Is Nvidia About to Buy Arm?

Acquisition would be a stretch, despite growing rumours

Speculation is rife that Nvidia is in advanced talks to acquire Arm. The British semiconductor design company was acquired by SoftBank for $32 billion in 2016. However, SoftBank is facing challenging times as several high-profile investments have fallen short of expectations. CEO Masayoshi Son has pledged to divest some major assets in an effort to cut debt and improve the balance sheet. To date, this has included selling part of its stakes in Alibaba and T-Mobile US.

If the reports are true, the move would reflect very badly on SoftBank. Arm is a licensing business and its architecture is central to a vast array of connected devices. So far, Arm partners have shipped more than 165 billion chips based on its technology and an average of 22 billion a year over the past three years. Arm’s role in a growing proliferation of connected devices lured SoftBank, but the Japanese company had to give clear assurances of its commitment to the business and its ongoing independence. If SoftBank were to divest Arm after just a few years, this would rightly result in more scrutiny of any future deals.

Nvidia isn’t an obvious candidate to buy Arm, but there’s some logic to the argument from Nvidia’s perspective, at least on first appearance. Nvidia has grown tremendously over the past few years thanks to the growing prevalence of its GPU architecture and platforms. In that time, its market capitalization has risen to $261 billion, higher than that of Intel, giving Nvidia the muscle to address a missing piece of its strategy and establish a CPU capability that can be more tightly integrated with its GPUs.

Its acquisition of Mellanox in 2019 showed that Nvidia is taking a broader systems view and wanting to control more of the elements that underpin its solutions — the networking interconnect between GPUs, in the case of Mellanox. Arm would strengthen Nvidia’s position in silicon for data centres, the industrial Internet of things, and especially smaller client devices.

But this reasoning is entirely theoretical. In reality, Arm is a licensing business. Although it could put Nvidia in a powerful position of control, some $35 billion would be a steep price to pay, and licensing alone offers little synergy. Nvidia could license Arm technology and build its own CPU cores, or partner to do so without acquiring Arm.

The argument that Nvidia would gain control over Arm’s road map doesn’t hold water either. Any intention to exert control over the future direction of the Arm architecture and bend it toward Nvidia’s own needs would be very unlikely to pass intense regulatory scrutiny.

A lot of stories have been written about whether Nvidia should or shouldn’t acquire Arm. I firmly believe that such a move would be detrimental to Arm and its ecosystem, and would not be good news for the industry if it were allowed to proceed. Here’s why:

Arm’s value is its ecosystem. The company has a huge range of partners that rely on Arm for an increasingly wide array of uses. Its independence is critical to its ongoing success, a point that Arm itself has emphasized in the past. The moment that control shifts to a rival, this independence is compromised. We suspect that Apple was approached by SoftBank and very wisely declined the opportunity.

The value of Arm’s independence is such that an acquisition by Nvidia would be likely to erode Arm’s value. Arm is facing growing competition from RISC-V, an open-source architecture. If its partners believed that Arm’s integrity and independence was compromised, it would accelerate the growth of RISC-V and in the process devalue Arm.

RISC-V is poised for rapid growth and poses a competitive threat. I believe this is a further reason for SoftBank’s determination to divest Arm. This should be a central consideration for Nvidia. A long line of Arm licensees are also part of the RISC-V community and steadily upping their investment. There’s a scenario in which Nvidia pays a rich premium for Arm, the very act of which sparks an erosion in its value as focus shifts to a more independent alternative.

Regulatory approval would be extremely difficult for all the reasons above. The deal would have a detrimental and destabilizing effect on the Arm community, take months, if not years, to go through the process and ultimately fail to get approval. This would be damaging to all parties and the uncertainty alone would hurt Arm regardless of the outcome.

This raises the question of who would buy Arm, if not Nvidia. The truth is, I don’t see any other interested parties or contenders that could make a deal work — at least not today. But a sale of Arm isn’t the only option available to SoftBank.

The alternative, and by far the best option, is that Arm returns to public ownership. This would ensure its ongoing independence and that the company retains a management team that understands its value and priorities. Despite the speculation, I’m optimistic that this is the more likely option. Let’s hope that SoftBank ultimately pursues the IPO path for Arm.