China wants to trade data across borders as AI fuels demand. Southeast Asia is a key focus

SHENZHEN: A smart cleaning robot trained to navigate an office building in Shenzhen may work flawlessly.

But place the same machine in a shopping mall in Kuala Lumpur, and it could struggle with unfamiliar layouts and consumer behaviour.

The difference is not the hardware, but the data used to train it.

As artificial intelligence (AI) systems spread across borders, access to local data is becoming increasingly valuable alongside algorithms and computing power.

That is driving efforts to find new ways for companies to share information safely and turn it into economic value - a field where China is positioning itself as an early mover.

Beijing in 2020 formally recognised data as a new economic resource, alongside traditional drivers of growth such as land, labour and capital.

It has been building mostly domestic exchanges where companies can trade access to data products and insights into everything from consumer trends to industry patterns.

A recent tie-up with a Southeast Asian firm is now testing whether cross-border data trading can unlock business opportunities on both sides.

A memorandum of understanding signed on May 22 between Shenzhen Data Exchange (SZDEX), a state-owned enterprise under the Shenzhen municipal government, and Malaysian technology firm Zetrix AI has put that ambition in focus, as both sides explore ways to connect China and Southeast Asia’s emerging data markets.

Zetrix AI chief executive Wong Thean Soon told CNA the aim is to build what could become a new model for cross-border data trading, allowing companies to securely exchange access to data products between China and the Association of Southeast Asian Nations (ASEAN).

“Think of it as a stock exchange, but for data,” he said.

But unlike goods moving through ports or money flowing through financial systems, data is far harder to trade, said analysts. It also remains unclear how data should be priced, shared and managed across different markets, experts and industry players added.

THE ECONOMICS OF DATA

Despite the name, data trading does not usually involve companies simply buying and selling raw databases.

“Raw data basically is not used for trading … it is mostly presented in the form of services and products,” said Ou who oversees cross-border data business at SZDEX.

Ou, who wanted to be identified only by his surname, added that this is done so after strict compliance checks.

Put simply, companies are often paying for insights drawn from data, rather than buying the underlying information itself.

For businesses entering new markets, such data could help answer practical questions - from what customers want to how products should be adapted for a new market.

For example, a Chinese maker of AI-powered cleaning robots expanding into Malaysia may require insights into local building environments to fine-tune its systems, without directly obtaining the underlying datasets.

Ou said exchanges like SZDEX aim to act as a middleman - matching companies with data providers, while reviewing whether information can be safely and legally shared.

For Zetrix AI, the goal of collaborating with SZDEX is to allow companies to access and monetise data securely, without necessarily handing over raw datasets.

“What’s actually exchanged is access rights and usage permissions,” Wong said.

Wong said potential applications could include, for example, a Malaysian logistics firm licensing shipping route data to a Chinese manufacturer for supply chain optimisation, without the buyer directly owning or accessing the raw datasets behind those insights.

Malaysian digital solutions firm WITO Technology, which became SZDEX’s first Malaysian data provider in 2024, told CNA a key challenge is explaining that data trading does not mean selling personal information or handing raw data overseas.

WITO Technology founder and chief executive Bryan Ng said companies initially questioned whether their information would be “sold” to China.

“Is our data exposed? Are you selling the data to China?” he said, describing some of the concerns raised.

Ng said data products have to go through reviews involving legal, technical and compliance checks before they can be listed on an exchange, with the focus on approved datasets rather than personal information.

AI AS THE KEY DRIVER

AI is also creating new demand for such services.

“Large language models developed mainly in China will still need broader overseas data, such as local languages, customs and industry-specific information,” Ou from SZDEX said.

China’s National Data Administration has made unlocking the value of data a key focus for 2026, with efforts centred on building a national integrated data market and increasing the supply of high-quality datasets for AI development.

But beyond computing power, experts said access to specialised datasets will also be critical in determining how AI systems are developed and deployed.

Alex Capri, a senior lecturer at the National University of Singapore (NUS) Business School, said data is becoming increasingly valuable in the AI race because its value differs from traditional commodities.

“Data is, in economic terms, a ‘non-expendable commodity’, meaning as compute power increases, the same data can be analysed again and again and continue to yield value,” he told CNA.

While data is already exchanged globally through private marketplaces and trusted data-sharing frameworks, China has taken a more state-led approach by building dedicated exchanges and treating data as a factor of production.

China has built more than 50 data exchanges nationwide over the past decade, forming a network led by Beijing, Shanghai, Guangzhou and Shenzhen, according to research published in May by Qiushi, the official publication of the Communist Party, and a research institute affiliated with China’s state planner.

Most of these exchanges focus on the domestic market, where companies, government bodies and other organisations list and trade data products and services within China.

Shenzhen has emerged as one of China’s main testing grounds for cross-border data trading. SZDEX currently works with partners across eight countries and regions on areas such as data products and compliance, including cross-border data guidelines with Singapore.

Kendra Schaefer, a partner at Beijing-based strategic advisory consultancy Trivium China, said China is essentially trying to build the market infrastructure needed for data to be treated like a tradable asset.

She pointed out that, unlike traditional assets such as land or capital, data still lacks mature systems to prove ownership, determine value and govern transactions.

“Data trading is actually quite experimental,” she said.

TURNING AMBITIONS INTO REALITY

Zetrix AI believes that ASEAN, a regional bloc comprising 700 million people with diverse languages, cultures and economies, could become more than just a consumer of technology.

“Ultimately, this positions ASEAN as a serious data hub, not just a consumer of technology,” Wong said.

For companies trying to enter new markets, WITO Technology’s Ng said cross-border data services could provide more grounded insights than traditional market research, with Chinese companies expanding into ASEAN among those seeking such information.

“When China wants to export products to ASEAN, they require a very, very comprehensive kind of analysis,” he said.

But turning that vision into reality remains complicated.

While China has experimented with cross-border data transactions, turning individual deals and partnerships into a broader international marketplace remains a work in progress.

SZDEX’s Ou said cooperation with Zetrix AI is still at an early stage.

The partnership currently focuses on product and channel cooperation, rather than the construction of a fully operational regional platform, he said.

“This is a commercial activity,” Ou said, adding that Zetrix had expressed interest in learning from Shenzhen’s experience.

“Building broader channels for data circulation will be a long process,” he said.

As for Zetrix AI, Wong, the CEO, said the company is not announcing a specific launch date yet, with both sides focused on building the partnership framework, technical infrastructure and compliance standards needed for cross-border data trading.

The company aims to have a pilot phase operational within the next 12 to 18 months, he added.

Schaefer said cross-border data trading allows China to test whether systems developed domestically can work in countries with different rules and attitudes towards data.

“It’s one thing to do it in a captive ecosystem where you get to write all the laws and you get to write all the rules,” she said.

“It’s another thing to explore economic exchange with countries that don’t necessarily have the same ruleset that you do.”

CONVINCING COMPANIES

Beyond building the infrastructure, another challenge is convincing more companies to participate, said analysts.

Lim May-Ann, director of Access Partnership Institute, a global think tank focused on ensuring emerging technologies benefit society, said the success of cross-border data initiatives will depend on whether companies see enough value and safeguards to participate.

“Commercially-viable trust-building opportunities are the biggest challenge,” she said.

“No company will expose its own data for no reason.”

Ng, the CEO of Wito Technology, said interest in his firm’s services has so far largely come from Chinese companies looking for insights into Southeast Asian markets as they expand overseas.

But he said many companies, including those in ASEAN, are still learning how the information they hold can have commercial value.

“It’s too new,” he said. “They need to understand the importance of data - why they need it, why it is important.”

Capri from NUS said ASEAN’s diversity - from languages and cultures to regulatory systems - could make the region’s data increasingly valuable for developing more localised AI applications.

The push comes as ASEAN works on frameworks such as the Digital Economy Framework Agreement to deepen digital integration and support trusted data flows across the region.

But Lim from Access Partnership Institute said building a wider cross-border data ecosystem will require not only technology, but also trust between companies and markets.

She pointed to safeguards such as data privacy frameworks, international standards and contractual agreements as ways to support safer data sharing.

For Simon Chesterman, a law professor and AI governance expert at NUS, the greatest challenge is political: ensuring that countries and communities generating valuable data have a say in how it is used.

“The risk is not simply that data moves across borders, but that accountability does not move with it,” he said.

Source: CNA/mc(ws)