement
On September 13, 2020, Nvidia and SoftBank announced that they had reached a definitive agreement for Nvidia to acquire Arm in a transaction valued at $40 billion. The deal would be paid for with a combination of $12 billion in cash and $21.5 billion in Nvidia common stock, with the remainder in other considerations, including a potential earn-out for SoftBank if Arm met certain performance targets. The acquisition was expected to close within 18 months, subject to regulatory approvals in the United States, the United Kingdom, the European Union, China, and other jurisdictions.
Nvidia's CEO Jensen Huang stated that the acquisition would create a company that could address the growing demand for AI computing across all scales, from edge devices to data centers. The combined entity would leverage Arm's extensive ecosystem and Nvidia's AI computing capabilities. SoftBank, which was facing financial difficulties due to losses in its Vision Fund, would retain a minority stake in Nvidia as part of the deal.
Regulatory and Industry Reaction
The announcement immediately drew concerns from industry players and regulators. Arm's neutrality was seen as critical to the semiconductor ecosystem, as many of its licensees, including Qualcomm, Samsung Electronics, Apple, and Intel, compete with Nvidia in various markets. These companies feared that Nvidia could restrict access to Arm's IP or favor its own products, undermining competition.
In the United Kingdom, where Arm is headquartered, politicians and regulators expressed concerns about national security and the potential loss of jobs. The UK government ordered a national security review of the deal in April 2021. In the European Union, the European Commission launched a formal investigation in October 2021, citing concerns that the acquisition could lead to higher prices, less choice, and reduced innovation in the semiconductor industry.
In the United States, the Federal Trade Commission (FTC) filed a lawsuit in December 2021 to block the acquisition, arguing that it would harm competition in several markets, including data center CPUs, automotive SoCs, and networking chips. The FTC's complaint noted that Nvidia had a history of using its GPU dominance to extract concessions from customers and that the acquisition would give Nvidia the incentive and ability to withhold or degrade Arm's IP to competitors.
Opposition from Competitors
Several major technology companies, including Qualcomm, Google DeepMind (though not directly involved), Microsoft (AI), and Amazon Web Services, reportedly voiced opposition to the deal in private meetings with regulators. Qualcomm, a major Arm licensee, was particularly vocal, with its president Cristiano Amon stating that the acquisition would be "catastrophic" for the industry. Google and Amazon, which design their own Arm-based chips for data centers, also expressed concerns about the potential for Nvidia to favor its own products.
Arm's co-founder Hermann Hauser launched a public campaign against the deal, calling it a "disaster" for the UK and the global tech industry. He argued that Arm's independence was essential for maintaining a level playing field in the semiconductor market. A petition against the acquisition gathered over 2,000 signatures from industry professionals.
Nvidia's Defense
Nvidia sought to address these concerns by making voluntary commitments. In October 2021, Nvidia offered to license Arm's technology on a fair, reasonable, and non-discriminatory (FRAND) basis, and to maintain Arm's open licensing model. The company also pledged to keep Arm's headquarters in Cambridge and to expand its workforce in the UK. Nvidia argued that the acquisition would actually benefit the industry by increasing investment in Arm's technology and accelerating innovation.
However, these commitments were not sufficient to assuage regulators. The UK's Competition and Markets Authority (CMA) concluded in its report that the deal could harm competition in the markets for data center CPUs, automotive SoCs, and other products. The CMA recommended a Phase 2 investigation, which was initiated in November 2021.
Termination of the Deal
On February 8, 2022, Nvidia and SoftBank announced that they had mutually agreed to terminate the acquisition. The decision came after it became clear that the deal would not receive regulatory approval in a timely manner. The termination was attributed to "significant regulatory challenges" that prevented the completion of the transaction. Under the terms of the agreement, SoftBank retained the $1.25 billion breakup fee that Nvidia had paid, and Arm announced plans to pursue an initial public offering (IPO) instead.
Nvidia's CEO Jensen Huang expressed disappointment but stated that the company would continue to collaborate with Arm as a licensing partner. SoftBank's CEO Masayoshi Son said that Arm was well-positioned for growth in the AI era and that the IPO would allow the company to realize its full potential.
Aftermath and Impact
The termination of the deal was widely seen as a victory for regulators and industry competitors who had opposed the acquisition. It also highlighted the increasing scrutiny of large technology mergers, particularly in the semiconductor sector, which is considered strategically important for national security and economic competitiveness.
Arm subsequently filed for an IPO on the Nasdaq in September 2023, raising $4.87 billion at a valuation of $54.5 billion. The IPO was well-received, and Arm's stock price surged on its first day of trading. The company continues to play a central role in the global semiconductor industry, with its architecture used in billions of devices.
For Nvidia, the failed acquisition did not hinder its growth. The company continued to dominate the AI GPU market, with its data center revenue soaring in the following years. By 2025, Nvidia had become one of the most valuable companies in the world, with a market capitalization exceeding $3 trillion. The company also expanded its own Arm-based CPU offerings for data centers, such as the Grace processor, which was announced in 2021 and launched in 2023.
The proposed acquisition of Arm remains a case study in the challenges of consolidating critical technology assets. It underscored the importance of maintaining neutral and open ecosystems in the semiconductor industry, and it set a precedent for future regulatory reviews of large tech deals.
Legacy and Lessons
The failed deal had several lasting effects. It reinforced the role of antitrust regulators in shaping the technology landscape, particularly in the European Union, the United Kingdom, and the United States. It also prompted other companies to reconsider their own acquisition strategies, with some opting for partnerships or licensing agreements instead of outright acquisitions.
For Arm, the episode led to a renewed focus on its independence and its potential for growth in new markets, including Artificial intelligence and Machine learning applications. The company's IPO in 2023 was a testament to its resilience and the continued demand for its technology.
For Nvidia, the experience highlighted the limits of its influence in the face of coordinated regulatory opposition. However, it also demonstrated the company's ability to pivot and continue its aggressive expansion in AI and data center markets. The company's subsequent success with its Grace CPU and other Arm-based products showed that it could compete without owning Arm.
The proposed acquisition of Arm was a landmark event in the history of the semiconductor industry. It brought together two of the most influential companies in computing, but ultimately failed due to the fundamental conflict between Nvidia's commercial interests and Arm's role as a neutral technology provider. The outcome was a clear signal that the industry and regulators would not allow a single company to control the architecture that underpins most of the world's digital devices.