The market for mergers and acquisitions in the artificial intelligence sector is growing rapidly. Liechtenstein is also increasingly positioning itself as an attractive jurisdiction for technology-focused investment structures, venture capital investments, and international holding models in the AI sector. Strategic investors, private equity firms, and technology companies are specifically investing in AI startups, data-driven business models, and proprietary models.

At the same time, AI M&A transactions differ significantly from traditional technology deals. The value of an AI company often depends less on traditional financial metrics and more on data quality, regulatory compliance, technological expertise, and actual control over the models being used.

For buyers, this means that a traditional financial and legal due diligence process is often no longer sufficient. In particular, issues relating to training data, IP rights, data protection, GDPR compliance, and regulatory or technical risks are becoming increasingly critical.

Who owns the data, models, and IP rights?

One of the central questions in AI transactions concerns the ownership and rights chain of the technologies being used.

Many AI companies rely on open-source models, third-party APIs, or external datasets. However, it is often unclear whether all rights have actually been validly transferred to the target company. Especially in fast-growing AI startups, proper documentation and clear licensing structures are frequently lacking.

The origin of training data is particularly critical. Numerous models have been trained using publicly available or automatically scraped content. Whether this is permissible under copyright law is currently being heavily debated across Europe and the United States.

This creates substantial risks for buyers, including injunction claims, damages, or regulatory proceedings. As part of the due diligence process, particular attention should therefore be paid to the origin of datasets, existing licensing agreements, and internal governance processes.

How independent is the business model really?

Many purported AI companies do not possess their own model infrastructure but are economically dependent on foundation models provided by major companies such as OpenAI, Anthropic, or Google.

In many cases, the actual value proposition consists merely of user interfaces, workflow automations, or prompt structures. Investors must therefore assess whether the company truly has a sustainable technological competitive advantage.

The more a business model depends on external APIs, the greater the risk of future price increases, technical restrictions, or margin pressure. Companies without their own data base or technological differentiation are increasingly exposed to competitive pressure.

EU AI Act, GDPR, and regulatory risks

The EU AI Act introduces the first comprehensive regulatory framework for artificial intelligence in Europe. Due to Liechtenstein’s EEA membership, the regulatory developments under the EU AI Act will also have immediate relevance for many companies and investors in Liechtenstein. Depending on the business model, extensive compliance obligations may arise — particularly for high-risk systems or general-purpose AI.

In addition to the EU AI Act, the GDPR also plays a central role in AI M&A transactions. Especially in data-driven business models, the key question is often whether personal data was lawfully collected, processed, and used for training purposes.

For buyers, it is therefore essential to determine whether regulatory requirements have been complied with, whether sufficient governance structures are in place, and whether potential compliance risks are identified at an early stage.

Regulatory violations can result in significant financial and operational consequences.

Key-person risks and post-closing integration

In the AI sector, company value often depends heavily on the know-how of individual developers or founders. If these individuals leave the company after closing, the buyer may lose the ability to further develop or efficiently operate the models.

As a result, retention programs, earn-out structures, and long-term incentive models are becoming increasingly important.

The integration of AI companies is also frequently underestimated. Particularly when acquired by larger corporations, agile development structures often clash with traditional governance and compliance processes. The first months after closing are therefore often decisive for the long-term success of the transaction.

Conclusion

The acquisition of AI companies requires a significantly deeper and more interdisciplinary due diligence process than traditional technology transactions.

In addition to corporate and transactional matters, data protection, GDPR compliance, copyright law, regulatory compliance, and technical reviews are becoming increasingly important.

Companies and investors should therefore structure such transactions at an early stage together with M&A lawyers, data protection experts, and technical specialists in order to identify risks early and safeguard the long-term value of the transaction.