€180 million in AI grants in 2026: what has been announced and what to have ready before applying for anything

What has been announced

The Government has announced €180 million for 372 artificial intelligence projects in companies. It was presented by Óscar López, Minister for Digital Transformation, and published on 1 September 2026 (source).

That is all we have confirmed about the allocation.

What we do not know yet, and why we say so

You are going to read this announcement in many places this week. Many will add amounts per company, application deadlines and specific requirements. We will not, because we have not checked them.

As of the date of publication of this article we do not treat as confirmed:

• How much a specific company can receive.
• When the deadline opens and closes.
• What exact requirements the call asks for.
• What percentage of the project is subsidised.

If any of those figures is going to make you move money or dedicate hours, look for it in the official call when it is published, not on a blog. And that includes us.

We say it for a practical reason: an accounting firm that plans on a wrong figure does not lose an afternoon, it loses a funding window.

The context that is measured, and a discrepancy worth looking at

Two recent measurements of AI adoption in Spain, and they do not say the same:

• AI adoption in Spanish companies has gone from 11.4% to 20.5% in a year (source, 1 September 2026).
• 61% of Spanish companies already use AI, seven points above the European average (source, 4 June 2026).

20.5% and 61% do not contradict each other: they measure different things. And that difference is precisely the problem the first headline points to: in our translation, that the problem is no longer adopting AI but knowing what to use it for.

Translated to an accounting firm: someone at the firm using an assistant to draft emails counts as "using AI" in a survey, and changes nothing in operations. Changing how a specific process is run, does.

A public grant does not fix that difference. It funds it.

What an accounting firm needs in place for a grant like this to be of use

This is the part the press releases leave out, and it is where the money is really won or lost.

1. An identified process, not "we want AI"

"Implementing artificial intelligence" is not a project: it is an intention. A project is "document requests to clients follow a fixed cadence for each type of procedure, and nobody chases them by hand".

If you cannot write in one sentence which specific process changes, any quote you are given will be impossible to evaluate — and so will any application you submit.

2. Data in order before automating anything

Automating on a disordered archive multiplies the disorder. If the case files are in three places, client emails in each person's inbox and deadlines in a sheet updated by whoever remembers, AI does not fix it: it inherits it.

This is not a preliminary phase that can be skipped with budget. It is the phase.

3. A "before" number

How many repeated queries come in per month. How many hours go on chasing documentation. How many deadlines are warned about by hand.

Without that number you cannot justify the project to anyone, nor know afterwards whether it worked. And measuring "before" can only be done before.

4. Who is going to use it on Tuesday morning

A tool that nobody on the team is assigned to is left switched on and unused. Before applying for a grant, be clear about which person at the firm changes the way they work and how.

The order we recommend

First the process and the number. Then the quote. And the grant, if it fits, last.

Doing it the other way round —applying for the grant and then looking for what to spend it on— is the quickest way to end up with a half-paid tool and nobody to use it.

If you want to know where to start

If you have an accounting firm and do not know which of your processes is the one to look at first, write to us and we will tell you where we would start — with the call or without it. Calls and demos are held in Spanish.