In the burgeoning landscape of digital innovation in China, the nation is witnessing a significant evolution in its data factor marketThis transformation was underscored during the recent Seventh Digital China Construction Summit held in Fuzhou on May 23. Notably, this event marked the inaugural summit since the establishment of the National Data Administration, highlighting the conference's critical theme: "Unlocking the Value of Data Elements to Develop New Productive Forces." This summit took a comprehensive look at China's current data industry trajectory.

As reported during the summit, China's data factor market is becoming increasingly dynamic in 2023, with the total data production soaring to 32.85 zettabytes (ZB), reflecting a robust year-on-year growth of 22.4%. This remarkable expansion is largely attributed to the rapid advancements in technologies such as 5G, artificial intelligence (AI), and the Internet of Things (IoT), coupled with the widespread deployment of smart devices that have collectively spurred a significant increase in data generation.

To put this into perspective, statistics reveal that the country’s total data storage capacity reached 1.73 ZB in 2023, with roughly 2.9% of the generated data being actively stored

Liu Liehong, the head of the National Data Administration, provided insight into the impact of AI, stating that China now boasts over 100 large AI models, each with over a billion parametersThese models have been intricately woven into various industries—including electronic information technology, healthcare, transportation, and finance—resulting in the emergence of hundreds of application models that empower numerous sectors and drive the transformation and upgrading of traditional industries.

Crucially, the summit also shed light on the current state of China's data resources and the challenges that lie aheadA report presented at the summit highlighted that while the structure of data production, storage, and processing ("production-storage-computation") has formed a substantial foundation, the various stakeholders in the "supply-flow-usage" chain remain underdeveloped

• OPEC+ Delay: What Lies Ahead for Oil Prices?
• Gold Weakens and Declines
• Gold & Silver: Volatile Ahead of Non-Farm Payrolls
• Hedging Options as Bitcoin Breaks $100K Barrier
• Coal Power's Low-Carbon Shift: External Aid & Internal Upgrades

Analysts emphasized that despite the abundance of data and diverse application scenarios, there’s a pressing need to unlock the potential of these vast reserves.

Officials from several data industry chain companies echoed this sentiment, emphasizing the importance of enhancing the efficiency of data circulation and tradingThey noted that attracting various financial stakeholders and data traders to engage in digital transactions is crucial for establishing a unified, open, and highly effective data factor market.

Despite the impressive growth statistics, the report addressed a primary challenge contending with untapped data valueIt detailed that approximately 40% of stored data has gone unused for a year, pointing towards a deficiency in data processing capabilities, resulting in many data sets being undervalued and insufficiently exploitedNationwide, the total data storage capacity has risen to 2.93 ZB, but the utilization rate remains just 59%, illustrating a significant gap between data generation and actual use.

An analysis of the underlying causes identified a dual issue: on one hand, vast amounts of data lack reusability, with much process-related data discarded shortly after creation; on the other hand, inadequate support from model training tools and processing capabilities has resulted in a situation where the costs of data storage exceed its potential value, leading to high-value data being lost in the shuffle.

Wang Qinglin, the manager of Ruida Heng Research Institute, expressed that the rapid development of China's big data market is notable, as it has swiftly permeated various facets of societal life, resulting in an immense pool of data resources

However, she highlighted that the focus of recent advancements has centered primarily on data production and storage, with most generated data being collected and categorized rather than being genuinely utilized for further processing and management.

"Data production and storage are foundational capabilities within big data applications and commercial uses, yet there remains considerable room for improvement in more sophisticated stages such as data processing and management, data value transmission, digital creation capabilities, and lifestyle integration, all of which require higher technical thresholds," Wang statedFurthermore, she remarked that there are currently very few institutions in China capable of efficiently processing data, and the commercial value of processed data remains uncertain—an aspect that leaves industries globally in a state of trial and error.

Within this context, Wang asserted that the overall development of China's data industry is still at an elementary stage, indicating that a large quantity of stored data remains inadequately explored and harnessed.

Complementing this perspective, a report published by IDC highlighted that only one-third of enterprise data in China is analyzed and leveraged for value generation

The root of this limitation lies in the inability to integrate data across physical or virtual spaces, restricting the scalability and value of data due to fragmented circulation and sharingAdditionally, the demand for data analysis is primarily driven by proactive initiatives from front-end business operations, hampering the likelihood of leveraging data's inherent value.

To address these concerns, IDC’s senior analyst Li Haoran suggested optimizing the avenues for extracting data value from two primary fronts: "First, we must achieve platform-level aggregation of data, creating a unified data resource management platformSecondly, it is essential to leverage generative AI technology to actively identify interconnections among datasets, uncover potential commercial opportunities, and initiate active management of metadata while using low-code and visual tools to empower employees across all levels with agile data querying and market insight capabilities," he contended.

As data continues to burgeon, the demand for processing power has also escalated swiftly, particularly with the revolutionary waves ushered in by AI

Currently, China operates over 2,200 computing centers with a combined processing capacity amounting to about 0.23 zettaflops (ZFLOPS), experiencing an approximate growth of 30% year-on-yearDuring this growth phase, the need for computational power to train large models has surged, further driving escalating requirements across sectors such as science, governance, finance, and industry.

This rise in demand has positioned proactive deployment as a central tenet of China’s current computational landscapeTaking telecommunications operators as a case in point, while China Mobile has announced plans to deploy three large-scale clusters nationally this year, China Telecom revealed intentions to establish super-large intelligent computing clusters in hotspot areas, alongside developing a large-scale green computing pool in the western regions aimed at achieving an aggregate computational capacity of 21 EFLOPS (exaflops) by 2024. This initiative seeks to lay the groundwork for a nationwide integrated computational internet.

Looking ahead, Li Haoran cautioned that while the national push towards infrastructural computing capabilities has been ongoing for several years, particularly amid the generative AI surge, it is essential to remain aware of three pressing risks: Firstly, computational demands tend to be cyclical, making it critical to enhance efficiency, foresee shifts in demand, and balance resource allocation effectively; secondly, managing the computational network involves more than simply providing chip servers, encompassing network bandwidth, storage optimization, caching management, data compression stratification, and heterogeneous scheduling; lastly, energy consumption must also be considered, with a commitment to sustainable, low-carbon development.

Meanwhile, Wang Qinglin reiterated the need for higher-level breakthroughs in big data applications, stressing that pushing the industry towards advanced applications is essential for high-quality growth and achieving broader commercialization

In this regard, she underscored that beyond national policy support and expedited infrastructure improvements promoting data and computing advancements, there is a need to explore additional commercial scenarios that stimulate engagement from businesses while attracting a wider array of participants in related research and development effortsA systematic approach must be adopted to guide resources for deep development, avoiding redundant construction and waste.

While concerns about resource waste and investment risks from ahead-of-curve computational planning are not unfamiliar to the telecommunications industry, which has seen extensive dialogue regarding 5G, Omdia's senior telecom strategist Yang Guang underscored that risks concurrently exist when aiming for optimal balance between preemptive expansion and market demand development, as well as while maintaining a diversified supply chain amidst a focus on domestic control of industrial chains.

"Being excessively ambitious can lead to operational risks, while a lackluster approach may heighten China’s competitive vulnerabilities on the global stage," Yang explained

Thus, moderate foresight in the allocation of computational resources reflects a synthesis of various factors, representing the most viable strategic solution in the current context, where managing diverse risks effectively is key.

As data continues to emerge as a new productive factor, the summit also prioritized discussions around the challenges entwined with its marketizationThe head of the Market Department at China Electronics Cloud articulated that the essence of turning data into an asset lies in accelerating its entry into trading and market transactions—an industry consensus on unlocking data resources.

Current findings from the National Data Administration indicate a robust demand for data circulation and trading within China, with vibrant data interactions especially observable in consumer sectorsIndicators from 2023 show that the total data flow across the four major telecom operators increased by 7.6% compared to the previous year, while rapid proliferation of data trading institutions has been observed across numerous provinces, with 19 regions establishing data exchanges that exhibit unique transaction models, business ecosystems, and technological underpinnings.

Significantly, among the various data categories being traded, public data has emerged as a crucial catalyst for development and utilization, with the National Integrated Government Affairs Data Hub facilitating access to 53 national departments and 31 provincial or municipal entities

This hub has successfully managed to process over 536.1 billion service requests to date.

However, the influx of substantial data, including public data into market circulation, brings to light challenges in data governanceLi Haoran noted that while government bodies and large state-owned enterprises are accelerating the construction of data resources, albeit with some efforts to explore pathways for data asset realization, there has yet to be a consensus on establishing fixed paradigms for building standards, value assessment, transaction pathways, asset management frameworks, and comprehensive lifecycle systems.

"The establishment of a robust data governance system must precede these actions to guide the process with national standards and trading guidelines, while exercising caution in data value assessments, and considering both immediate and derived values from data transactions, as well as the dynamics of asset flow throughout their lifecycles," Li advised.

Parallelly, data security emerges as a formidable challenge in public data circulation, necessitating the creation of secure and compliant platforms that link all stakeholders involved in data management and utilization.

Wang Qinglin concurred, emphasizing that the intricacies of deep data processing and trading highlight merely the tip of the iceberg in the present state of the industry