For years, businesses have described data as an asset — and the expression is attractive, and often true in an economic sense. Data can improve decisions, reduce costs, identify customers, detect risk, support innovation and create entirely new products. Yet most businesses do not treat data the way they treat their other important assets.

An asset without the usual infrastructure

Asset classes such as real estate, purchased IP and plant and machinery are generally recognised on a company's balance sheet as tangible or intangible assets (“Recognised Assets”). Recognised Assets such as real estate and IP are often recorded on public registers, with ownership and characteristics clearly documented for third parties to review.

Data is different. Despite being widely seen as an asset, it is currently often absent from the balance sheet altogether. Businesses have had no way to register their valuable data or record its ownership in the way they can with Recognised Assets.

What has been missing is the legal and institutional infrastructure that Recognised Assets enjoy — the machinery that lets them function and opens them up to third-party trust and recognition. This question formed an important part of DataVision EMEA 2026. Alongside data products, trusted data and valuation, the conference considered the Isle of Man's developing Data Asset Foundation regime. For me, it also raised a wider issue: what has to exist before data can operate as a dependable, transactable asset?

Valuable does not mean transferable

Land is supported by title, boundaries, registration, transfer rules and systems for recording mortgages and other rights. Securities are supported by legal definitions, custodians, settlement systems and regulated markets. Intellectual property has recognised forms of ownership, licensing and enforcement. Data, despite being seen as an important asset, has none of these supports in a form that makes it a genuinely monetisable asset.

A business may say it owns a customer dataset, but the legal position is rarely that simple. The business may control the database while individuals retain rights over their own personal information. Parts of the data may have been supplied or licensed by third parties. Contracts, confidentiality obligations and data protection law may all restrict how it can be used.

A business may therefore hold a valuable collection of rights relating to data, without owning the information in the same straightforward way it owns an office, a vehicle or a piece of machinery. That does not mean data lacks value — it means that value, ownership, control and lawful use are separate questions. Until those questions can be answered, a third party will struggle to work out precisely what it is being asked to acquire, license, finance or rely upon.

The asset is more than the data

One of the most important ideas discussed at DataVision was that the asset should not be viewed simply as the underlying dataset. Instead, the dataset has to be considered together with the rights, obligations and governance that surround it.

Good governance should identify matters such as:

• what the data is;
• where it came from;
• who is responsible for it;
• how reliable and complete it is;
• what licences or third-party rights apply;
• whether it contains personal or confidential information;
• how it may be used;
• whether it may be accessed by AI systems;
• whether it can be licensed or commercialised; and
• what security interests or other claims have been created over it.

Why governance must travel with the data

Taken together, this creates a more complete commercial object than raw data on its own — information combined with evidence of its identity, provenance, permitted uses and continuing integrity. That distinction matters because a large dataset is not automatically a valuable one. Its value depends on its accuracy, relevance, uniqueness and lawful usability, as well as its ability to create income, reduce cost or manage risk.

A collection of inaccurate or unlawfully obtained records may have little value; once remediation costs and legal or reputational risk are taken into account, it may even be a liability. A structured, well-governed dataset, by contrast, has real value — and, with the right legal framework, that value can be monetised.

Within many organisations, the information needed to understand a dataset is spread across several different systems: the data itself in one platform, the licence in a contract repository, the privacy assessment with the legal or compliance team. An experienced employee may be able to bring those pieces together. A third party — and increasingly a machine — cannot safely be expected to reconstruct them. This matters more by the day, as AI systems and autonomous agents select, combine and process information without continuous human supervision.

Before using a dataset, an AI system may need to know whether it is authoritative enough to train a model. The owners of that AI may in turn want to know whether the resulting output can be commercialised. Without the data already being structured and governed to a recognised standard, these questions can currently only be answered through human intervention and inspection. A party can claim its dataset meets a recognised standard — for example, that it is suitable for AI training — but until now there has been no simple way for a third party to verify that claim.

The role of an authoritative register

The Isle of Man's Data Asset Foundation framework includes a statutory Data Asset Register. Its purpose is not to store the underlying data, but to provide an authoritative record of the data asset and the rights and obligations attached to it.

A register entry could record matters such as the identity of the asset, its stewardship, provenance, licences, permitted uses, encumbrances, and the involvement of AI in its creation or management. This would not remove the need for due diligence — a buyer, lender, investor or licensee would still want to consider data quality, commercial relevance, privacy and contractual risk. But it could change the starting point. Rather than reconstructing the nature and legal status of an asset from disconnected documents, a third party could begin with an authoritative record of what the asset is and which rights attach to it. Parts of the due diligence process could become verification rather than investigation.

The owners of an AI system would have an authoritative, independent record of the data asset and its provenance to point commercialisation partners to. The data feeding the AI could be formatted to a standardised, recognised framework — reducing the need for constant human intervention and helping to unlock AI's fuller potential.

From recognition to economic use

It is important not to overstate what a statutory framework can achieve. Registration cannot make poor data valuable, nor can it override data protection law, intellectual property rights or contractual restrictions. It cannot guarantee accounting recognition, and it cannot create commercial demand where none exists.

What it can do is create a clearer, more dependable object around which commercial activity can take place. Once a data asset can be identified, and its rights, restrictions and governance verified, it becomes easier for third parties to consider licensing or acquiring rights in it, valuing it, taking security over it, including it in a corporate transaction, or permitting controlled access to it.

In Summary

The role of the Isle of Man's data register is not to price every data asset. It is to provide enough legal certainty and institutional trust for valuers, insurers, lenders, investors and commercial counterparties to do their own jobs — the same infrastructure that other established classes of Recognised Assets already enjoy.