China now treats data as an asset that can be traded, put on a balance sheet, borrowed against and registered. A US congressional advisory body has set out how it got there. It has also asked whether the United States should follow.

From where we sit as advocates, the more important question is legal rather than accounting. Putting a value on data is one thing. A lender also needs to know what the asset is, who holds rights in it and what it could enforce. That is the problem the Isle of Man's Data Asset Foundation regime was built to address.

What has the US Commission published?

In August 2026, staff at the US-China Economic and Security Review Commission published an issue brief titled The People's Republic of Data: How China Is Turning Data into Capital. Congress set up the Commission to advise it on the economic and security sides of its relationship with China. Most of its work concerns trade, technology and defence. This brief is about something quieter, though it may prove just as important. What happens when a country decides that data should be treated as capital?

I have spent much of the last two years working on the Isle of Man's Data Asset Foundation regime. So I read the report with more than passing interest. I am not a China specialist, and this is not an article about Chinese politics. Yet the report raises exactly the questions that businesses, lenders and regulators everywhere will soon be asking. They are worth setting out plainly.

Why does the legal question come first?

Property law has long drawn a line between knowing what something is worth and proving that it is yours. Anyone who has bought a house knows which of the two the bank asks about first. A valuation tells a lender how much it might recover. A title tells it whether it can recover anything at all. Keep that distinction in mind, because it runs through everything that follows.

Is computing power still the real bottleneck in AI?

For several years, Washington has tried to keep its lead over China in AI by restricting access to the most advanced chips. The reasoning is simple. Limit access to the best semiconductors and, in theory, you limit the ability to train the best models.

The report suggests that compute is no longer the only constraint that matters. Frontier AI developers have largely worked their way through the open internet as a source of training material. As a result, public text is fast becoming a commodity. What remains scarce is data that cannot be scraped. That means company records, operational logs and data from machines, sensors and factories that never touch the public web.

That is precisely the data China's policies are designed to capture, organise and put to work. However, it sits in a layer that chip export controls simply do not reach.

What has China actually done with data?

The starting point came in April 2020. China's leadership formally named data a 'factor of production', alongside land, labour, capital and technology. In the West that might read as a line in a strategy document. In China it set off six years of practical follow-through on four fronts.

Exchanges

China has built a network of government-backed data exchanges where datasets can be listed, bought and sold. Around them has grown an industry of services: cleaning and labelling data, legal and compliance checks, valuation and storage.

By the end of 2025, China's National Data Administration reported at least 4,000 data exchanges, infrastructure operators and data merchants. Together they offered more than 13,000 data products and services. The largest exchanges, in Guiyang, Shenzhen, Shanghai and Beijing, each handle more than RMB 1 billion (around US$150 million) a year. Official exchanges handled under 5 per cent of data transactions in 2021, but around 20 per cent by 2024.

Accounting

This is the part that caught my attention most. Since January 2024, Ministry of Finance rules have let Chinese companies record data on their balance sheets. It can sit there as either an intangible asset or inventory.

Western rules, on the other hand, work differently. Under IAS 38 and US accounting rules, the cost of generating data internally is largely expensed as incurred. As a result, much of the value of that data never appears as a separately recognised asset.

China, however, took a different path. By December 2025, 136 companies on China's A-share markets had recognised RMB 3.8 billion of data assets. A further 417 unlisted companies had done the same by March 2026.

Finance

Accounting recognition does not by itself create collateral. It does give companies, valuers and lenders a clearer starting point for structuring finance. Indeed, a small number of companies have already borrowed from banks against their data assets.

By May 2026, securities described as data asset-backed had reached RMB 20 billion in cumulative issuance. By comparison, issuance for the whole of 2025 was RMB 5 billion. However, the report cautions that most were not pure data securitisations, because repayment depended on other cash flows or guarantees. The same month brought China's first so-called 'pure' data asset-backed security. This was a RMB 532 million issue linked to data-asset income streams, and even it was supported by a state-owned guarantor.

Standards and registration

China has presented its data accounting framework to the International Accounting Standards Board. That body is currently reviewing how intangible assets are treated. In March 2026, Beijing also launched a new international body, the World Data Organization. According to the report, it has already attracted more than 200 members from 40 countries.

Since the report was published, China has gone a step further. On 21 September 2026, its National Data Administration announced the first batch of data property-rights registration certificates. Fifty-six enterprises and institutions received certificates evidencing rights to hold, use and operate specified data. The system is meant to provide authoritative evidence of data rights for trading, balance-sheet recognition, financing and investment.

Taken together, that is a remarkably complete attempt to give data the machinery other assets already enjoy. There is a place to trade it, a way to account for it, a means of financing it and now a register. Notice the order, though. China spent the best part of six years on the valuation. Only now has it begun issuing something that looks like title.

Has China's approach actually worked?

What I found refreshing about the report is that it does not oversell any of this. It is candid about the difficulties.

What has gone wrong?

Much of the data offered on the exchanges is duplicated or of poor quality. In addition, most data in China still changes hands through private agreements rather than on exchanges. Meanwhile, private technology companies remain reluctant to share what they see as a competitive advantage.

There is also a gap in the law. China has created rights to hold data resources, to process and use data, and to manage data products. However, it has not created a clear general ownership right in the underlying data. The report compares this to the way China treats urban land.

Then, in June 2026, regulators paused approvals of new data-backed securities. They had concerns about underwriting standards. There were also concerns that some issuers were using the securities to get round limits on local government borrowing.

Problems of title, not value

None of that should surprise anyone. Anything built at this scale runs into problems, and the Chinese authorities have adjusted as they go. The report's own conclusion is that China has still laid the foundations for treating data as an asset. As a result, it is now well placed to help set the international norms.

The Commission's Vice Chair, Mike Kuiken, made a similar point to Reuters when the report came out. Over the last half decade, he said, China has consolidated, labelled and refined its data. It has also made new data quickly available to those who can use it.

Yet the weaknesses the report lists are telling. Poor-quality data, unclear ownership, securities that leaned on guarantees rather than on the data itself. Almost without exception, these are problems of title rather than problems of value.

What does the report suggest Congress should consider?

The closing considerations for Congress are the part of the report likely to matter most, and for longest. To be clear, this is a staff research paper. These are suggestions for Congress to consider, not law or policy. Still, they show clearly where the conversation in Washington is heading.

A national data strategy

China has made data a strategic economic asset. By contrast, America has made no comparable national decision. So the report suggests Congress consider a data strategy that treats data as an economic asset. Mr Kuiken told Reuters this was the Commission's 'number one recommendation'.

Accounting recognition

US companies have no equivalent to China's rules, which leaves the value of corporate data invisible in financial reporting. The report suggests Congress consider whether US accounting standards should recognise data as an asset. It also suggests encouraging the Securities and Exchange Commission and the Financial Accounting Standards Board to examine it quickly.

Many of us have spent time explaining to a finance director why their most valuable resource is not on the balance sheet. For us, that is a significant sentence to find in a congressional report.

International standards

China is already presenting its approach to the IASB as that body works on intangible assets. The report suggests the United States work with allies to lead on data interoperability and transaction standards. In other words, it should not cede that ground. Its reasoning is blunt: standards are difficult to revise once adopted, so engaging now beats trying to unwind one later.

Knowing what data you hold

Perhaps the most practical suggestion concerns federal agencies, which hold extensive agricultural, geological and health research data. The report suggests requiring them to inventory and value their data holdings within six months. That would give a baseline for protecting the data and putting it to productive use.

Why should businesses outside the US and China pay attention?

It would be easy to read all of this as a story about two superpowers. It is more useful to read it as a sign of where everyone is heading. The direction of travel is the same, whoever happens to be driving.

Accounting and standards

The accounting point is the obvious one. For years, the accounting treatment of internally generated data drew little mainstream attention. That now looks far less settled. China has adopted its own rules, and the IASB has begun a broader review of IAS 38. Meanwhile, a US congressional staff paper is asking whether American standards should change too.

The IASB's project is not specifically about data, but the direction of the debate is worth watching. Boards that assume data will stay off the balance sheet indefinitely may want to revisit that assumption.

Standards matter here too. The rules for how data is defined, valued, registered and traded are being written now. Once embedded in accounting practice, audit methodology and lending criteria, they will be very hard to unwind. So the report's warning about engaging early applies to every jurisdiction, not just the United States.

The legal point

For a law firm, though, the more interesting issue is the legal one. Getting data onto the balance sheet matters, but it is not the whole answer. In many common-law jurisdictions, lenders can already take security over categories of intangible property and contractual rights.

What they struggle with is the detail. Exactly what is the data asset? Who has rights in it? Has it been properly governed? And what could a lender actually enforce if things went wrong?

A lesson from land

Land went through the same transition once. In 1858, Robert Torrens introduced registered title in South Australia, borrowing in part from the way merchant ships were registered. In effect, the register itself became the proof of ownership. As a result, land became markedly easier to buy, sell and lend against.

The value of the land did not change overnight; the certainty around it did. China's own sequence, adding a data-rights register only in September 2026, suggests the same holds for data. Legal certainty and a reliable record of the asset have to sit alongside the accounting, not trail behind it.

Where does the Isle of Man's Data Asset Foundation fit?

This is where the Isle of Man's approach becomes relevant. The Foundations (Amendment) Act 2026 received Royal Assent in May 2026. It created the Data Asset Foundation, a legal structure designed to hold data as a defined asset.

How the regime works

Where China began with the valuation, the Manx regime begins with the title. Under the statutory model, data is first defined and dedicated to a Data Asset Foundation. Next it is submitted for provisional registration and independently accredited against the prescribed Data Governance Framework.

Once the statutory requirements are met, it is fully registered on the Data Asset Register. At that point the Act vests a personal property right in the registered data asset in the Foundation. The regime also provides for a specialist Data Enforcer to oversee compliance with the Foundation's governance obligations.

In many ways the regime tackles the same practical problems the report describes, from a different starting point. How do you make data visible? What do you actually hold? Can a third party rely on it with confidence?

A measured view

I want to be measured about this. The regime is new. Its secondary legislation and governance framework are still being finalised following consultation. Several firms on the Island are building expertise in the area, as they should. For that reason, it will take time to prove itself.

Even so, it is one of the first attempts outside China to build a legal framework specifically for data assets. Moreover, the report suggests demand for that kind of framework will only grow.

What should boards do now?

For boards and finance directors, the practical starting point is the same wherever the standard is eventually set. Start by knowing what data you hold. Then establish the terms on which you hold it, since licences, third-party rights and data protection all bear on that. Finally, be clear about how it is governed. After all, those are the questions any lender will ask, and a Data Asset Foundation is designed to answer them on the record.

So what is the real question now?

The argument has moved on. The difficult question is no longer whether data has economic value. Instead, it is how that value should be recognised in law, in accounts and in finance.

China has built the valuation and is only now writing the title deeds. Meanwhile, America is asking some of the same questions. From the Isle of Man, we started with the title and are working outward. That makes the next few years particularly interesting.

Which brings us back to the bank and the house purchase. Suppose your organisation's data were offered as security tomorrow. Could you show a lender not just what it is worth, but that it is yours?