UK’s National Semiconductor Strategy - a plan begins to form

When the company was trying to fight off a hostile acquisition by a temporary alliance of GEC and Siemens in the late 1980s, Plessey used its newly built semiconductor fab near Plymouth as one of the arguments to press politicians as to why it was important for the company to stay in one piece. Managers at the site pointed out the advanced features, for the time, that the fab employed and the care that went into its design. Even the site was carefully chosen for its geological properties: stable rock formations in a country not known for earthquakes, unlike the other epicentres of chipmaking at the time on the US west coast and in Japan.

The acquisition went through anyway and the Roborough fab did not remain under GEC management for very long. However, it is still running. After passing through numerous owners following the initial sale to Canadian telecom supplier Mitel in 1996, the Plessey name eventually returned to the fab, though this was a completely different company. Today, output from the plant has been turned over exclusively to production of microdisplays for Meta's virtual-reality business.

In the meantime, hardly anyone else considered Britain's geology to be a major pull when it came to deciding where to locate their fabs. Though the UK still has more than 20, they are small and dedicated to speciality processes, not the mainstream chips pumped out by the megafabs operated by Intel, Samsung and TSMC. Most of them are in earthquake-prone zones; but that still is not the problem.

Home to much of the chip capacity is Taiwan, but its proximity to an increasingly belligerent China has seen the governments of some western nations look at what it takes to finance chipmaking plants and consider the cost to be worth bringing at least some production home.

In the US, the CHIPS Act passed last year has pledged almost $280bn to semiconductor-related investments by the end of 2026, with just over $50bn of that earmarked specifically to support local manufacturing.

The EU is not that far behind with its €40bn-plus for manufacturing in addition to the funding for the various Horizon Europe projects already in place. The subsidies are immense.

The £1bn pledged by Prime Minister Rishi Sunak's administration for the UK's semiconductor industry in the National Semiconductor Strategy is a comparative drop in the ocean. "The headline figure is almost a negative if an investor compares this to what is happening elsewhere," says Fhaheen Khan, senior economist at Make UK. More important is the signal there is at least a change in direction in a country that few expect to compete directly in advanced semiconductor manufacturing.

"What we often find is that investors who might be willing to bring finance ask, 'where is the UK going?'," Khan says. "The important thing is to say that the UK is in the race. It at least provides some confidence to investors as to where they can put their money."

'What we find is that investors who might be willing to bring finance ask, 'where is the UK going?'. The important thing is to say that the UK is in the race. It at least provides some confidence to investors.'

With an election likely next year, a key risk is that there might be a further change in direction and that it would make more sense to align a plan for semiconductor with a wider industrial strategy. "We feel there should be a Royal Commission and that a semiconductor strategy would ideally be absorbed into that."

The political risk for a semiconductor strategy is perhaps lower than last time the country had one. By the time GEC and Siemens took aim at Plessey, a full decade had passed since the last national plan for semiconductors had been drawn up by a Labour administration, displaced soon after by a Conservative government that had little time for active industrial intervention. The main result of that strategy, Inmos, was ultimately absorbed by European chipmaker STMicroelectronics, but the project created a sizeable design cluster around Bristol as well as a fab that may yet find a place in the new strategy.

Sean Redmond, managing partner of Silicon Catalyst, a group that helps semiconductor start-ups obtain funding and support, points to the engineers trained in this cluster as having a major influence on the UK's chip-design sector, feeding into the early iterations of the Arm processor at Acorn Computers and the formation of Cambridge Silicon Radio, Element 14 and others.

"It isn't really identified by government as a tremendous success, but had we not had that strategy there we wouldn't have the same semiconductor design industry today, and the impact lives on. These policy interventions need to have a very long-term view," says Redmond.

In some ways, the UK's weak position in semiconductor manufacturing compared to the Far East, the US or even the EU means the country can easily hold back from being tempted to stump up the same kind of expensive subsidies. However, full details will have to wait until at least the autumn. A consortium led by the Cambridge-based Institute for Manufacturing (IfM), and which includes Silicon Catalyst, won a contract from the government to work out what options the government could pick to support R&D and small-medium enterprises working on chips.

One clear direction is a call for the creation of the Semiconductor Infrastructure Initiative, which will take responsibility for setting up a national institution for the sector. "The beating heart of the strategy is the firm announcement that we will be creating that body," says Redmond, who points to the important role research institutes such as CEA-Leti, Imec, the various Fraunhofer Institutes and Ireland's Tyndall Institute have played in EU countries. "Having something like that is a tremendous asset in a semiconductor strategy," he adds.

Actions that might be recommended by the IfM include building incubator facilities for early-stage start-ups, providing cheaper access to design tools and prototyping for compound semiconductors, silicon photonics and 3D packaging. Redmond points to Cambridge-based power-electronics specialist QPT as an example of how packaging for multiple chips using compound semiconductors as well as silicon has become a key specialism for some UK companies.

The report nods to the potential of expanding "open ecosystem" foundry activity in the UK. Companies such as IQE have called for the creation of a dedicated foundry for compound semiconductors, an area where British suppliers have succeeded in manufacturing locally, as well as design. Such a foundry would probably cost around £1bn to build and equip.

Because of the high upfront costs for manufacturing, foundry access may instead come through deals with offshore specialists, particularly for 3D packaging, where the UK has far less direct experience. But when it comes to packaging, most advanced plants are in the Far East, though the investments by TSMC and others will introduce similar production to the US.

A less immediately obvious factor lies in whether the workforce is ready. Across the Atlantic, the CHIPS Act recognises the big increase the US will need in engineering numbers for it to regain its former leading position in semiconductors. The act allocated $13bn to the National Science Foundation to spend on science and technology education. It also called for a report by the National Academies to study gaps in microelectronics teaching to school-aged children and to find ways to promote hands-on projects with electronics in competitions and various informal events.

Shortly ahead of the strategy's publication, the UK Electronic Skills Foundation (UKESF) issued its analysis of the skills shortage that already exists in Britain. This shortage is felt most acutely in chip design. According to UKESF, more than 80 per cent of companies involved in chip design, which covers many more sectors than photonics and compound semiconductors, had unfilled vacancies at the time of the report. "Our view is that we need to double the number of students on electrical and electronics engineering-related degree courses to close the gap," says Stewart Edmondson, CEO of UKESF.

So far, the strategy's approach to education rests predominantly on plans for STEM that are already in place, although the strategy looks to increase the number of PhD-level engineers through the EPSRC's Centres for Doctoral Training programme. Make UK's Khan points to aid for schools relying largely on the Levelling-Up Premium. "Quite a lot of the solutions highlighted by the strategy are not semiconductor specific," he says.

"The semiconductor skills pipeline needs to be coherent and joined-up, all the way from secondary school through to Centres for Doctoral Research. All elements in the pipeline need focus, attention and investment," Edmondson says. UKESF also proposes an industry-led initiative to form a skills academy that can deliver the additional training needed to provide recent graduates with the experience needed for industrial chip creation.

For once, however, there is at least a feeling that viable intervention may be forthcoming for the first time in decades.