The Framework
Wales is about to become one of the UK’s main testbeds for AI-driven growth. Two AI Growth Zones are expected to attract major investment, new infrastructure, and significant construction activity. The enduring economic test, however, is whether they help Welsh SMEs adopt AI, raise productivity, build specialist capability, and widen access to high-value work.
Why this matters
Wales is becoming one of the UK’s main testbeds for AI-driven growth, with one AI Growth Zone planned in North Wales and another along the M4 corridor in South Wales. Together they are expected to attract very substantial private investment, major infrastructure build-out, and thousands of jobs during construction.
But the longer-term prize is different. The lasting economic test is whether the benefits diffuse into Welsh SMEs, supply chains, and new specialist firms. If the gains remain concentrated in a handful of fenced-in sites, Wales will miss the wider opportunity.
Wales has the lowest productivity of any UK nation or region. If AI policy does not help close that gap, infrastructure alone will not be enough. The real question is whether Growth Zones improve productivity across the firms where most Welsh citizens actually work.
SMEs account for around 62.3% of employment in Wales. That means any serious AI growth strategy has to work for tens of thousands of smaller firms, not only large operators, hyperscalers, or a narrow group of specialist employers.
AI adoption in Welsh SMEs is still low, with estimates around 7–10% and Senedd-commissioned research finding 7.4% adoption in 2022. If AI does not reach the wider SME base, it will not solve the productivity problem, no matter how many data centres are built.
The three tests
These tests are not arguments against AI Growth Zones. They are criteria for whether Growth Zones and AI Cymru are actually working for Welsh businesses and the wider Welsh economy.
Test 1
Growth Zones will make AI-related infrastructure easier and cheaper to build through faster planning, grid access, discounted electricity, and potentially wider public support. In return, Wales should secure measurable commitments to local procurement, Welsh SME participation, skills transfer, supplier development, and where fiscal relief is offered, clear local conditions.
Test 2
Each Growth Zone comes with adoption and skills funding, but the risk is that this gets diluted into short-run pilots, generic bootcamps, or superficial trials with little business impact. The better approach is to back 6–12 month transformation projects in real Welsh firms, organised around sector-based cohorts, measurable productivity gains, rural participation, Welsh-language support, and voucher-style access to accredited Welsh providers.
Test 3
AI Growth Zones are fundamentally about power. With South Wales alone expected to draw more than 1GW in the early 2030s, grid reinforcement must not prioritise hyperscale sites at the expense of SMEs. Strategic planning should reserve room for local demand, support flexible connection models, and use major AI infrastructure as an anchor for renewables and storage that also benefit smaller firms and communities.
The missing fiscal lever
One important lever is still largely missing from the current debate: fiscal design. If AI is as strategically important as ministers say, AI Growth Zones should be designed to be meaningfully tax-advantaged for both firms and scarce talent.
This need not mean open-ended subsidy. The better model is a conditional exchange: targeted relief in return for Welsh jobs, local procurement, skills transfer, supply-chain participation, and measurable productivity gains. Used this way, fiscal policy becomes a serious behaviour-changing lever rather than a blank cheque.
Relief should support AI activity and SME adoption, not just new data-centre construction or land assembly.
Capital incentives should support productivity-raising investment in established Welsh SMEs, not only start-ups or headline infrastructure projects.
Income tax or NIC relief for scarce AI roles should be tied to mentoring, apprenticeships, or formal skills transfer into Wales.
What government should do now
The next phase needs to be practical, visible, and easy to understand from the perspective of Welsh firms. The priority is not more abstract strategy, but early delivery that proves the model works.
Commit in year one to support an initial wave of SMEs in priority sectors in each Growth Zone, with funding and technical support for 6–12 month AI projects, and publish outcomes on productivity, quality, cost, resilience, and workforce transition.
Set up a cross-government team with licence to move quickly, unblock procurement and energy issues, and work directly with ecosystem partners on short delivery cycles.
Use existing business networks, sector bodies, and regional innovation partners as a shared front door that routes SMEs into funded projects within weeks rather than months, with light-touch paperwork and pre-approved provider lists.
Do not just track jobs, capital spend, and headline GVA. Track SME productivity improvement, local SME AI adoption, the share of spend reaching Welsh SMEs, and the number of new specialist Welsh firms created or scaled.
Set out the reliefs Wales wants Westminster to consider, the conditions attached, and the metrics that will govern them.
What Welsh SMEs and founders can do now
The success of AI Growth Zones cannot depend on government action alone. Welsh SMEs and founders have a critical role in shaping, testing, and accelerating practical AI adoption across the economy.
Identify where AI can deliver measurable gains in productivity, quality, customer experience, cost reduction, or new niche products and services. Clear, scoped projects will make firms ready when funding or cohort programmes appear.
Manufacturing, construction, logistics, tourism, care, agriculture, food and drink, and the creative industries can all benefit from collective learning and shared adoption programmes.
Use existing support from universities, applied innovation programmes, business networks, and regional ecosystem partners rather than waiting for new Growth Zone structures to be fully built.
AI adoption is easier when firms have cleaner data, stronger processes, and staff who understand how tools can support their work. Small steps now can materially improve readiness later.
Planning policy, procurement, grid reinforcement, and Growth Zone design will all shape outcomes. SMEs should show up early and collectively so real needs are reflected in delivery.
One of the biggest barriers to adoption is uncertainty. Practical examples from Welsh firms can build confidence and make AI feel real, useful, and attainable.
Closing view
If AI Growth Zones exist mainly to host racks for global tech giants, generate substantial construction activity, and provide a relatively modest number of long-term specialist roles, they may still look impressive on a press release. But that alone will not solve Wales’s productivity gap.
If, instead, they are used to put AI tools, skills, clean power, and targeted incentives into the hands of tens of thousands of small Welsh businesses, and to help more specialist Welsh firms form and scale, they become something far more important: a practical answer to Wales’s productivity challenge rooted in the real economy.
That is why these three tests, and the cross-cutting fiscal principle behind them, matter so much. If Wales is to be an AI nation, AI must work first and foremost for its small businesses and entrepreneurs.