How geopolitical strategy decisions impact industry, trade, technology and research
A US$40-billion acquisition of ARM – the UK company which supplies design and intellectual property to the semiconductor industry – by Nvidia, a famous GPU (graphic processing unit) manufacturer, has sent seismic tremors throughout the technology industry. The deal must be examined through the geopolitical lens, as much as through the technology prism. In fact, technology has never been as central to the ongoing global trade war and the Nvidia-ARM deal creates an entirely new dimension to the battle for technological supremacy between the US and China. ARM’s design architecture is a global standard for mobile devices.
This is a story of powerplay between global tech giants and how political forces drive them across borders. Gone are the days when technology specialists could concentrate on research and innovation, unmindful of political tensions. Now, every major political decision has far reaching impacts on industry, trade, technology and research. More so for the IT industry, because of its multinational nature. It is, therefore, prudent for every aspiring technology professional to keep abreast of the winds that can swing the turn of events to disrupt research, jobs and careers.
Semiconductors or chips are the brains ofany computing device – from laptops to massive Cloud servers, Edge computing, artificial intelligence (AI), mobile phones, telecommunication networks, consumer electronics, the Internet of Things, 5G and absolutely everything which needs electronic processing. The acquisition gives unequalled global technological dominance to Nvidia, the US$12 billion technology giant that invented the GPU and dedicated high-performance chips for the gaming and professional markets, as well as chip units for the mobile computing and automotive markets.
The Technology War
Currently, a technology war is going on between the US and China to control the semiconductor industry. From midnight of September 14, 2020, the Chinese telecom giant, Huawei has been cut off from essential supplies of semiconductors, creating an existential crisis for this flagship company. Without chips, it cannot make the smartphones or mobile-network gear on which its business depends.
America’s latest rules, finalized on August 17, 2020, prohibit companies worldwide from selling chips to Huawei if they have been made with American chipmaking kit. Semiconductors, in turn, depend on ARM design which enjoys near total global supremacy in this business. The US controls both the machines which make the chips and the largest company which creates the designs for the chips.
Semiconductor manufacturers license ARM’s architecture, or “instruction sets”, which determine how processors handle commands. This option gives chip-makers greater freedom to customise their own designs. In other cases, manufacturers license ARM’s processor core designs – which describes how the chips’ transistors should be arranged. These blueprints still need to be combined with other elements – such as memory and radios – to create what is referred to as a system-on-chip. ARM has over 500 licensees to its technology, and those licensees can tweak the designs and customise them for their own products and customers. Among the biggest of its customers are Huawei Technologies and Qualcomm Inc.
An Overwhelming Market Dominance
ARM has an overwhelming sway over the mobile application processor market and it is expected to increase exponentially over the next few years. According to Statista, ARM is expecting to grab over 90% market share within both the mobile application processor, and the embedded computing markets worldwide by 2028. It has also set a target of securing greater than 25% market share of the data centre and Cloud markets – a dramatic increase from the 4% market share they had in 2018.
It is expected that 72% of the global enterprise server market will be using ARM-based servers by the end of the decade, with notable revenue growth expected within matrix workloads – those which require large amounts of data to be rapidly processed. By 2030, it is forecast that ARM-based processor revenue for matrix workloads within the enterprise server market will reach US$35 billion, a dramatic increase from US$100 million in 2019, as per Statista.
ARM has also designed specialised AI processing chips and an AI computing platform for developing and running industrial AI applications. Although both Nvidia and ARM have fingers in many pies, they have market dominance in their own particular field of specialisation – GPUs for Nvidia and mobile processors for ARM; the most obvious crossover in what they do is in the field of AI. Combining these strengths creates a partnership that is well-placed to lead in a world where smart, connected devices are becoming increasingly important to our lives; in fact, it has the potential to reshape our lives.
Giving Nvidia the ‘Edge’
ARM’s CPUs are already present in billions of devices worldwide, thanks to the growth of mobile and IoT. Strategically, this makes them a great acquisition for Nvidia, as a vehicle for rolling out its deep learning technology to “the Edge.” This is the fast-growing share of compute power dedicated to analysing and interpreting data onboard the actual devices that capture it, rather than sending it to the Cloud for processing by remote computers.
Over the next few years, moving compute workloads to the Edge is expected to lead to increased speed, energy savings, and security improvements across all processes and operations that are driven by data collection and analytics. In short, a merger of resources and capabilities between Nvidia and ARM makes them uniquely positioned to capitalise on what will be the most important and impactful tech trends of the next decade. Much of this may be transparent to us, as end-users, who, unless we want to build a high-specification gaming PC, might never buy a product or service marketed directly by either brand.
The China Angle
ARM has a lot of customers in China – the entire Chinese mobile industry, in fact. It was a British company, then owned by SoftBank of Japan which has now sold its stake to Nvidia. China isn’t really in an active trade war with ARM. But Nvidia is an American company, and the US fired the first shots in a mobile-tech trade war by preventing Huawei from using Google’s mobile software and services. This merger would need to be approved by regulators in every country where ARM wants to do business.
The Nvidia-ARM deal poses a huge challenge to some of the largest providers of CPUs to Cloud data centre companies like Intel and AMD. The bulk of the world’s machine learning workload so far has been carried out using Intel or AMD processors, often working alongside Nvidia GPUs in the Cloud. This includes the core business functions carried out by web giants such as Amazon (shopping), Google (searching), and Facebook (socialising) as well as that carried out on their platforms by third-party service providers such as Netflix or Uber.
With the move to Edge and mobile, where ARM is the undisputed leader, this is likely to change. More data collection and analytics can be carried out directly on your phone, such as the personal data-driven calculations carried out by Amazon to work out what to try and sell you, or those carried out by Netflix or Spotify to predict what you want to watch next. This could lead to greater levels of personalisation and more useful predictions – with no need for the tech companies carrying out the work to even see the data that the decisions are based on.
The new owners of ARM hopes to avoid getting caught the in the technology war between the two superpowers by claiming that the IP of ARM was originated, created, developed over three decades in Cambridge, and hence the amount of code, the amount of innovation measured in thousands of human years and the IP will essentially stay in the UK, as the headquarters of ARM will be in the UK. However, it remains to be seen whether China appreciates this view or has other ideas. Whatever be the case, the Nvidia-ARM deal will transform the technology industry in many ways.