(Seeking Alpha) – Nvidia (NVDA), the dominant force in GPUs, is supposedly in talks with Softbank to acquire ARM Holdings: the dominant force in smartphone and tablet chip-design.
In this article, I will provide my analysis of this proposed acquisition and share Nvidia’s logic for going after ARM. We will explore the deal from the perspective of Softbank and from the perspective of regulator scrutinizing this mega-deal.
Further, I will share a potential deal structure and estimate a price that Nvidia is likely to pay for ARM. Towards the end of this note, we will discuss the inherent risks of ARM being acquired by a semiconductor company like Nvidia.
Why Is Nvidia Interested In Buying ARM?
As most of you may know, more than 90% of global smartphones and tablets are powered by ARM-based chips. Low-power consuming chips and a neutral business model (ARM only sells IP) have led to ARM becoming the ubiquitous chip design provider for all portable device manufacturers.
Moreover, Apple’s recent decision to move its Mac computers away from Intel’s (INTC) chips to custom ARM-based chips could be an early sign of a long-term architectural shift towards ARM from x86 in the CPU market. Furthermore, ARM-based supercomputers are gaining traction in the data-center markets.
From Nvidia’s perspective, the acquisition of ARM makes a lot of sense for the following reasons:
Nvidia’s data-center ambitions have been boosted by recent additions of Mellanox and Cumulus Networks. However, adding ARM to its repertoire could propel Nvidia into an unassailable lead in the high-performance computing space in the data-center market.
Nvidia’s major competitors, Intel and AMD (AMD), seem to have mastered the synchronous functioning of their CPUs and GPUs for optimal performance. This is an area where Nvidia is lacking, but the addition of ARM could help Nvidia close this gap, thereby enabling it to make an entry into the CPU markets.
Nvidia’s presence in the smartphone and tablet market is limited to a few Tegra processors. If Nvidia adopts a different business model for ARM, it could suddenly become a giant player in the portable device markets.
In my opinion, buying ARM is an obvious move for Nvidia on its path to becoming a $1 trillion company, which its CEO, Jensen Huang has long touted as the company’s future destination. The deal prices being reported by media outlets are ranging from $32-52 billion. I think the lower end of that range would make it an excellent deal for Nvidia, but more on that later.
Why Is Softbank Selling ARM?
Well, the simple answer is to raise cash. However, when I looked under the hood, I found out that ARM hasn’t performed very well since Softbank took over in 2016. ARM’s core-business revenues are somewhat flattening, and its venture into IoT products and services (Softbank’s vision) has failed to get any market traction so far.
Softbank has attributed the faster rise in costs to higher R&D investments, but that does not explain flattening revenues. After having said that ARM is central to the center of Softbank’s vision fund, Masayoshi Son’s willingness to offload ARM Holdings is an indicator of trouble brewing at the company, likely begotten by poor leadership, as can often derail companies in this industry (Looking at you, Intel).
The official word from Softbank is that they are looking to sell ARM fully or partially to raise cash and would be open to taking the company public if a suitable buyer cannot be found. Now, according to a report, Softbank has already broken off two IoT services departments from ARM Holdings, and it plans to hold onto those companies in the future. In other words, Softbank is looking to sell only the legacy ARM business that they acquired in 2016. After recent mishaps like WeWork, Softbank’s CEO, Masayoshi Son, has been cornered by activist investors into share repurchases and debt reduction via the sale of stakes in Alibaba (BABA) and Sprint (TMUS). The next big bet to be liquidated (partially or fully) from Softbank’s portfolio appears to be ARM holdings.
Since Softbank is a neutral player in the semiconductor industry, it would try to squeeze out the maximum possible monetary value it could gain for ARM. In my opinion, the most likely outcome would be a majority (or complete) stake sale because any buyer (e.g., Nvidia or Qualcomm (QCOM) or Samsung) would want complete control over future business decisions.
Further, Nvidia buying ARM raises a ton of anti-trust issues, and regulators may never approve such a deal. However, for the sake of this article, let’s say the deal were to happen. What would the Nvidia-ARM acquisition deal structure look like, and what would be the deal value?
Nvidia’s ARM Acquisition Would Most Likely Be A Cash Plus Stock Deal
As I mentioned earlier, reported Nvidia-ARM deal values are varying from $32 billion to $52 billion. At this point in time, your guess is as good as mine. However, here’s my estimate of what Nvidia would be willing to pay for ARM:
According to Softbank Group’s estimates, its 75% ownership in ARM is worth $25 billion. And naturally, the remainder (25% stake owned by Softbank Vision Fund) is worth nearly $8 billion. Hence, Softbank’s estimated value of ARM holdings is $33 billion.
Some sources have said that Softbank would be willing to sell ARM at its acquisition price of $32 billion. This would indicate a no profit, no loss sale. However, after raising cash from Sprint and Alibaba stake sales, I feel Softbank is in a better position to negotiate a premium price. In all fairness, ARM Holdings could be worth a lot more to Nvidia. Hence, I am expecting a higher deal price.
Four years ago, Softbank acquired ARM for $32 billion at a price-to-sales multiple of ~20x. If the Nvidia-ARM acquisition deal were to be pegged at the same multiple, then the deal size would be ~$38 billion considering FY19 revenues of $1.898 billion. I expect Nvidia will need to cough up this amount if it wishes to snag ARM from Softbank.
Nvidia has rocketed to a level at which it competes to be the most valuable semiconductor company on the face of this planet by market cap at any given time. Hence, buying ARM in a multi-billion dollar deal is plausible for Nvidia. However, Nvidia’s current cash and debt position would certainly require the ARM acquisition deal to be a mix of cash and stock.
Nvidia’s debt-to-Ebitda ratio of 1.75x would need to be raised to 6-10x for completing an all-cash deal for ARM. I do not believe Nvidia’s management would want to do such a deal, and creditors would likely not let Nvidia stretch its debt covenants to such an extent. Hence, a cash plus stock deal is the only way for Nvidia to acquire ARM from Softbank. (Note: Softbank has announced its willingness to accept a cash plus stock deal for ARM.)
Fallacies Of The Nvidia-ARM Deal
Here are some of my primary concerns regarding this deal:
(1) ARM’s success is down to its neutrality towards all of its customers and that neutrality stems from its business model. If Nvidia buys ARM for $38 billion, I am sure they will want to alter ARM’s business model to gain a competitive advantage over rivals like Intel, AMD, and all other semiconductor companies (customers of ARM) from ARM’s innovations.
Any such unfair practices could result in ARM’s customers fleeing away to neutral competitors. Hence, Nvidia could end up losing revenues from ARM. However, on the flip side, Nvidia could develop new products powered by future generations of ARM technology and enter the CPU market to challenge the duopoly of AMD (AMD) and Intel (INTC). Moreover, a combination of ARM-Nvidia tech could become the dominant force in high-performance computing and data-centers.
(2) The deal is still at the negotiation stage, and nothing has been finalized yet. Hence, we can’t rule out the possibility of Softbank selling ARM to someone other than Nvidia (they have pitched the ARM opportunity to giants like Samsung, Qualcomm, TSMC (TSM), and Apple (AAPL)). Nvidia may yet decide against putting in a serious bid.
(3) Nvidia and Softbank could strike a deal; however, regulators are sure to have a word or two to say about this deal. Since ARM serves almost the entire semiconductor industry (including Nvidia’s rivals – Intel and AMD), Nvidia buying ARM is a hornet’s nest of anti-trust concerns. To be fair, I can’t imagine this deal going through; however, if it were to be approved, several stipulations would block Nvidia from using ARM as it would like to, the reality of which might reduce the logic of the acquisition for Nvidia.
Conclusion
Now, the proposed Nvidia-ARM deal could fail for several reasons. However, if the deal goes through, I feel that Nvidia could become the most dominant semiconductor company in the world, which would thereby more sufficiently justify its market valuation, which I view as extremely stretched at the moment.
