Tech Highlights From Around the Globe – Focusing on Semiconductors and AI

Article by Ralitsa Stancheva

Introduction

Tech is one of the fastest-moving industries worldwide and the UK is certainly no exception. Whilst the UK’s tech sector does not function in isolation from international events, it is important to consider domestic developments since “supporting innovative businesses” has been announced to form part of the government’s new legislative agenda for economic growth. This article aims to provide an overview of both aspects by focusing on two strategic areas of the UK’s International Technology Strategy (UKITS) – semiconductors and AI.      

The art of risk management – semiconductors and microchips

Semiconductors have been included in the UK’s International Technology Strategy, published in March 2023, as a priority technology. Their role is seen as critical for ensuring the continued supply of electronic devices and realising the UK’s strategic advantage in other technological focus areas such as net-zero, 5G/6G, quantum computing, AI, space technologies, and defence equipment. The strategy recognises the highly complex and interconnected supply chain landscape required to produce chips for consumer, industrial, military, and other applications. It also necessitates close international cooperation between governments and companies for the successful mitigation of geopolitical risks. 

The government’s Fifth Report on the UK semiconductors industry characterises the sector as comparatively small in relation to the US and Asia. UK products have global reach due to areas such as design, intellectual property as well as compound and advanced materials. The major vulnerabilities arise from the lack of a national end-to-end supply chain and the dependencies on supplies from Taiwan. This exposes UK companies to the risk of potential disruptions, for example, as seen during the pandemic or in case of a potential invasion by China. One such risk was avoided in November 2022 when the government blocked the takeover of the country’s largest chip manufacturing facility by Nexperia, a Chinese-owned entity. 

The wider international context     

How the UK’s strategy will interlock with Vietnam’s establishing itself as a semiconductor hub remains to be seen. At present, Vietnam is the third most significant chip supplier for the US following Malaysia and Taiwan. In August 2022, Samsung Electronics, the world’s largest memory chipmaker, announced its intention to invest $ 850 million in local production facilities, thereby making Vietnam its fourth source of semiconductor components after the US, South Korea, and China. 

In September 2023, the US announced steps towards shifting parts of American chip companies’, such as Intel, Synopsys and Amkor, supply chains to Vietnam. Vietnam is also focused on building up semiconductor clusters as their lack of this technology has been identified as the major roadblock to attracting even greater foreign direct investments. This will include training 50,000 engineers until 2030 to compensate for the present shortage of skilled workforce. 

It is unclear, however, to what extent Vietnam has already successfully position itself as the hub of choice in the industry. It became known that Intel has abandoned plans to invest nearly USD 1 billion in the country to expand its existing local facility of 2,800 workers for the assembling, packaging, and testing of chips. Despite the absence of an official statement from the company, some media reports speculate that the decision could have been taken in favour of Singapore or Malaysia as alternative locations for establishing manufacturing facilities. It has also been suggested that during meetings with Vietnamese government officials, Intel had expressed concerns over the feasibility of its investment given the country’s heavy-duty bureaucracy and the ongoing blackouts that hinder the reliable supply of power. 

Earlier this year, the UK has experienced a similar setback in its efforts to attract foreign direct investments. In June, Intel announced its plan to build two factories in Magdeburg, Germany and create 10,000 new jobs. In 2021, the company stated that it can no longer consider Britain as a possible production site due to the aftermath of Brexit and will limit its short list to members of the European Union. With this move, Intel aims to capitalise on the possibilities for growth in the internal market including a multi-billion subsidy by the German government, subject to approval by the European Commission. Other plants will be built in Poland, Ireland, Spain, Italy, and France.

Yet, the connection between Intel and the UK does not end here. An October announcement by Nvidia and AMD for the release of computer chips based on Arm’s intellectual property by 2025 has caused fluctuations in Intel’s stock price. Apple has also recently transitioned from Intel to Arm processors to boost PC computer battery life. A similar shift may follow for Microsoft devices according to alleged plans of AMD to deploy Arm chips for the consumer PC segment. The move has been seen by market analysts as a direct challenge to Intel’s leading position in the PC chips segment. Intel’s CEO Pat Gelsinger commented that these alternatives have “pretty insignificant roles in the PC business”. He emphasized the strength of Intel’s roadmap, including due to the potential of AI.     

The dawn of a new regulatory and research era – frontier AI safety

According to UKITS, the key benefits of AI are unlikely to be realised without addressing the barriers to ethical AI. These barriers include the limited data quality and availability; shortage of skilled workforce; public mistrust; market disincentives and data monopolies, increasingly, the lack of coordinated governance, clear regulatory standards, and quality assurance has also been cited as a barrier. In parallel, the Stanford University’s 2021 AI Index Report ranked the UK as the country with the third highest private investment in AI worldwide after the US and China totalling $1.9 billion. Against this background and in the context of the UK’s 2021 enhanced trilateral security partnership with the US and Australia, overcoming the identified barriers becomes paramount for the successful adoption of AI technologies. 

As a step towards realising this AI strategy, the UK hosted the AI Safety Summit 2023 in early November. The summit has brought together key stakeholders in the process of developing and implementing AI technologies, such as national governments, technology companies, researchers, and civil society groups, to discuss the risks of AI and their mitigation at international level. Following the summit, Downing Street announced the launch of the world’s first AI Safety Institute that will coordinate cross-border efforts in the development and pre- and post-release testing of the safety of frontier AI. The institute is the successor of the Frontier AI Taskforce that has been tasked with identifying and estimating the risks inherent to frontier AI models. The participants in this initiative range from the US AI Safety Institute and the government of Singapore, to the national data science organisation Alan Turing Institute and       industry heavyweights like Google DeepMind, Microsoft, and OpenAI. The risks in scope include bias, misinformation, losing control of AI, unexpectedly rapid advances in AI models, and even the “humanity’s extinction” as warned by Elon Musk. 

To power its research and computing capacity, the Institute will be granted access to the newly created AI Research Resource (AIRR), a network of Europe’s supercomputers at an estimated worth of $366 million. The UK supercomputer to become part of this network will be created by Hewlett Packard Enterprise at the University of Bristol by next summer. Named Isambard-AI, it will incorporate over 5,000 of the latest Nvidia GH200 superchips. This will translate into a capability to complete up to 200 quadrillion calculations per second, thereby ranking the technology within the top ten fastest supercomputers and open-science system worldwide. Isambard-AI will also be used for the purposes of robotics, big data, climate research, training large language models ad AI-based drug discovery. 

Another academic institution and AIRR network member, the University of Cambridge, will connect its own cluster of supercomputers called Dawn with the Isambard-AI. Once Isambard-AI becomes operational, it is expected to surpass the computing power of Dawn as the UK’s current fastest supercomputer by a factor of ten. This will mark a new milestone is the marrying of AI and high-performance computing (HPC). The first phase of Dawn has been deployed in early November following a successful co-design collaboration between the University of Cambridge, the UK Atomic Energy Authority (UKAEA), UK Research and Innovation (UKRI) and the US industry giants Dell Technologies and Intel. 

Dawn’s capabilities find application not only in academic research but also in various scientific fields including fusion energy development and climate modelling, particle physics and engineering, cosmology and healthcare. The second phase of the supercomputer will have increased performance capabilities and is expected in the course of 2024. Dell sees the collaboration as vital for unlocking “the high-growth AI potential of the UK” and encourages the UK government to invest in the technologies and infrastructure that can support this trajectory. In the field of sustainability, the University of Southampton received a USD 18 million government grant for the creation of an AI centre and training over 70 PhD students.      

These developments and concerns over the safety of AI could hardly be timelier in view of concurrent events around the globe. In the UK, Addenbrooke’s Hospital in Cambridge has announced the introduction of AI into one of the relatively common procedures for liver cancer treatment to increase the diagnostics accuracy. In other European countries, the technology has already been added to an identical procedure for the treatment of lung and kidney tumours. 

Conclusion 

Both the semiconductor and AI fields continue moving at a higher-than-average pace. The strategic implications of the developments in the UK and at international level are merely beginning to take shape as national governments and major market players regroup and increase investments in research and new product lines. Following the related development, therefore, remains paramount.