SoftBank Up in Arms

End of the road for Nvidia’s acquisition of Arm

When Nvidia’s proposed acquisition of Arm was announced in September 2020, there were two distinct camps of opinion. One side viewed it as a fabulous move that ensured the long-term prosperity of Arm through investment and alignment with Nvidia’s broader strategy; the other side, including CCS Insight, believed it compromised the very basis of a business model built on Arm’s independence. As we wrote in March 2021, “We continue to view regulatory approval as an uphill struggle for Nvidia, with the better outcome for Arm being a return to a public listing”; see also our subsequent views published in May, August and September 2021, as well as our original article prior to the acquisition announcement.

It was difficult to see how Nvidia could maximize its original $40 billion offer and gain the competitive advantages it needed while maintaining its independence. The concerns raised by the UK Competition and Markets Authority on 20 August 2021 outlined this clearly. “Behavioral undertakings” designed to address concerns about competition and independence were dismissed. This included an open licensing regime based on equal access and interoperability, adding protections against disclosure of competitively sensitive information. Given that it included no mechanism to ensure “effective monitoring and enforcement”, this wasn’t a surprise.

Concern was also raised about the ability to divest parts of Arm’s intellectual property. Early on, we flagged this as a sticking point, given that the breadth of Arm’s patents is what made the company valuable to Nvidia. Carving out intellectual property to a level necessary for appeasing regulators ultimately proved impossible for Nvidia. This challenge was compounded by the size of the deal and SoftBank’s desire not to see the return on its investment diminish.

Regulatory resistance also highlighted industry opposition to the deal. Although Nvidia emphasized support from Broadcom, Marvell Technology and MediaTek, this was a drop in the ocean compared with the broader set of Arm users, many of which submitted evidence in opposition. Part of the challenge in this debate is the breadth and complexity of Arm’s ecosystem.

The huge number of uses for Arm technology — and the wide range of licensees with different business motivations — meant that Nvidia’s argument that it didn’t compete may have been true for one licensee but false for another. The overwhelming concern was that Nvidia’s business interests throughout artificial intelligence, automotive, edge computing and the Internet of things would inevitably clash with companies seeking to compete using Arm technology.

Arm and Nvidia both stated that the regulatory situation became more challenging further on. This is perhaps a generous interpretation. Although the regulatory environment has stiffened to high-value acquisitions, the winds of change were already blowing a gale in September 2020. The size of the deal and its strategic significance to Nvidia meant that it was worth pursuing despite the likely obstacles.

But the result was that Arm suffered almost 18 months of uncertainty and provided an opportunity for other architectures, notably RISC-V. The Nvidia deal was based on investment in Arm at a level that would dwarf that of a return to a public listing. Indeed, Arm CEO Simon Segars stated in a July 2021 blog post that “We contemplated an IPO but determined that the pressure to deliver short-term revenue growth and profitability would suffocate our ability to invest, expand, move fast and innovate”. His commitment to the Nvidia deal is likely to have been a factor in the appointing of Rene Haas to take on a new path to public ownership as CEO and rebuild the Arm vision with licensees. What this has meant in practice is over a year of underinvestment.

The upside of this is the unquestionable value that Arm still holds. The fact that Arm, SoftBank and Nvidia remained committed to the deal for so long highlights Arm’s value in areas spanning automotive, artificial intelligence, the Internet of things, data centres, smartphones and beyond. An IPO is the best path for Arm, providing that critical balance of investment and neutrality.

Nvidia may have lost the chance to own the CPU architecture that’ll power the next phase of computing, but this wasn’t without sizeable execution risk. The lack of movement in Nvidia’s share price following the news suggests investors are responding magnanimously. The big loser is SoftBank — the price tag was such a large premium that it left the Japanese firm with no plan B. Given the growth in Nvidia’s share price since September 2020, it’s difficult to see an IPO raising a comparable level of about $70 billion. This is compounded by the fact that Arm is still transitioning from the company SoftBank acquired in 2016.

With March 2023 targeted for an IPO, it’s still early days, and SoftBank’s determination to maximize Arm’s value could yet provide another twist to this tale. Regardless, the industry will be taking a salutary lesson — it can be under no illusion that the tide has definitively turned on the friction-free era of merger and acquisition.

A version of this article was first published by FierceElectronics on 9 February 2022.