The National Data Administration symposium requires clarifying AI data usage needs; the physical AI data collection segment may lead the way

Cailian Press Financial News Agency

Open the APP

STAR Market Daily Song Ziqiao, Editor Ji Sheng 2026-11-02 12:18

① Data collection provides training "raw material" for embodied large models such as VLA, and is the foundation for building a data flywheel and supporting a commercial closed loop; ② In early 2026, the gap between globally compliant and usable real physical interaction data and what is needed to train a single general-purpose embodied large model is still 20 times; ③ Institutions say the business model of data collection equipment and services is the clearest.

STAR Market Daily, November 2 (Editor Song Ziqiao) On the morning of October 28, the National Data Administration organized a symposium on the work of chain leader units in building high-quality industry datasets. The meeting emphasized that the construction of high-quality industry datasets must adhere to the principles of "determining supply by demand, promoting construction through use, and improving quality through evaluation," among which clarifying the data needs for artificial intelligence was mentioned as a key task.

The meeting stated that chain leader units should base themselves on their own needs, proactively expand external demand in the upstream and downstream of the industrial chain, adhere to a demand orientation, and precisely advance dataset construction.

Previously, on October 20, China National Data Group Co., Ltd. was officially inaugurated.

Data collection provides training "raw material" for embodied large models such as VLA, and is the foundation for building a data flywheel and supporting a commercial closed loop. Unlike text large models, which can scrape corpora from the internet, physical AI cannot rely on web crawling to complete data accumulation. It needs to establish temporal associations among task instructions, environmental observations, robot states, control actions, and execution results. The data must include not only vision and language, but also information such as joint poses, force perception, and touch, and meet the requirements of being "accurate, complete, broad, and real."

Guojin Securities said that the industry estimates that mature embodied large models require at least tens of millions of hours of data, and further progress toward general intelligence may require data on the order of hundreds of millions of hours, while current high-quality industry training data is only hundreds of thousands to one million hours, leaving a supply-demand gap of about 90%.

According to a Zheshang Securities research report, in early 2026, the gap between globally compliant and usable real physical interaction data and what is needed to train a single general-purpose embodied large model is still 20 times. The industry is currently benefiting from four favorable factors: accelerated construction of standards systems, marketization of data elements, industrialization of data scale, and rapid influx of capital, which are driving the rapid development of embodied intelligence data collection.

In the physical AI field, high-quality data comes first, and leading companies in the industry are accelerating commercialization. Scale AI is a global leader in AI data infrastructure, positioned as an AI training data service provider, with core business covering the three major areas of data, evaluation, and applied AI. In 2026, the company's valuation reached $29 billion, and it has order cooperation with artificial intelligence companies, hardware manufacturers, medical institutions, governments, and others.

From the perspective of the industrial chain, Zheshang Securities said that under a neutral expectation, the upstream hardware equipment market is expected to reach 15.3 billion yuan by 2030. Most of it is reuse of embodied intelligence bodies, and the industrial chain is already relatively mature. While barriers and concentration are relatively higher, as upstream shovel sellers, the certainty of performance for leading companies is expected to be greater. It is recommended to prioritize the camera and IMU segments, which have rigid demand attributes, as well as some segments with a better competitive landscape and higher barriers; under a neutral expectation, the midstream data collection market is expected to reach 174.1 billion yuan by 2030, with a CAGR of 203% from 2026 to 2030. The industry is in a rapid development period from 0 to 1, with clear long-term growth space and greater elasticity. It is recommended to pay attention to dataset suppliers, data collection providers with integrated software and hardware, data post-processing vendors, and simulation platform vendors.

Guojin Securities said that the supply gap and technologies that have not yet converged have spawned many players to enter. From the perspective of business paradigms, the business model of data collection equipment and services is the clearest, and it is the segment with the highest order visibility and the most stable cash flow in the entire data collection industry. Leading companies will continue to benefit from the widening supply-demand gap and the consolidation of the landscape brought about by route convergence. In the short term, pay attention to data collection service providers with orders, production capacity, and delivery capabilities; in the medium and long term, prefer companies that can platformize data processing and extend toward model evaluation, post-training, and deployment feedback.

The institution focuses on three types of companies: (1) traditional leaders with experience in traditional AI data services; (2) simulation leaders that use simulation assets as a foundation to build integrated software and hardware platforms; (3) full-process platforms that layout a closed loop of "data—simulation—training—migration—scenarios."

Artificial Intelligence

Robots

Cailian Press statement: The content of this article is for reference only and does not constitute investment advice. Investors who act accordingly bear their own risks.

Business cooperation

Popular unlocks

In the morning session, the index fluctuated and strengthened, rapidly pulling up near midday, with volume significantly enlarged. With the triple resonance of funds, sentiment, and policy, the quality of the rebound was relatively high, and after short-term consolidation it is expected to reverse the downward trend. Short-term sentiment has fully warmed up, with only a few unusual stocks showing negative feedback. Sectors saw broad gains, with large financials taking the lead. AI applications became the strongest direction in the morning session, with short dramas and games hitting the daily limit in batches and driving a return of funds to AI hardware. In addition, consumption, especially baijiu, performed better than expected, and the sheep concept remained the strongest bonus attribute.

[Institutional research] This company's special electronic cloth layout continues to land, and it has mass production supply capacity for second-generation cloth

This company's special electronic cloth layout continues to land, and it has mass production supply capacity for second-generation cloth. Its products have entered the supply chains of many leading copper-clad laminate manufacturers, and supply is still mainly to strategic customers.

[Research selection • industry data] Analysts compile a list of A-share "dividend aristocrats"

According to analyst statistics, a total of 24 companies in the A-share market meet the threshold of raising DPS for five consecutive years. Among them, only one company has raised DPS for 10 consecutive years. Companies meeting the five-year condition have formed a certain scale, but long-term sustained dividend increases remain relatively scarce.

The National Data Administration symposium requires clarifying AI data usage needs; the physical AI data collection segment may lead the way

Topics

Tags

Body · AI translation