Spanish FDI and Artificial Intelligence

Background

Artificial intelligence (“AI”) is increasingly important for companies and, consequently, as potential trigger for the Spanish FDI regime.  This note sets out the legal framework (Section 2, below), the interpretation of the authority (Section 3, below), and the levels of intervention as regards AI (Section 4, below).

Legal Framework

Article 7bis (2)(b) Act 19/2003, of 4 July, on the legal regime governing capital movements and economic transactions with foreign countries and on certain measures to prevent money laundering (the “Spanish FDI Act”) refers to AI in the context of “critical and dual-use technologies, key technologies for industrial leadership and training, and technologies developed under programs and projects of particular interest to Spain”.

Royal Decree 571/2023, of 4 July, on foreign investments (the “Implementing Regulation”) further elaborates on the above and brings the following types of AI within scope:

1.      AI falling within the definition of dual-use items in Article 2(1) of Regulation (EU) 2021/821 of the European Parliament and of the Council of 20 May 2021 setting up a Union regime for the control of exports, brokering, technical assistance, transit, and transfer of dual-use items.

2.      AI that has been developed with a substantial amount or percentage of funding from the European Union or Spanish budget, including the programs referred to in the annex to Regulation (EU) 2019/452 of the European Parliament and of the Council of 19 March 2019 establishing a framework for the screening of foreign direct investments into the Union.

(a)               There is no pre-determined threshold as to what constitutes a “substantial amount or percentage”.  The authority tells us that we have to analyze both (i) the amounts by themselves and (ii) how large a proportion of the project they finance.

3.      Key Enabling Technologies (“KETs”), including AI, as referred to in Council Decision (EU) 2021/764 of 10 May 2021 establishing the Specific Program implementing Horizon Europe - the Framework Program for Research and Innovation, and repealing Decision 2013/743/EU.

(a)               From a practical perspective, the latter is the most problematic category.  The Council Decision only includes general references to the European Commission’s priorities in funding projects in line with EU policies.  In this sense, the definition of KETs refers to AI without limitation to any specific type of AI.

Interpretation of the Authority

Despite the Spanish FDI Act subjecting the whole AI “sector” to a screening obligation (see Point 3 of Section 2 above), we gather from the Spanish FDI authority that the focus is on (i) the development of AI and (ii) ownership of proprietary AI, i.e., the mere use of AI would not be sufficient.

In addition, in a meeting of FDI practitioners in late January 2025, Marta Font, the head of the authority, stated that (i) on the one hand, AI and cloud computing are becoming increasingly important, meaning that there are very few companies that do not use these systems (i.e., in comparison to 2020 and 2023 when the screening mechanism and the Implementing Regulation entered into force, respectively); (ii) on the other hand, data, cybersecurity, and critical technologies are in the spotlight.  According to Ms. Font, this probably means that they will need to place less emphasis on the analysis of (i) and put more emphasis on the review of (ii).

Levels of Intervention

According to the statistics published by the authority, in 2024 36%, i.e., approximately 49 of all authorization requests were made under the technology category.  Even though this category is broader than AI, it gives an indication as to the importance of this sector.

In general, remedies under the Spanish FDI regime are possible, but they remain rare.  According to data from 2024, out of the 136 authorization requests, the Spanish authority imposed remedies in only eight cases.  Spanish FDI decisions are not publicly available, so it is not possible to ascertain the grounds on which the remedies were imposed.   However, according to the authority, they have indeed imposed remedies in transactions related to AI.  Again according to the authority (and as referenced in the statistics published on the authority’s website), mitigation measures in the field of technology are aimed at preventing the loss of sovereignty in this area.  Though not expressly stated, we tend to assume that this also applies to AI.