💾 Mann Virdee, Head of Science and Technology
The year is 1978. The TV show Dallas has premiered on CBS. Over on BBC1, Grange Hill has started, and Kate Bush has released her debut single, Wuthering Heights. The UK Government under prime minister Jim Callaghan has just put £50 million into INMOS, a new semiconductor company in Bristol.
This is the UK’s industrial strategy. Public money, channelled through the recently established National Enterprise Board, to build a British chip industry from scratch. As with most chipmakers of its time, INMOS made much of its early money manufacturing memory.
But it didn’t last. INMOS opened a fabrication plant in Newport in 1982, was sold to SGS-Thomson in 1989, and by 1994 the brand was gone (although the building is still there). The Newport site passed through several owners before being bought by Nexperia, then made subject to a government divestment order on national security grounds. A US company, Vishay, acquired it in 2024.
I thought about this while reading the Semiconductor Sector Study 2026, published this summer and updated last week. On computer memory, which accounts for almost 30% of the global market and is an area where the report sees an opportunity for Britain, it warns that leading British firms face scale-up barriers that could see them move operations overseas — risking a repeat of the UK’s INMOS experience.
Semiconductor design accounts for roughly 80% of UK sector revenues. Manufacturers report that government support is concentrated on the R&D side rather than the capital-intensive work of scaling production. They also stress that high energy costs are damaging their competitiveness.
Manufacturers responding to the 2026 sector survey also reported 18% of their cleanroom floorspace as under- or unutilised, and said that absent labour constraints their production could more than double. One suggestion in the report is to look at Fraunhofer in Germany and Tyndall in Ireland, where shared facilities get long-term support for running costs, not just capital.
🛒 Ian Ng, Researcher
The Government spends roughly £400 billion a year through public procurement and much has been said about how it should use that spending to support wider policy objectives, most recently on backing home-grown innovation.
The Competition and Markets Authority published two reports on public procurement yesterday. Both are filled with case studies of the procedural burden facing SMEs. A digital startup was rejected from the G-Cloud 15 framework, the Government’s key digital marketplace, because it used a Simple Agreement for Future Equity (SAFE) contract. SAFE is commonly used by early-stage startups but in this case was mistaken by the Government as a form of debt. The Startup Coalition later identified over fifty suppliers that had been rejected for the same reason. Separately, an SME with only five employees decided against bidding for an NHS England medical equipment tender due to the administrative burden of having to publish a current Carbon Reduction Plan, undertake an Evergreen supplier assessment, complete a Modern Slavery Assessment, and undertake Cyber Essentials certification. It reported that exemptions for SMEs from some of these requirements can themselves be burdensome to complete.
Loading every tender with competing secondary objectives produces everythingism: nothing much gets achieved, and the bar rises until only firms with the manpower to tick every box can clear it. The Government should heed the CMA’s advice to narrow its priorities to a meaningful number and to be honest about the trade-offs between them. It has been striking the right tone lately, from exempting smaller tenders from social value requirements to limiting the measurement of social value to job creation, skills and local opportunities.
Every procurement review since the Procurement Act 2023 has arrived at broadly the same conclusion: the system has an incumbency bias and a preference for the safety of a giant consultancy. Since April 2025, central government departments and their arm’s-length bodies have been required to set and publish their own three-year targets for direct spend with SMEs, and to report against them annually — a weaker mechanism than a single mandated figure, and one that has so far done little to change who wins the work. The problem, as the CMA identifies, is that reform has focused on rules and processes without touching the incentives facing procurement teams. Changing the rules alone will not be enough; unless the incentives on decision-makers change too, SME targets will go the same way. British investors are already seen as more risk-averse than their US counterparts — the Government cannot afford to be the same.
📈 Sophya Mashkoor, Research Intern
Two studies published in June push back on the idea that total founder control is good and investor oversight is just friction to be minimised.
Researchers at Imperial College London and Emlyon Business School examined 12 companies across 27 US securities fraud cases brought between 2000 and 2023. A separate study from the University of Toronto looked at 654 fraud cases against US venture-backed startups over the same period. Its headline finding is that fraud remains rare overall — but that startups with founder-controlled boards were twice as likely to face fraud charges as those with investor-controlled or shared-control boards.
The Imperial and Emlyon paper’s contribution is a mechanism it calls “façading”, a slow escalation that often starts innocuously. In the first stage, surface façading, founders overstate how successful the company is or is becoming. In reinforced façading, they manufacture evidence to support those claims — one company in the study fabricated customer contracts and invoices to book revenue that did not exist, then used it to raise at a unicorn valuation. In deep façading, the deception extends to the product itself, complete with staged demos.
Investors aren’t always the victims. The researchers describe some of them setting growth expectations so aggressive that founders feel there is no honest route to hitting them. The investors who push founders towards hockey-stick projections are also the ones best placed to catch the deception that follows — but in hot markets they compete for access by conceding exactly the governance rights that would let them do it, which is how founder-controlled boards become common in the first place. The paper argues that investors can unwittingly “co-create” fraud, and that continuing to back founders previously accused of it normalises the behaviour.
The Toronto study also found that startups launched during overheated markets, with weak oversight and thin due diligence, were 19% more likely to commit fraud later — which suggests board control is less a question of founder autonomy than of what market conditions do to governance. The proposed remedies are stronger; holding directors liable when façading happens on their watch, and formal audits of later-stage private companies. Both carry costs, not least making good people warier of taking early-stage board seats, and neither maps neatly onto the UK, where nothing polices private markets the way the SEC does.