A new ‘King of Chips’ would be both exciting and terrifying
Posted on August 14, 2020
Adam Gibson, CEO of Konduit and VP of Skymind Global Ventures
It could be history in the making. Tech giant ARM, the world’s major supplier of microprocessor technology and one of SoftBank Group’s most treasured holdings, could have potential takeover interest from Nvidia, the largest graphics chipmaker on the planet.
If this happens, we will witness the world’s largest ever semiconductor deal.
SoftBank has been investigating exit options for part or all of its stake, which it paid $32 billion for in 2016 – the highest price ever paid for a European technology company. The British chip designer has only increased in value since, as its chips are part of nearly every major mobile processor from Apple to Huawei.
If the deal goes through, this would be a huge win for Nvidia – and it’s both exciting as well as potentially terrifying.
This deal would herald ‘a new king of chips’: until a few years ago, Nvidia’s reputation was for creating graphics units for gaming computers. The purchase of ARM, whose designs are used in billions of processors, would crown Nvidia as the top chip maker in America.
It goes without saying that the deal would boost Nvidia’s position in an almost unprecedented capacity. The company’s expansion has already seen it move from video games to AI and giant computer servers – but now, they would have the chance to compete with reigning semiconductor king Intel and other general–purpose CPU makers.
The deal could dramatically alter the semiconductor landscape and could have ripple effects throughout the chip industry, threatening its rivals: an ARM acquisition could enable NVIDIA to provide complete platform solutions along with its GPU technologies, from machine learning, AI, cloud and big data applications, to its own CPU offering as well.
How much should one company be allowed to own? Is this deal one step too far?
Currently, NVIDIA does not have much in the way of CPU design or mobile hardware beyond its Tegra line of mobile chipsets. But it does have dominance in hardware accelerators – its GPU strengthens the ability of edge computing technology to process terabytes of data per second.
This means it can facilitate support for a wider range of applications. In fact, last year’s Mobile World Congress (MWC) in Los Angeles saw NVIDIA present a range of initiatives aiming to cement its leader status in edge computing.
The deal could have negative effects on the ecosystem which are unlikely to be easily overlooked. This could be an overstep that lands the company in a mine of regulatory hurdles as it faces competition scrutiny. The Verge has pointed out that Nvidia is a customer of ARM – whose competitors also use ARM’s designs.
We think there’s a strong possibility that Apple and other ARM users might move away from ARM if NVIDIA owns it. One of ARM’s most attractive features is that its technology is used by all the major chip firms – but if a deal like this goes through, we believe it could be dangerous for players like Apple and Samsung who both deeply customise ARM processors.
For instance, the chips inside an Apple iPhone – like those in nearly each smartphone all over the globe – are based on the instruction-set architectures from ARM. That lost opportunity to bid on Apple’s iPhone business was one of the reasons Intel is thought to have lost the mobile market – alongside its fatal error to sell its ARM division and XScale processor.
Up until recently, Huawei – which has surpassed Samsung to become the world’s largest smartphone maker – was also using Arm chips but was forced to cease production of its custom Kirin ARM chips.
Let’s not forget that the deal might not close for a number of reasons: there is still the potential for there being different bidders. Also, Nvidia’s interest may not end in a deal, or the potential regulatory hurdles could prove too significant a risk to ignore. Lastly, SoftBank could choose to list the business instead.
Whatever the outcome, it will be interesting to see how everything unfolds.