On the same day England’s greediest football clubs were brought to heel by the threat of government intervention against their European Super League cartel, the Culture Secretary Oliver Dowden blocked the planned sale of Cambridgeshire silicon chip designer Arm to US chip manufacturer Nvidia. Unsurprisingly, the football story overshadowed a squabble over the arcane business of semiconductors, but it represents another dramatic break with past economic policy, and one with global implications as the trade and tech war between the US and China builds once more.
Arm is a novelty: a homegrown British technology success story that is now a genuine world leader. A spin-off from Acorn Computers, whose BBC Micro became a familiar sight in 1980s schools as a result of a Thatcher government purchasing scheme, Arm was an early designer of innovative, low-power consumption silicon chips. Its business model was a similarly radical departure from industry conventional wisdom: eschewing the construction of silicon fabrication plants (a phenomenally expensive and complex undertaking), Arm proposed not to make chips, but only to sell their designs, for a royalty fee. The gamble paid off spectacularly, with low-power chip architecture proving critical to the development of mobile phone technology. Arm chip designs now sit in more than 95 per cent of smartphones worldwide.
Those same chips are now appearing in a growing range of devices as the “Internet of Things” becomes a reality. The number of internet-enabled devices is expected to rise from more than 15 billion today to 75 billion by 2025, creating the tantalising prospect of huge and profitable new sales for Arm’s designs. Japanese company SoftBank, Arm’s current owner, cited this as a primary motivation for buying the company in 2016. If SoftBank has a business model, it is to get in early on businesses with (often tenuous) prospects for radical expansion.
The US firm Nvidia, however, would be a very different sort of owner. The company has expanded from a leading producer of graphics processors to a supplier of high-tech equipment for machine learning. The kind of data-intensive processing needed to produce pixel-perfect computer graphics is similar to the kind of data-intensive processing needed to, for example, persuade a computer to learn a natural language. Nvidia has branched out rapidly from its product niche into becoming one of the world’s leading suppliers of AI processors. The acquisition of Arm would mean the closer integration of Arm’s technologies into its own offer.
This has provoked some concern within the industry: the production of customised processors, of the kind Arm specialises in, means working closely with the purchaser to meet precise specifications and delicately balancing customer requirements with the current possibilities of chip fabrication technology. As a result, alongside its design expertise, Arm has acquired a deep knowledge of the wider industry. That knowledge would transfer to Nvidia in the event of a sale; as a chip manufacturer, Nvidia would therefore gain an edge on its rivals. As Arm’s co-founder Hermann Hauser has argued, Nvidia has every incentive to keep Arm’s intellectual property to itself, using it to lever an advantage over its competitors but blocking Arm’s own potential for growth. This quasi-monopoly potential has fired up competition investigations across three continents, with US, European and Chinese regulators launching inquiries.
But the prospect of a dramatic expansion of data technology – from its collection via smart devices, through its transmission on 5G, and then its processing in AI – is helping push the US and China into an increasingly tense economic conflict. The tit-for-tat tariff and export control battle between the two has resulted in around $450bn of bilateral trade being subject to tariffs since 2018. The US has also extended export bans across the sector, largely motivated by the desire to prevent Chinese telecoms specialist Huawei accessing the software and hardware it needs to manufacture 5G base stations.
China, for its part, is investing heavily in domestic chip production as a core part of its “Made in China 2025” strategy to become the global leader in high-tech manufacturing. With the world’s leading semiconductor manufacturer, TSMC, based in Taiwan – claimed by China since 1949, but disputed by Taiwan’s government – and the steady build-up of competing naval hardware in the South China Sea (Germany being the latest country to dispatch a warship), the economic tensions sit dangerously close to existing military and diplomatic livewires.
Arm’s Chinese subsidiary, part-owned – as a condition of operation in the country – by Chinese shareholders, is currently in the middle of a messy internal power struggle, its CEO refusing to leave office despite a 7-1 board vote to remove him. Arm was a major supplier to Huawei, causing some jitters about its ongoing access to US markets, but has recently confirmed it can continue to sell designs to the company, suggesting some of the “security concerns” cited by Dowden may be less pressing than he implied. More likely is that the UK government wants to use the block on selling Arm as a negotiating tactic. China’s government, for its part, is said to be strongly opposed to the Nvidia deal as potentially allowing the US to restrict access to Arm’s technology.
Arm is a product of an earlier age – a “Switzerland” (in Hauser’s phrase), neutrally selling its chip designs to whoever wants them. Its business model is predicated on being able to do that across the globe, without significant government interference. But in a world of great power conflicts over the control of technology, that position puts Arm close to getting caught in the crossfire. Neoliberal neutrality is no longer viable as a business proposition in a world of economic nationalism.
Active protection is needed, in the broader public interest. Arm could be the key link in an industrial strategy for semiconductors and computing more generally, with the government insisting, for example, that the company use its own market power as a purchaser to build up local and regional suppliers, building new centres of expertise across the UK, and using government spending to help create new markets for its own products on the other side, just as Margaret Thatcher did for Acorn’s BBC Micro, 40 years ago. It could insist on retaining ownership in Britain, including a government-held stake, giving it crucial leverage in future. The status quo isn’t going to work for Arm any longer. It’s time for the government to step up, or risk losing a critical national asset.