Big announcements are easy to measure. A site is named. An investment figure is quoted. A jobs number makes the headline. But if Wales is serious about turning AI Growth Zones into long-term economic value, it needs a better way to judge success than announcement politics.
That is why the framework is built around three practical tests.
These tests are not anti-investment and they are not anti-data-centre. They are a way of asking whether public support, planning priority and scarce energy capacity are being turned into measurable value for Wales beyond the site boundary. In other words, they ask whether AI Growth Zones are being designed for diffusion rather than display.
Test one: do data-centre deals clearly benefit Welsh SMEs?
Major AI infrastructure will often benefit from accelerated planning, energy access, public support and political attention. That is a powerful package. The first test is whether Wales gets something concrete in return for smaller firms and local communities.
That means visible conditions around Welsh supplier participation, not vague aspirations. It means targets for spend with Welsh firms, including SMEs. It means transparent reporting. It means skills programmes that widen access to specialist work instead of only training people for the needs of large operators. And it means thinking creatively about community benefit, including access to cloud credits, compute, partnerships or procurement routes that smaller firms can actually use.
If a Growth Zone attracts large capital but leaves little trace in local supply chains, it may still look successful from the outside. The first test asks whether that success is real enough to reach the Welsh business base.
Test two: does adoption funding back real SME projects rather than short pilots?
Every Growth Zone is likely to come with adoption and skills funding. The risk is that this money disappears into light-touch pilots, short courses and generic innovation language without ever changing how firms actually work.
The second test is whether support reaches SMEs in a form that matters. That means sector-based cohorts, practical implementation work over six to twelve months, trusted providers, and support that measures real outcomes such as time saved, defects reduced, output improved, margins strengthened or resilience increased.
It also means designing for the economy Wales actually has. Rural firms should not be an afterthought. Welsh-language support should not be decorative. Smaller businesses need clear routes into trusted help, not complicated schemes that require them to become policy experts before they can access support.
If adoption funding creates only a handful of showcase case studies, it has probably failed. If it helps hundreds of firms make practical changes that stick, it begins to move the wider economy.
Test three: is the grid being planned in a way that leaves space for smaller users?
AI infrastructure is also an energy story. The South Wales Growth Zone alone is expected to involve very large power demand. That makes grid planning central to whether AI Growth Zones work for Wales or around Wales.
The third test is whether new power and reinforcement plans leave room for SMEs, public services and smaller industrial users. If large projects get priority connections while everyone else faces long delays, the result may be more than unfairness. It may actively slow the spread of productivity gains across the economy.
A better approach would reserve space for local demand, encourage flexible connection models for smaller users, and use large AI sites as anchors for new renewable and storage investment that improves resilience for surrounding businesses and communities as well.
This matters because a Growth Zone that consumes local capacity without widening local opportunity risks becoming an energy island: impressive, strategic and largely disconnected from the economic reality around it.
How the tests should be used
Taken together, the three tests do something important. They change the question from “How big is the announcement?” to “What is actually changing for Welsh firms?”
They also give policymakers, business groups and the public a practical way to hold delivery to account. Are local procurement commitments being published? Are SMEs completing meaningful AI projects? Are smaller users seeing faster, cheaper routes to power? Those are the kinds of questions that reveal whether the framework is working.
Used properly, the tests could shape funding conditions, public scorecards and procurement choices. They could also support a wider fiscal argument: if government is willing to offer tax advantages or fast-track treatment for AI infrastructure, those benefits should be earned through visible Welsh outcomes.
If Wales uses these tests seriously, AI Growth Zones can become more than infrastructure policy. They can become a disciplined model for turning major investment into local capability and wider productivity gains.
If not, the country risks celebrating inputs while missing the harder, more important outcomes.