The UK has a high value computing sector but its productivity growth has slowed since the financial crisis and it has few large companies. Consequently, a new government will need to develop policies to address the challenges it faces and sustain an area of strength in the UK economy, write Filip Mandys and Diane Coyle.
After the Second World War, the UK computing sector was at the frontier of technological innovation, pioneering the development of the first programmable digital computer and the first electronic-stored programme computer in the 1940s. Today, the sector – both hardware and software – remains an important contributor to new fields, such as artificial intelligence (AI) and quantum computing. It also has some areas of strength in manufacturing hardware, symbolised by the recent stock market flotation of Raspberry Pi. Yet despite its strengths, the sector now faces significant challenges, including shortages of skilled labour, intensifying global competition, and increasing cybersecurity threats. Furthermore, in addition to a significant trade deficit, the sector has experienced a considerable slowdown in productivity growth since the 2008 financial crisis, affecting both computer manufacturing and software development. This decline contrasts sharply with its high productivity growth through the early 2000s.
In a new report, we analyse the prospects for a sector generally considered to be an area of strength in the UK economy, but one displaying concerning areas of weakness. In 2022, the sector as a whole generated approximately ₤86 billion in gross value added (GVA) at current prices, amounting to 3.4% of GDP. The sector can be broadly categorised into hardware and software, with over 80% of the GVA attributed to software, particularly computer programming, but also including the games industry. The hardware subsector, primarily composed of computer wholesale, as well as computer manufacturing and repair, contributes significantly less to the overall GVA.
Figure 1: (a) total GVA and (b) productivity in the UK computer manufacturing subsector. Data is taken from INDSTAT, EU KLEMS, and DCMS.
The computer manufacturing subsector has, in fact, been shrinking in terms of its total GVA, and stagnating in terms of labour productivity (measured as GVA per employee). In 2020, it ranked 12th globally, and 16th in productivity (Figure 1). This puts the UK significantly behind leading countries such as South Korea, US, and Taiwan, as well as behind major European economies like Germany, France, and Italy in hardware manufacturing. Conversely, over the past decade, software GVA has grown at an annual rate of 7.3%, compared to 2.3% for hardware (Figure 2). Similarly, productivity in the software subsector grew 3.3% a year in comparison to 2.2% in hardware. A particularly productive activity is video games (Figure 3).
Figure 2: Growth in the total GVA of computing (a) hardware and (b) software. Data is taken from the DCMS.
The sector overall employed over 1.3 million people in 2023, and comprised about 140,000 active firms. Despite a 4.6% annual decline in the number of firms in recent years, employment in the sector has grown at a rate exceeding 7% per year. Consequently, UK computing firms are increasing somewhat in size, but the average firm still employs only 8.8 people and has an annual turnover of less than ₤100,000. The firms are concentrated in the south, where London and South East England in particular stand out for having a large number of computing and video game companies. But there are just a handful of large companies and many minnows that do not scale.
Figure 3: GVA per employee of computing (a) hardware and (b) software. Data is taken from the DCMS.
In terms of trade balance, the UK imports more computer goods and services than it exports, resulting in a net trade deficit of approximately ₤13 billion in 2021. However, the UK is a net exporter of computing services, with software constituting the majority of these exports. The gaming sector, in particular, is strongly export-oriented, with 45% of game firms’ turnover derived from international sales. Overall, the primary trade partners include the US, the Netherlands, and Germany (Figure 4).
Figure 4: Top five UK trade partners in the computing sector. Data taken from DCMS.
Opportunities and challenges
The UK does demonstrate a strong commitment to research and development, investing around ₤11 billion annually – nearly double the proportion of the GDP compared to the European Union (EU). There are strengths in semiconductor research and development and design, as well as in the fabrication of compound and legacy semiconductors. It is also a leader in quantum computing, with an established industry and supply chains. The UK ranks second globally in the number of quantum firms and private investment, and third in scientific outputs related to quantum computing. The UK is also a global AI leader on some metrics, having the third highest number of companies and amount of private investment. It ranks second in machine learning systems and AI breakthroughs (Figure 5). The UK’s video game sector is the largest in Europe, and the cloud computing industry is expanding rapidly.
Figure 5: Total number of significant machine learning systems by country, 2022. Data taken from Epoch.
There are several well-established tech clusters, each specialising in different activities. South West England is notable for photonics and electronics, while London is home to a significant number of video game firms. South Wales has compound semiconductor manufacturing, while the Cambridge cluster stands out as Europe’s largest tech cluster and a centre for electronics firms.
Figure 6: (a) number of semiconductor fabrication plants across regions 2022, and (b) semiconductor fabrication capacity by logic chip size across regions, 2019. There is a total of 157 300mm semiconductor fabrication plants, where the UK has one. Data from SEMI.
Despite their strengths, however, these clusters have experienced a slowdown in productivity growth and contribute relatively little GVA by international standards. In relation to hardware, the UK’s semiconductor fabrication facilities are well behind the frontier, with only one 300mm fabrication plant (Pragmatic), and no plants capable of producing advanced chips below 90nm (Figure 6). Additionally, the UK lags in computing power capacity (‘compute’), restricting researchers’ access to advanced computing resources. The UK’s most powerful system ranks only 39th globally and is over 60 times slower than the world’s most powerful system (Figure 7).
Figure 7: Ranking of the overall computing power (Rmax) by country. The Rmax score refers to the maximum achieved performance of the system under a specific benchmarking standard. Values of Rmax are in teraflops. Data from Top500.
In cloud computing, significant challenges include cybersecurity, data privacy issues, and the dominance of US service providers. Additionally, high equity stakes taken by some universities act as a disincentive for the commercialisation of novel AI research. The whole sector faces labour shortages and talent outflows, exacerbated by Brexit. Addressing these issues is crucial for sustaining and enhancing the UK’s position as a leader in software innovation.
Policy recommendations
The UK’s opportunities lie in leveraging its well-developed tech clusters and technological strengths, to promote commercialisation, enable companies to scale up, and increase exports and productivity growth. The next government will need to address the specific weaknesses, including barriers to commercialisation and scale-up, skill shortages, and the need for investment in computing.
A key focus should be on scaling up the average firm size to take advantage of economies of scale and enhance international competitiveness. This could be achieved through targeted policies, such as grants, regulatory reforms, and tax incentives. Additionally, the government should prioritise talent development to address the domestic shortage of skilled labour. Higher education institutions should be encouraged to expand and emphasise programmes in computer engineering, quantum computing, and AI.
The UK government should also prioritise the expansion and enhancement of the nation’s computing infrastructure. This includes increasing investments in semiconductor fabrication and constructing plants capable of producing more advanced chips. Similarly, the UK should invest in more high-performance computing systems to ensure that researchers have reliable access to advanced computing resources. Future investments should also focus on advancing 4G, 5G, and eventually 6G networks to maintain and improve connectivity and technological capabilities.
Furthermore, the UK should continue to fund research and support industry-academia collaborations, particularly building on its well-established tech clusters. Expanding research collaborations beyond Cambridge and London could foster the development of new tech clusters in other regions. Funding should prioritise research in existing areas of strength such as areas of AI, compound and advanced material semiconductors, chip design, and quantum computing.
The key point, however, is the need for a strategic and sustained approach, rather than a piecemeal set of policies, with a stable policy framework extending from compute and hardware manufacture through to software services, gaming, and a national data strategy consistent with growing a high value and export-oriented sector. The UK’s two main political parties have made some commitments during the election campaign: Labour is pledging to support investment in data centres and the creation of a National Data Library; the Conservatives plan to increase public spending on R&D by ₤2 billion and invest over ₤1.5 billion into the development of large-scale compute clusters. In a fast-moving area of technology, the next government will need to prioritise a strategy for the sector if the UK is to sustain one of its areas of economic strength.
Report: Competitiveness, productivity and innovation in the UK’s computing sector
The views and opinions expressed in this post are those of the author(s) and not necessarily those of the Bennett Institute for Public Policy.
