It may have cut its teeth on building specialised chips to accelerate gaming, but chipmaker Nvidia has now flagged artificial intelligence (AI) as a major growth area – and put its money where its mouth is by making one of the tech industry’s largest-ever acquisitions.
Nvidia’s $54 billion ($US40b) acquisition of mobile and embedded-chip giant Arm Holdings from Japanese investment giant SoftBank Group – which bought Arm in 2016 for $42 billion ($US31.4b) – will give it a massive base of expertise in the fast-growing mobile market, where the need to balance power and performance has driven extreme innovation.
Cambridge, UK-based Arm designs and licenses chips for embedded systems – such as set-top boxes, Internet of Things (IoT) devices and gaming consoles – as well as working with a range of vendors that draw on its expertise to deliver all kinds of smartphones and tablets.
Nvidia, which has sidestepped once-dominant Intel by harnessing the superlative processing power of gaming-focused graphics processing units (GPUs), will gain access to Arm’s broad range of reference designs and expertise – and its top-shelf client list.
Arm earns unit-based royalties from licensing its chip designs to companies including Samsung, Qualcomm, VMware, Amazon Web Services, Broadcom, Dolby and many others – and Nvidia has pledged to continue licensing to third parties after the acquisition is complete.
Arm reference designs power over 180 billion devices and tens of billions more ship every year, including Fujitsu’s Fugaku – the fastest supercomputer in the world upon its launch in June – and ‘Bionic’ processors that Apple has been co-developing with the company for many years.
Arm-based chips have gotten so fast that Apple is ditching 15-year partner Intel, announcing in June that it will transition its desktop iMacs and macOS operating system to run on Bionic processors that are also compatible with iPhone and iPad applications.
Shaping the age of AI
The move is a significant step forward for development of purpose-built AI chips that will, Nvidia said in an investor presentation, expand its reach to 15 million developers working to integrate learning capabilities into cloud platforms, proprietary computer systems, and connected and standalone devices including self-driving cars, robotic automation tools, and more.
“In the years ahead, trillions of computers running AI will create a new internet-of-things that is thousands of times larger than today’s internet-of-people,” Nvidia founder and CEO Jensen Huang said in announcing an acquisition that he said “will create a company fabulously positioned for the age of AI”.
Supporting its vision of the future, Nvidia – which spends 24 per cent of its revenues on R&D – has committed to building a “state-of-the-art AI supercomputer” running ARM processors.
In the long term, the integration of increasingly powerful chips will enable the integration of AI – a highly compute-intensive task due to the volume of data involved – will reshape the functioning of all manner of computing devices large and small, with Nvidia’s AI capabilities integrated into Arm’s own product portfolio.
AI’s strategic importance has driven a new ‘space race’, and US-based Nvidia’s purchase of Arm is being seen as an offensive move against China – which has poured money into the technology as controversial tech company Huawei doubles down on AI and Chinese AI chipmaker Cambricon prepares for an initial public offering.