Nvidia sells final $140M Arm stake, ends acquisition chapter

The divestment, which involved 1.1 million shares, was executed in the final months of 2025, with the transaction only recently coming to light through a regulatory filing, as reported by Tom's Hardware. This final sale marks the conclusive step in unwinding what was once one of the technology sector's most ambitious and scrutinized proposed mergers.

This move formally concludes a saga that began in 2020 when Nvidia announced its intention to acquire Arm from SoftBank for $40 billion. The deal was ultimately scuttled by stringent regulatory opposition from authorities in the United Kingdom and the European Union, who expressed significant concerns over the potential for reduced competition in the global semiconductor market. The forced abandonment of the acquisition left Nvidia with a substantial minority stake in the chip designer, which it has now fully liquidated.

Despite the complete dissolution of its financial interest, the commercial relationship between Nvidia and Arm remains firmly intact. According to the reports, the two companies will continue their critical technical partnership. Nvidia’s development of its Arm-based data center CPUs, including the Grace Hopper and Blackwell platforms, is expected to proceed unchanged. Furthermore, Tom's Hardware notes that the upcoming Vera Rubin AI platform will also leverage Arm’s CPU architecture, underscoring a continued strategic alignment.

The enduring partnership highlights a pragmatic industry reality where collaboration often supersedes ownership. Nvidia’s reliance on Arm’s energy-efficient architecture has become a cornerstone of its strategy to dominate the AI data center market, where pairing its powerful GPUs with sophisticated Arm-based CPUs creates a compelling integrated solution. This technical synergy has proven too valuable to be disrupted by the collapse of the acquisition.

The context of this divestment extends beyond a simple corporate restructuring. As noted in additional coverage, Nvidia has recently entered into a significant partnership with Intel, planning to utilize Intel’s foundry services and its x86 architecture for future products. This diversification of partnerships suggests a strategic calculus where Nvidia is opting for flexibility rather than exclusivity, securing access to the best available technologies from multiple leading providers rather than tying its fate to a single architecture through acquisition.

The sale of Nvidia’s stake also arrives as the competitive landscape for AI infrastructure intensifies. Other cloud infrastructure providers are making substantial investments, such as Vultr’s recently announced plan to invest over $1 billion in an Ohio AI cluster utilizing AMD chips. This competitive pressure necessitates that Nvidia remain agile and focused on its core competencies of GPU and AI accelerator design, rather than managing the broader complexities that would have come with owning a foundational IP company like Arm.

Looking ahead, the closure of this chapter allows both companies to move forward with clarity. For Arm, it operates with continued independence under its parent company, SoftBank. For Nvidia, the path is one of deepened collaboration through licensing rather than ownership, a model that has already proven highly effective for its business. The focus now shifts entirely to execution, as the company continues to develop its next-generation processors that will power the advanced AI workloads of the future.