AUS
NZ
UK
CA
USA
IND
Highlights
London Tech Week put sovereign computing power and major infrastructure commitments at the centre of Britain's technology agenda.
The UK's AI supply chain spans compound semiconductors, IT services, data-centre property, enterprise software and telecoms networks.
IQE has been among the year's most dramatic London gainers as investors connect its wafer technology to data-centre and sensing demand.
Every technology boom eventually stops being about the headline names and starts being about the supply chain. The artificial intelligence build-out is reaching that stage in Britain. London Tech Week brought a parade of commitments to domestic computing capacity, supercomputing projects in partnership with leading universities, and a political push for what policymakers call sovereign AI, the idea that the nation should own a meaningful slice of the infrastructure on which the technology runs. For investors in UK growth stocks, the significance is straightforward: a theme that once seemed the exclusive property of American mega-caps now has a genuine London expression, running from Welsh semiconductor fabs to warehouse parks on the capital's fringe.
The timing is notable because the broader market mood is anything but exuberant. The FTSE 100 and FTSE 250 have drifted towards multi-week lows on Middle East tension and pre-inflation-data caution, yet the AI-infrastructure cluster has retained a striking degree of investor attention. That persistence is the hallmark of a structural theme rather than a passing trade, and it justifies a closer look at the layers of the UK's AI stack.
Which companies supply the physical layer?
Start with the most tangible end of the chain. IQE (AIM:IQE), the Cardiff-based compound semiconductor specialist, has been one of the most dramatic gainers on the London market this year as investors have reappraised its advanced wafer technology in the context of data-centre construction, optical connectivity and sensing applications. The company has pinned its recovery narrative explicitly on AI-related demand, and its trajectory illustrates how quickly the market can re-rate a business once it is mapped onto a powerful theme. Raspberry Pi (LSE:RPI) occupies an adjacent niche, embedding low-cost computing into industrial and edge applications where AI workloads increasingly need local processing. Both names carry the volatility typical of hardware stories, but they give London something it has often lacked: listed exposure to the silicon end of the boom.
Who builds and houses the computing capacity?
Computing power needs somewhere to live and someone to install it. Segro (LSE:SGRO), the industrial and urban-warehouse landlord, has emerged as the property market's principal AI proxy, with data-centre development land and power-connected sites around major cities becoming some of the most sought-after real estate in the country. The economics of the theme, scarce grid connections, long leases and deep-pocketed tenants, map neatly onto the company's existing strengths. On the services side, Computacenter (LSE:CCC) sits in the path of corporate technology budgets as enterprises retool their infrastructure for AI workloads, sourcing, integrating and managing the hardware that underpins deployment. These businesses rarely feature in breathless AI commentary, yet they monetise the build-out with a directness that many higher-profile names cannot match.
How do software and data companies capture the value?
Above the physical layer sits the application economy, and here the UK has genuine champions. RELX (LSE:REL) has spent years converting its legal, scientific and risk databases into analytics products, and AI tools deepen the moat around proprietary data that cannot easily be replicated. Sage (LSE:SGE) is embedding AI assistants into accounting and payroll software used by vast numbers of small businesses, a route to monetisation that depends on trust and distribution rather than frontier research. The pattern in both cases is the same: companies that already own customer relationships and unique datasets can charge for AI-enabled productivity without bearing the capital costs of the infrastructure itself. For growth investors, that is an attractively asymmetric position, though it demands evidence that AI features translate into pricing power rather than merely defending existing revenues.
What about the networks underneath it all?
No AI strategy functions without connectivity, which is why BT Group (LSE:BT.A) keeps appearing in discussions of the national build-out. Fibre networks, low-latency links between data centres and the security architecture around them are foundational to everything else in the stack. The investment case here is more contested, since telecoms carry heavy capital requirements and competitive pressures, but the strategic logic of the network layer is hard to dispute. The week's policy emphasis on sovereign capability strengthens that logic: a country that wants domestic AI infrastructure needs domestic pipes, power and property as much as it needs algorithms, and each of those needs has a London-listed answer.
The UK's AI-infrastructure cohort cuts across several Industry Classification Benchmark sectors. IQE is classified within technology hardware and semiconductors on the Alternative Investment Market, where it ranks among the larger constituents tracked by the FTSE AIM 100 Index. Raspberry Pi and Computacenter sit in technology hardware and computer services respectively within the FTSE 250 and FTSE 100 universes, Segro is a real estate investment trust specialising in industrial and data-centre property, RELX is classified under media and data analytics, Sage under software, and BT under telecommunications. The breadth of classifications underlines a defining feature of the theme: AI infrastructure is not a sector but a value chain, and London offers listed exposure at almost every link.
What could make or break the theme from here?
Three tests lie ahead. The most immediate is macro: AI-linked growth shares trade at premium ratings, and the path of inflation and interest rates will govern how much optimism the market is willing to fund. The next is delivery: announcements made at conference podiums must convert into contracts, capacity and earnings, and companies that over-promise will be punished severely, as the sharp treatment of stumbling growth names elsewhere in the market this week demonstrated. The final test is power and planning, the unglamorous constraints of grid connections and permitting that will determine how fast Britain's data-centre ambitions can physically be built. The theme has momentum, political backing and a genuine industrial logic. Whether it has staying power will be decided not at tech conferences but in earnings statements, and that is where attention now turns.
