The government's implicit acceptance of semiconductor irrelevance

If industrial policy is uninspired, it might as well be cheap

Photo by Anne Nygård on Unsplash

After an extensive delay, the British semiconductor strategy has been unveiled. It has inevitably disappointed onlookers with the paltry size of its funding. Executives of Graphcore, an AI chip designer, called the commitments modest, while compound semiconductor developer Paragraf’s CEO went a bit further and called the strategy ‘flaccid’. It truly is a disappointing semi-strategy of Tim Henman-esque proportions.

£1 billion over ten years, or £100 million annually. The U.S. CHIPS Act provides $280 billion over 5 years. Intel is receiving potentially €7 billion from the German government to build a fabrication plant.

Of course, Britain does not have the money of these countries, but it is a tiny sum even when factoring in our limited role within semiconductors. Compare with British spending on other things. £4.5bn in Covid-19 support has been lost to error and fraud since 2020. The UK is set to spend over £15 billion on aid and development from 2022 to 2024. The annual cost of the asylum system is officially £1.5 billion. This twenty-year Semi strategy is even more threadbare than the projected £1.7 billion in exports resulting from joining the Comprehensive and Progressive Agreement for Trans-Pacific Partnership (CPTPP). The committed spending over a decade is about as much as annual losses to warehouse theft and half the annual losses in retail revenue from shoplifting.

Given the continual waste permitted by HMT throughout the government, it is a bit rich to argue the figure has been based on a rigorous devotion to return on investment. This is a portion of money that sounds like a lot but is ultimately trivial.

The document is most notable in that it avoids interrogating why Britain is such a small player in semiconductors, or indeed in most industries that are intensive in capital, brainpower, and energy. A large university sector means little when there is no lab space for commercially minded graduates to spill into. Energy costs are higher than anywhere else. Semiconductors require cheap electricity but also extensive water infrastructure. TSMC consumes 10% of Taiwan’s water annually. UK industries, particularly food & drink, and chemicals, are already facing demands to reduce water consumption as crumbling infrastructure strains supply. The UK’s freshwater resources per capita are about half that of Taiwan. Our economic fundamentals preclude any serious industrial strategy, especially in something as hard as semiconductors. Aware of this, the document makes nearly no mention of building factories, lowering electricity costs, reforming planning, or reducing tax on capital expenditure.

Only an industrial policy that answered these problems could be successful, and the modern British state has given up trying to solve them. The country’s political economy of cheap borrowing to finance chronic current account deficits was nearly turned upside down when Lizz Truss tried to issue some peripheral tax cuts. Fixing energy costs is beyond such a fragile system.

The strategy is at least honest, ignoring all pretenses that the UK can build modern fabrication facilities. This is a limited strategy aimed at juicing R&D, funding certain types of compound semiconductors, and greasing the wheels of ‘international cooperation’ and ‘skills’. It is, to all intents and purposes, a surrender.

Could it have been anything else? Possibly not. Professor Richard Jones is one of the foremost authorities on British technology policy. His analysis of the British semiconductor space settled on tinkering with procurement to get supercomputers to use British-designed accelerators, and getting government control of the semiconductor architecture designer ARM. The former is not emphasized and the latter is not considered. Buying ARM would show intent, and the investment could be recouped. But it would require concentration and long-term planning. There is reasonable skepticism that such a company could flourish under government control. Ultimately, even for well-informed people, the prospects of the UK semiconductor industry are limited.

My overriding thought is this. No industrial strategy is probably not an option, but if the industrial strategy is going to be so lackluster, it might as well be this cheap. Our policy documents, whether it be this or the risible 10-point plan for Net Zero, fails in two ways. It does not tackle the structural issues that make Britain such a lousy place to do productive capital-intensive industries (housing, planning, energy, capital spending). Secondly, all the focus is on established opportunities where larger powers are already siphoning billions in subsidies. This process of ignoring systemic weaknesses and shadowing other nations’ industrial strategies is guaranteed to disappoint people.

There will likely be a Labour government soon. Their policy whitepapers will be longer and more detailed. They will probably get some more approval from industry leaders, and their cheques will be larger. But were I to be advising them, I would stress this; meta-problems or industries no one is paying attention to. No more strategies focused on Net Zero, electric cars, or semiconductors.