The U.S.-China Economic and Security Review Commission, a body under the U.S. Congress, has released its annual assessment naming China’s push to commercialize and monetize data as a strategic national asset as a factor that could shape the direction of the AI race. At the release briefing, Commission Vice Chairman Mike Kuiken said drafting a national data strategy is the commission’s top recommendation.
His framing: the commission is urging Congress to think through a national data strategy and how the U.S. government should treat data as an economic asset. The report itself is more blunt, saying Beijing is “marshalling data to drive productivity” while also using it to improve intelligence-gathering and military capability.
What Can Be Scraped, and What Can’t
The report’s most concrete finding centers on a split between types of data.
The public internet text corpus that leading U.S. AI companies have relied on to train language models is running low. What China is systematically gathering instead is a different category: enterprise data, operational data and physical-world data — data the report says “cannot be scraped.” That data feeds enterprise-facing AI tools, autonomous vehicles and humanoid robots.
That split is reshaping the contest. Getting text corpora is a networking problem — whoever crawls fastest, cleans best, or can afford the licensing gets ahead. Getting physical-world data is an industrial problem — it requires production lines running, vehicle fleets on the road, hospitals in operation, so the data can be generated piece by piece. Following that logic, the report concludes China’s advanced manufacturing and industrial robotics ecosystem supplies large volumes of high-quality scenario data, potentially putting it ahead of the U.S. in robotics software development.
A Policy Six Years in the Making
The China timeline the report lays out isn’t new. In 2020, data was designated a factor of production alongside land, labor, capital and technology; in 2023, the National Data Administration was established; since then, data-asset entry onto balance sheets, industry standardization, and the rollout of data exchanges have followed. The report singles out manufacturing, transportation, finance and healthcare as priority sectors, and notes Chinese companies have already begun listing their own datasets on exchanges in Shanghai, Shenzhen and Beijing, with thousands of data-trading institutions now operating nationwide.
The difference lies in how each side reads it. Domestic discussion of data as a factor of production has long centered on institutional issues — rights confirmation, pricing, balance-sheet entry — with mixed views on how fast things are moving. This report skips the institutional detail and treats the whole thing as a supply mechanism already running, then estimates the consequences. The same development looks like a policy rollout’s progress bar from the inside, and like a rival’s growing capacity from the outside.
The Standards Piece
Beyond a data strategy, the report offers a warning about international standards: once a standard is adopted, it’s hard to change, so participating now is more effective than trying to overturn an established standard later.
That points to infrastructure-level rules — data formats, interoperability interfaces, industry data dictionaries. They don’t make headlines, but they determine whose systems can plug in later and who needs an extra layer of conversion. Standards fights in industrial internet, vehicle-to-everything coordination and medical imaging have been running for years, and the report’s implication is that the U.S. hasn’t shown up often enough at those tables.
For people working in the field domestically, the report’s value isn’t in its conclusions — it’s in the list it hands over of which cards outside observers think China is holding: density of manufacturing-scene data, the supply of robotics training data, and first-mover rights on industry standards. Whether those cards get played well depends on how much friction exists in getting data to actually flow between companies — and that part won’t show up in any outside assessment.
Sources: U.S.-China Economic and Security Review Commission report, Reuters, CocoLoop, South China Morning Post; quotes and the recommendation ranking are drawn from the Commission Vice Chairman’s public remarks and the report text, with the data-factor and National Data Administration policy timeline checked against public sources.