Nvidia acquisition of ARM is facing challenges from the UK and Europe.
Even if the UK and Europe clear the deal, there are no signs that China is even willing to take up the deal.
It is nearly certain that Nvidia will not be able to close the deal and will have to pay SoftBank a break free and/or restructure the deal.
ARM aside, Nvidia’s prospects are shaky.
Looking for a portfolio of ideas like this one? Members of Beyond The Hype get exclusive access to our model portfolio. Learn More »
Justin Sullivan/Getty Images News It has been over a year now since Nvidia ( NVDA ) announced its intended acquisition of ARM. There were rumors floating prior to the intended acquisition announcement that Nvidia was about to make such a move. Over a month prior to the announcement, we wrote an article about why the acquisition was unlikely to occur and an alternate path Nvidia could take instead of the acquisition. (see ” Nvidia Is Likely To Make A Strategic Investment In ARM Instead Of Buying It Outright ” ). Of course, the prognostication appeared inaccurate when Nvidia went on to make the acquisition announcement. Now that more than a year has passed since the announcement and with regulators in no hurry to approve the deal, is Nvidia’s acquisition of ARM likely, or were we correct with our previous assessment that Nvidia is likely to make a strategic investment instead of buying ARM outright?
This article reviews various issues surrounding the acquisition and the likely outcome.
In evaluating the ARM acquisition effort, investors need to understand Nvidia’s current business model and its key markets. The two most important markets for the Company, revenue-wise, are its client GPUs which are used for gaming and cryptocurrency mining, and server-side GPUs which are used for Machine Learning and professional content development.
On the client front, Nvidia has been the lead dog for nearly two decades and had very little competition at the high end for nearly a decade. However, this has been changing rapidly with long-time rival Advanced Micro Devices ( AMD ) becoming an increasing threat at the high-end of the GPU market and with Intel ( INTC ) about to get into the discrete GPU market after many prior failed attempts. This threat increasingly puts Nvidia’s client revenue stream at risk because, among other tactics, both Intel and AMD have their own CPUs and can bundle their GPUs with CPUs. Nvidia needs its own x86 CPU to counter this threat from Intel and AMD, but getting access to strong x86 IP is not an option. The next best CPU option, ARM, is officially supported by Microsoft ( MSFT ) for Windows and Google ( GOOG ) ( GOOGL ) for Chromebook. Given this dynamic, ARM becomes the default choice for Nvidia in the client space.
Nvidia is also witnessing intensifying on the server front. Especially in the critical data center market, in addition to competition from Intel, AMD, and several startups, many cloud service providers are starting to deploy their own CPUs, GPGPUs, and AI accelerators. Nvidia has now lost several high-profile HPC accelerator designs to AMD and Intel as it cannot effectively mate its GPUs with Intel and AMD CPUs for these applications. To compete well in the future, Nvidia needs its own CPU. Again, given the lack of access to x86 IP, ARM becomes the de-facto CPU choice. This is the market dynamic that makes Nvidia interested in ARM in the data center.
Now that ARM’s need is established, the question is if Nvidia should acquire ARM to control its CPU destiny and make it difficult for other server CPU aspirants to compete. Licensing ARM would be easy and Nvidia can license the appropriate cores for its use. But licensing is limited in its use. ARM will advance its CPU development based on its view of market needs and not necessarily what Nvidia wants. Also, other competitors will have access to the same ARM IP at the same time as Nvidia and that makes it difficult for Nvidia to differentiate.
Nvidia also gets other benefits by owning ARM. Given ARM’s widespread use, Nvidia could benefit in promulgating Nvidia GPU IP alongside ARM IP. ARM’s distribution channel could get Nvidia GPU IP into several segments of the market including servers, networking, mobile phone, IOT, etc. Having it’s GPU IP gain widespread use has strategic value to Nvidia as it can target a much larger application space and motivate many more developers […]