10. Scotland Leads the Future 🟡

Scotland already has a dense cluster in photonics, quantum, semiconductors, and sensing; a nationally hosted supercomputer; Europe’s leading small-satellite manufacturing base; and a licensed vertical spaceport. Planning, skills, universities and enterprise agencies are devolved. Corporation tax, the largest industrial subsidies, immigration, space-launch licensing and the UK compute budget are not. The question is how far those foundations become scaled jobs here, not whether the laboratories exist.

Current reality

Legal and institutional position

Planning, further and higher education, apprenticeship frameworks, Scottish Enterprise and Highlands and Islands Enterprise, business rates and a Scottish research budget are devolved. The National Manufacturing Institute Scotland and university clean rooms sit inside that system.

Reserved or UK-administered: corporation tax and the main subsidy-control regime; immigration; Civil Aviation Authority and UK Space Agency licensing of launch; UKRI ownership and funding of the national supercomputer; designation of AI Growth Zones and National Wealth Fund guarantees. The Equality Act 2010 bars hard nationality or race quotas in private employment. Positive action is narrower than a domicile reservation of leadership posts.

The latest official numbers

Critical Technologies Supercluster (photonics, quantum, semiconductors, sensing and connectivity).
Industry and Scottish Government descriptions in 2025–26: more than 150 companies; about 10,900–11,000 jobs; turnover cited at £4.2 billion (2022 baseline in the Technology Council report) and, in an August 2026 ministerial statement, £4.5 billion. Exports are a large share of output. GVA per employee is described as about three times the Scottish average. The published ambition is £10 billion turnover and around 17,500–18,000 jobs by 2035. The Central Belt is stated to produce about 20,000 engineering and physical-science graduates a year from HNC to PhD. Two of the five UK Quantum Technology Hubs are led from Scotland. Fraunhofer Centre for Applied Photonics is the UK’s only such centre. Open-access fabrication includes the Scottish Microelectronics Centre and the James Watt Nanofabrication Centre.

NASPIC.
Announced in 2025 at Inchinnan, led by NMIS / University of Strathclyde, described as Europe’s first open-access advanced semiconductor packaging centre of its kind. Projected to support about 300 jobs and enable around £800 million of revenue for users over a decade. Backed by Innovate UK, Scottish Enterprise and Glasgow City Region Innovation Zone money within a larger UK regional package.

Compute.
ARCHER2, the current UK national supercomputer, is at Edinburgh’s Advanced Computing Facility. The university is the UK’s National Supercomputing Centre. Construction of buildings for the next UKRI-owned national machine began in June 2026 at a Midlothian site. UK funding is up to £750 million. The university’s published comparison is about fifty times ARCHER2, with hoped-for use around spring 2028 after buildings are completed at the end of 2027. Chip-cost inflation has led UK ministers to drop some earlier “exascale” language. That is a specification risk, not a cancellation in the official record.

Lanarkshire AI Growth Zone.
UK designation, January 2026, around DataVita and CoreWeave. The UK announcement cited more than 3,400 jobs and more than £8 billion in private investment, plus a large community fund. In August 2026, DataVita closed a £300 million debt facility (ING, ABN AMRO, Santander, Scottish National Investment Bank, Siemens), with a £202 million National Wealth Fund guarantee, to expand DV1 and build DV3 — 40 MW each, both leased to CoreWeave for 15 years. The operator’s own figures for those two buildings are about 600 construction jobs and about 100 permanent roles. On-site generation of more than 500 MW within four years was part of the designation narrative. Grid connection and consented generation at that scale are not evidenced as complete.

Space.
Scottish Government export-plan figures: around 246 organisations, more than 7,000 jobs, income around £298–300 million, nearly 20 per cent of UK space employment, and the largest European producer of small satellites by volume (Glasgow). SaxaVord holds a full vertical-launch licence. The CAA licensed RFA One for up to ten orbital flights a year. A five-week window from 10 August 2026 was cancelled after RFA destacked the vehicle in late July to investigate tanks. The operator still spoke of a first test later in 2026. Sutherland’s primary operator has faced delays and, in 2026, insolvency in related launch activity. Prestwick remains a horizontal/aerospace hub.

Storage, not cells.
Invinity manufactures vanadium flow batteries in Motherwell and Bathgate. Grid-scale batteries (Blackhillock and others) are operating or consented. No Scottish gigafactory for automotive cells is in the official pipeline comparable to English sites.

How it is experienced in daily life

A photonics or quantum engineer in the Central Belt can change employer without leaving the cluster. A packaging engineer still often waits months for offshore assembly until NASPIC is fully online. A Shetland community prepared for a launch window that never opened. A Lanarkshire construction worker may see a data hall; the 3,400-job figure is not the same as the 100 permanent roles attached to the first two buildings. A researcher using ARCHER2 is on a machine that has fallen down the world rankings, while the next hall is a building site.

How it got here

University physics and engineering, not a single industrial plan, produced photonics and small satellites. UK quantum hubs and UKRI compute landed where the groups already were. Enterprise agencies assembled a “supercluster” label over firms that already exported. Launch is a UK regulatory path on Scottish land. AI data centres followed power, fibre and a UK zone designation. Cell manufacturing followed automotive plants that Scotland does not host.

What has already been tried, and what that showed

Shared infrastructure (Fraunhofer, NMIS, NASPIC, Robotarium) shortens the path from lab to package. It does not substitute for late-stage capital or a domestic automotive OEM. AI Scotland and SME adoption programmes address diffusion, not frontier model training at hyperscale. Spaceport licensing proved possible; first orbital flight has not. Growth-zone designation and a wealth-fund guarantee can bring debt into a data hall. They do not, by themselves, deliver 500 MW of private-wire power or 3,400 permanent jobs. Clas-SiC’s 2026 capital grant is the routine enterprise tool working at factory scale.

What needs to happen

The improvement in measurable terms

• Supercluster turnover and jobs moving toward the 2035 figures without counting construction as permanent AI employment.
• NASPIC in operation with published lead times in days, not months.
• National supercomputer on the stated 2028 path with a published specification.
• A first licensed orbital attempt from SaxaVord that either flies or fails in public, not an open-ended pause.
• Data-centre jobs reported as construction versus operations, with grid MW actually connected.
• Graduate and apprentice flows into named cluster occupations, not only into general STEM.

Sequence

Keep shared kit running. Consent power and grid for compute before repeating investment headlines. Finish packaging capability so wafers do not leave. Treat launch as a regulated experiment, not a tourism date. Align college and MA places to the trades RAND-type studies already name for advanced manufacturing. Do not wait for a reserved tax holiday to do those things.

Workforce, capital and standards

The cluster already pays above-average salaries. The gap is technicians, packaging engineers, launch operations and mid-career movers, not only PhDs. Clean-room and packaging capital is the manufacturing bottleneck. Launch safety standards are CAA standards. Equality law is the standard that bars a Scots-only quota.

What would count as success in twelve months

• NASPIC equipment and first user projects on a public timetable
• Supercomputer building progress against the end-2027 hall date
• Supercluster survey update that is not only the 2022 £4.2 billion baseline
• A restated SaxaVord window or a dated reason it cannot fly
• Lanarkshire MW connected and jobs split in an official update
• No new public claim that domicile quotas will allocate private leadership posts

What can be done under current powers

Who holds the lever

Scottish Ministers and planning authorities: sites and pace. Enterprise agencies: grants, premises, cluster coordination. Universities and SFC: places and translation. SDS: apprenticeships. UK Government, UKRI, CAA, Ofgem and the National Wealth Fund: compute budget, launch licences, grid, guarantees, corporation tax.

The statute, budget or direction

Town and Country Planning (Scotland) Acts; enterprise agency directions; higher and further education funding; Scottish AI Strategy 2026–2031; UK Space Industry Act licensing via CAA; Equality Act 2010.

What is already possible and unused

Faster, honest planning for energy-intensive sites. Skills programmes tied to NASPIC, photonics and satellite lines. Public procurement of sensing, health-AI and space data that builds Scottish suppliers. Heat reuse from compute into hospitals and homes where the physics works. Publishing job and MW figures that match company filings, not zone brochures.

What would still be true if management improved tomorrow

Chip prices would still threaten the supercomputer’s original specification. RFA’s tanks would still be an engineering problem. CoreWeave would still be a US customer on a 15-year lease. Equality law would still forbid a nationality quota. A gigafactory would still need an OEM and reserved-scale incentives.

Where the limits are

Law and reserved powers

Tax incentives at gigafactory scale, immigration of specialists, launch licensing, UKRI compute and the Wealth Fund are not Holyrood instruments. A Scottish visa for cluster staff is reserved.

Money

£750 million for the supercomputer is UK. NASPIC’s larger envelope includes UK regional funds. SNIB can co-lend; it cannot replace late-stage global capital. Enterprise grants work at Clas-SiC scale, not at full cell-plant scale.

Workforce and capital

Eleven thousand cluster jobs are real and still small next to global AI and semiconductor employment. Packaging and launch skills are thin. Housing and infrastructure around Unst and Lanarkshire constrain how fast people can move.

Time, geography, UK-wide systems

A supercomputer hall is 2027–28. A first orbital flight is not a planning permission. Grid queues are GB. Polar launch latitude is an asset only if a vehicle flies.

Political difficulty, kept separate from legal impossibility

Inflating 100 operational jobs into 3,400 is a communications choice. Admitting that on-site 500 MW is not consented is politically harder than repeating the designation text. Neither changes the Equality Act.

What is not yet known

Whether the national supercomputer meets the fifty-times-ARCHER2 claim once chips are bought. When SaxaVord actually launches. How many of the 3,400 Lanarkshire jobs are permanent operations. 2025/26 supercluster turnover in a consistent official series. Whether NASPIC’s day-scale packaging is achieved in practice. Orbex/Sutherland’s residual position after operator failure.

Honest summary

Achievable now. Run the cluster as one pipeline. Finish packaging and clean-room capacity. Align skills. Plan power before headlines. Report jobs honestly. Use enterprise tools at factory scale.

Only partly achievable. A European-scale AI training campus, a cell gigafactory, routine orbital launch, and a supercomputer whose final spec is still being bought on a global chip market.

Would require reserved change. Corporation-tax holidays of the usual industrial-strategy size, a Scottish immigration route, CAA replacement, or UKRI’s compute budget becoming a Scottish line.

Colour mark: yellow. The science and the firms are already here under Scottish and UK co-funding. Planning and skills can raise the conversion rate into jobs. The largest cheques, the launch licence and the labour-market border are not Scottish. Yellow does not claim the cluster is weak. It is a claim that leadership in the next decade is only partly a Holyrood choice.

What this item does not settle. It does not treat 3,400 jobs as a fact. It does not treat a paused rocket as a failed space sector — manufacturing is the present strength. It does not say quotas would be wise if they were lawful. It does not argue that compute should be refused because power is scarce; it argues that power must be real.

The test is whether the next official survey shows more packaged devices, more satellites, more operational compute staff, and a launch that happens or is explained—not another ambition date.

Connection

Item 7 is the same skills and immigration split. Item 8 is the power and grid constraint on data centres. Education (item 6) is the graduate pipeline. Shipbuilding and energy share the technician pool. Public finances and SNIB appear in the Lanarkshire facility. The Independence Debate focuses on reserved tax, borders, and whether a later state would fund compute and launch itself—not on whether Edinburgh already hosts ARCHER2.

People’s Future Scotland