When governments say a technology is strategically important, they usually back that claim with more than warm words. They design the tax system around it. Yet in the conversation about AI Growth Zones, fiscal policy is barely mentioned. The debate focuses on planning priority, grid connections, and grants for adoption and skills. Those matter, but they leave one of the most powerful tools government has almost untouched.

This article explains, in plain terms, what the missing fiscal lever is, why it belongs in the debate, and how it could be designed so that it delivers wider Welsh value rather than becoming an open-ended giveaway.

What "fiscal lever" actually means

A fiscal lever is simply the use of tax to encourage or discourage behaviour. Reliefs, allowances and credits make some activities cheaper; taxes make others more expensive. Governments already use these levers all the time, for research and development, capital investment, and specific sectors and regions.

Applied to AI Growth Zones, the idea is straightforward. If the goal is to raise productivity by getting AI into Welsh businesses, then the tax system could be designed to make AI adoption, productivity-raising investment, and the attraction of scarce skills cheaper in the zones, on conditions that protect the public interest. At the moment, the incentives on the table are mostly about making it easier and cheaper to build data centres. That helps infrastructure. It does not, on its own, help a Welsh SME adopt AI.

A conditional exchange, not a subsidy

The obvious worry about tax relief is that it becomes a subsidy for things that would have happened anyway, or a windfall for large firms. That is a real risk, and it is why the framework argues for a conditional exchange rather than open-ended relief.

The principle is simple: targeted relief in return for measurable public benefit. If a firm wants the benefit, it accepts the conditions, Welsh jobs, local procurement, skills transfer, supply-chain participation, and measurable productivity gains. Relief is tied to outcomes and can be withdrawn if the conditions are not met. Used this way, tax stops being a blank cheque and becomes a genuine behaviour-changing lever. The public gets something specific and verifiable in exchange for the support it provides.

Three practical reliefs

The framework points to three areas where conditional relief could do real work.

The first is targeted tax relief for AI activity and SME adoption, not just for building or land assembly. The point is to reduce the cost of the thing Wales actually wants more of, which is firms adopting and using AI, rather than only rewarding the construction of infrastructure.

The second is capital incentives for SME productivity. Allowances that reward productivity-raising investment in established Welsh SMEs, not only start-ups or headline projects, would help ordinary firms buy and integrate the equipment and systems that make them more productive. Much of Wales's productivity gap sits in everyday businesses that never feature in announcements.

The third is talent support tied to transfer. Relief on income tax or national insurance for genuinely scarce AI roles could help attract senior specialists, but it should be conditional on mentoring, apprenticeships or formal skills transfer into Wales. That way, incentivising a handful of experts today helps build a much larger Welsh talent base tomorrow, rather than simply importing skills that leave again.

Who holds the levers, and the ask

Fiscal policy in the UK is complicated by devolution. Wales controls some levers, such as business rates and certain devolved taxes, while many of the largest ones, including corporation tax and most of income tax and national insurance, remain with Westminster. That is a reason to be clear and specific, not a reason to stay silent.

The practical step is for Wales to publish a specific fiscal ask: the reliefs it wants considered, the conditions attached to each, and the metrics that would govern them. A clear, conditional, evidence-backed proposal is far harder to ignore than a general call for support, and it signals that Wales is serious about turning AI infrastructure into productivity across the wider economy.

The missing fiscal lever is not about giving money away. It is about designing tax so that the benefits of AI Growth Zones are tied, explicitly and measurably, to Welsh jobs, Welsh suppliers, Welsh skills and Welsh productivity. If AI really is as important as ministers say, that lever is too useful to leave sitting idle.