Leading GPU chip maker Nvidia is trying to acquire leading chip intellectual property (IP) licensor Arm, from Japanese conglomerate Softbank. For some background please see our posts, and here, and here, and here, this post in going to get deep in the weeds.
To start, we think it is safe to say that very few of Arm’s customers want to see Nvidia get its hands on Arm. Many of them already compete with Nvidia, and everyone wants a piece of the auto market. So having a competitor like Nvidia own the underlying engine of pretty much all our non-PC chips is going to be a problem. Despite that, very few companies have spoken out against the deal publicly, or from what we can tell, privately. Qualcomm is notable for being very vocal in its opposition, but many of the others seem content to participate in some sort of Silent Majority.
Their reasons for doing this vary. Many just do not like to engage in anti-trust processes for fear their comments will come back to haunt them when it’s their own turn. A fair number of large companies are just far enough away that they do not see themselves as having a dog in this fight. But this deal is different, with potentially far higher stakes, so it seems possible that something else is at work.
We have a few theories on this. First, there are a lot of companies that are just afraid. They worry what happens if Nvidia wins, and then the opponents turn up for their first post-closing licensing conversation. Qualcomm is large enough that they do not have to worry about this as much, but many others feel they do not have the same heft and so keep quiet.
Another possibility is that they did not wait for the deal to close to go renegotiate their Arm license. Sure would be a shame if something happened to your deal Mr. Arm, and I have my reservations about how your acquisition will impact our next license….We recently spoke with someone from a chip company and we joked about “Jensen [Nvidia’s CEO] calling your CEO”, and were met with an awkward silence. We strongly suspect that Nvidia is out there right now making phone calls to potential deal opponents with pricey Arm contracts up for renewal who might be persuaded one way or another.
But corporate views are just one part of the regulatory approval process, and the regulators often go their own way on these matters. There are three regulators that really have a voice here – the US, China and the UK (where Arm is based). The UK regulator recently launched a second review of the deal, rejecting many of the companies’ initial positions. From a political standpoint, we think the UK regulator’s primary concern are jobs in the UK, something that a promise of a new R&D lab and continued investment could quickly assuage. Beyond that, there are very few Arm customers who would be affected by this deal based in the UK. PRC regulators are sending mixed signals. They have accepted Nvidia’s application for review, which presumably was prompted by some sort of positive signal, but as far as we can tell, they have not actually set a timeline for that review. China’s regulators are notorious for sitting on deals until they are abandoned, saying no without actually saying no. The US has so far said nothing.
Our best guess is that US regulators are fielding a lot of inbound opinions on the deal, and these are probably mostly negative. By contrast, China’s regulators are probably getting much more mixed messages from their corporate constituents. Nvidia has been able to navigate China’s market largely unscathed avoiding the troubles of many other foreign chip companies there. We do not know how they managed this, but someone deserves a raise. Softbank also has had a mostly positive relationship with China’s government. If you want to go really far down a conspiracy rabbit hole, consider that the Chinese government appears to be in the process of acquiring stakes in its leading Internet companies, and that Softbank two weeks ago sold down a significant stake in its holding in Alibaba on the open market. For its part, Arm has done a credible job of building up a favorable presence in China. It has a lot of licensees there (but so does Risc V, a topic for another day), it has also done its fair share of investing in Chinese R&D.
Which leads us to Arm China. This is a whole other soap opera, and the topic for our next post, but the original intent of the Arm China joint venture (JV) was to create a Chinese entity that owned considerable IP. Today Arm owns 47% of Arm China, with the other 53% held by a collection of Chinese private equity funds. There is a lot of confusion about what exactly Arm China does. Corporate Arm says it collects about a third of its revenue from China, let’s say that’s roughly $1 billion. Does that mean Arm China does $1 billion in revenue? Maybe not. Many people we have spoken with say that Arm China focuses on emerging Chinese chip companies in IoT and embedded devices, and NOT Arm’s relationship with China’s large, established licensees (Huawei, ZTE, Xiaomi, BBK, etc.). We have no way of telling without Arm China disclosing its financials. So what would prompt them to disclose those numbers? Funny you should ask. Prior to the Nvidia deal, there was speculation that Softbank was planning an IPO for both global Arm and Arm China. That chatter died away, but with the deal potentially losing momentum as opposition mounts, the rumor about an Arm China IPO has resurfaced.
Could this IPO happen? Here it is important to consider the context. No one likes to walk away from $40 billion, so Softbank is unlikely to change course without reason. On the other hand, no one likes to ride a sinking ship all the way down. So should the Nvidia deal start to look unlikely, or should regulators start demanding unpalatable concessions, Softbank’s best interest might shift to consider its options. We can almost guarantee that some banker has a draft presentation of “alternatives” ready to fire off to Softbank. The world has also changed a lot since the deal was announced in September 2020. Crucially, at the time Softbank was reeling from a number of high-profile blow-ups in its private company portfolio, notably but (not exclusively) WeWork. Today, Softbank is benefiting from a number of well timed trades it did late last year and is under considerably less pressure from investors. Add to that, valuation multiples for semis companies in China are much higher than multiples in the US. So if Arm China is really doing a $1 billion in revenue (a big if), at 80%-90% gross margins, it could be valued north of $50 billion, possibly much more, and thus more than Nvidia is paying for all of Arm.
For their part, so far, Nvidia has shown nothing but resolute interest in closing this deal. They contain to apply pressure and persuasion wherever they can. Last weekend, Arm China essentially declared independence (see our upcoming post). This seems the textbook definition of an adverse material breach in the deal terms, but there was nary a whisper from Nvidia. That being said, we have never understood Nvidia’s rationale behind the deal. Arm is not a natural complement to Nvidia’s business. Nvidia already gets virtually everything it wants from Arm as a supplier. All that talk about access to new markets like AI and automotive are markets that Nvidia can already pursue without Arm. True, Arm brings Nvidia exposure to mobile, but Nvidia gave up on mobile almost a decade ago when they shuttered their Icera acquisition, and it is not exactly a growth market. And while Nvidia has the financial wherewithal to withstand Arm suffering a massive loss of customers post deal, this is exactly the situation that the regulators want to avoid.
Obviously, there is not a lot of solid information here, hence our title. That being said, there are a lot of forces at work here: vocal and possibly mounting customer complaints about the deal, rising regulatory scrutiny, China’s ambitions for its semis industry, control of the Arm China JV, and Softbank’s financial position to name just a few. It is not hard to construct scenarios from this where Softbank is better off from a broken Nvidia deal. Our view is that the big swing factor here will be US regulators and their view on the deal. We should probably get some hint of that soon.
If it were not for the fact that this is one of the most important strategic assets in semiconductors, we would find all of this highly entertaining.
Photo Credit; Blade Runner
Surprised you do not mention the EU. I think that’s the toughest regulatory hurdle.
Fair point, but they have not said much so far, and I wonder how much effort they will put into a non-EU company.
Pingback: Start Up No.1699: inside Apple’s design labs, the ransomware front company, Paul Dacre spills Ofcom fix, Arm’s likely future, and more | The Overspill: when there's more that I want to say·
Pingback: Start Up No.1699: inside – The Hypocrilist·