Current Development Trends of China’s AI Industry

In recent years, artificial intelligence (AI) has become the core engine of a new technological revolution, penetrating deeply into the fabric of human production and life, profoundly reshaping global economic structures, innovation paradigms, and social governance logic. China has entered the first tier of global AI development and is at a critical opportunity period transitioning from catching up to leading. In the face of increasingly fierce international competition and the intrinsic demand for high-quality development, we conducted field research to understand the current state of China’s AI industry.

Current State of China’s AI Industry

General Secretary Xi Jinping pointed out that “artificial intelligence is a strategic technology leading this round of technological revolution and industrial transformation, with a strong ’leading goose’ effect.” AI is not merely a linear iteration of a single technology or a partial upgrade of an industry; it represents a comprehensive and disruptive reconstruction of the underlying logic of economic and social operations. To assess its development level and trends, we must break away from traditional technology evaluation and industry analysis frameworks, and conduct comprehensive evaluations from dimensions such as technological capability, industry scale, element support, and integrated applications.

From a technological capability perspective, AI technologies led by open-source have achieved collective breakthroughs, forging new standards within the global developer network. During our research at a laboratory, we observed that the research team introduced a self-criticism mechanism for AI, allowing the model to improve its problem-solving accuracy significantly through multiple rounds of self-play without human intervention. AI has evolved from “being able to hear and see” to “thinking, reasoning, planning,” and ultimately to “mastering how to learn.” Overall, the gap between China and international top levels in key indicators such as model performance, training efficiency, and multimodal integration continues to narrow, with some areas achieving parity or even leading, forming a unique technological route of open-source leadership and ecological co-prosperity. By 2025, China’s share of global downloads of open-source models is expected to reach 17.1%. Recent statistics show that among the top 10 open-source models globally, 8 are from China. The performance of the DeepSeek—V4 model is comparable to that of leading international models, with API prices as low as 1% of those of the GPT—5.5 model. This signifies a break from the technological monopoly of a few tech giants, allowing millions of global developers to conduct secondary development based on Chinese open-source models. Open-source not only benefits developers but also accelerates the flow and spillover of knowledge, continuously enhancing China’s AI technology within an open ecosystem.

From an industry scale perspective, the AI industry has experienced nonlinear explosive growth, with significant value spillover effects behind the trillion-dollar blue ocean. By 2025, the global AI market size is projected to reach $757.58 billion, while China’s core AI industry size is expected to exceed 1.2 trillion yuan. This figure reflects not only the number itself but also the underlying growth logic. Traditional industries follow the law of linear input and diminishing marginal returns, while AI breaks this curse, with technological breakthroughs and application diffusion mutually reinforcing, forming a positive feedback loop of “the more it is used, the stronger it becomes.” Research indicates that Beijing, as an innovation source, will see its core AI industry size reach 450 billion yuan by 2025, with mature algorithm models acting as “digital technology pumps,” continuously delivering intellectual energy to factories in Hebei, ports in Tianjin, and pastures in Inner Mongolia. Shanghai is leveraging the “Moulding Shanghai” initiative to create an ecological attraction through the “Mould Speed Space,” while Shenzhen focuses on industrial implementation, aiming to build a highly concentrated and precisely serviced enterprise ecosystem for the real economy. Ultimately, the AI industry exhibits a clear multiplier effect, where an investment of one yuan can leverage several yuan, with the trillion-dollar scale representing a complete industry chain from underlying computing power to upper-level applications, from core algorithms to intelligent terminals, giving rise to new services, new divisions of labor, and new markets.

On April 24, 2026, DeepSeek officially released a new series of models, the DeepSeek—V4 preview version, and simultaneously open-sourced it. This model adopts the MoE architecture and natively supports ultra-long contexts of 1 million tokens. The model has achieved domestic and open-source field leadership in agent capabilities, world knowledge, and reasoning performance.

From the perspective of element support, China’s core AI resources have achieved a strategic leap, with institutional innovation accelerating the release of element vitality. The competition in AI is not only about how fast models can run but also about how solid the computing power foundation is and how smoothly data flows. China has established significant scale advantages in these two core resources. In terms of computing power, 42 intelligent computing clusters have been built, and as of the first quarter of this year, the scale of intelligent computing power has reached 188.2 quintillion floating-point operations per second, ranking among the top globally. In terms of data, there are over 100,000 high-quality datasets nationwide, with a total volume exceeding 890 petabytes, equivalent to 310 times the total digital resources of the National Library of China. Additionally, institutional advantages are gradually becoming apparent. In a leading area for data infrastructure in Beijing, a “regulatory sandbox” mechanism has been established to break the deadlock regarding companies’ reluctance to open their resources, allowing them to enter a protected “testing ground” for integrated training without transferring data ownership. A technology manager at a company stated, “Previously, training on our small data led to increasingly biased models; now, the sandbox aggregates real data from over 10 industries, significantly improving accuracy, making data more valuable as it is used.”

From the perspective of integrated applications, AI in China is accelerating its penetration into various industries, with the breadth of applications and depth of integration constructing new global competitive advantages. By the end of 2025, the CNC rate of key processes in major industries in China is expected to reach 68.6%, with AI integrated applications transitioning from “point blooming” to “full-chain intelligence.” First, the penetration areas continue to expand, covering most major categories of the national economy and forming a number of benchmark applications in manufacturing, healthcare, transportation, finance, and energy. Second, the empowerment level has significantly improved, advancing from auxiliary roles to core processes such as R&D design, production manufacturing, and operational management. In a heavy equipment manufacturing company in Shandong, we observed that an industrial large model system comprehensively took over the entire process from blueprint analysis, process planning, to quality inspection, compressing the time for new process design from several weeks for multiple senior engineers to less than 72 hours, with a 5% increase in yield rate. Third, new business models and new patterns are emerging rapidly, with intelligent connected vehicles, AI pharmaceuticals, and embodied intelligent robots flourishing, continuously forming new trillion-level industrial tracks. Throughout the research, it was evident that in this global intelligent competition, whoever has the richest application scenarios, the closest integration, and the most concentrated industrial feedback will define the standards and application paradigms of how AI is “used, where it is used, and how deeply it is used,” thus gaining the initiative in the intelligent era.

Challenges Facing China’s AI Industry Development

Currently, the global AI technology competition is becoming increasingly intense, and China’s AI industry is at a critical juncture of application leadership, foundational catch-up, and ecological breakthrough. In the face of ongoing external pressures such as computing power blockades and talent competition, we still face many “bottleneck” links and points of blockage, from high-end chips to foundational algorithms, from original innovation to industrial transformation.

International competition is squeezing the development space of the AI industry. Research indicates that some Western countries have upgraded their policies towards China from single technology restrictions to systemic ecological blockades. First, the “hard” blockade continues to intensify. The United States has increased its control over AI chip sales to China, forcing many domestic innovation teams to slow down their large model development due to “computing power hunger.” Second, the “soft” ecological barriers are being constructed. NVIDIA’s graphics processing units (GPUs) dominate over 90% of the global market share, and its unified computing device architecture (CUDA) ecosystem, built over more than a decade, has formed a closed-loop system of “hardware + software + developer community.” At a domestic chip company in Shanghai, we learned that although its hardware computing power metrics are close to international mainstream levels, customers are primarily concerned with “whether it can be compatible with CUDA.” The crux is that replacing chips is not a simple hardware swap but involves a complete system migration of the development framework, operator library, debugging tools, and development habits. Millions of developers are deeply bound to the CUDA ecosystem, making migration costly and time-consuming, and even if domestic replacements meet performance standards, large-scale applications still face obstacles. Third, the competition for discourse power over rules is fierce. Global AI technology standards, governance norms, and cross-border data rules are predominantly led by Western countries. At the beginning of 2025, the DeepSeek large model caused a stir in the global market due to its technological breakthroughs, prompting several Western countries to issue bans or initiate strict reviews. The reality warns us that technological leadership does not guarantee market access; lacking discourse power means that the industry will be constrained when going global.

Large models face reliability crises in specialized scenarios. While large models perform impressively in general dialogue, their capability deficiencies become apparent when entering fields such as industrial inspection, medical diagnosis, and financial risk control, where precision and reliability are critical. A manufacturing company reported that its AI visual inspection system misjudged good products as waste due to slight changes in lighting, leading to waste being released, necessitating manual re-inspection. The phrase “stunning during demonstrations but failing on the production line” has become a reality for many companies deploying AI. The issue lies in the generalized capabilities exhibited by large models in open-domain tasks, which do not naturally transfer to specialized scenarios with near-zero tolerance for error. The gap from “being able to speak” to “being able to use reliably” presents a significant engineering challenge. The “hallucination” problem cannot be overlooked either. In general scenarios, such errors may be minor flaws, but in contexts such as medical dosages, legal judgments, and financial risk control, every instance of “seriously talking nonsense” could trigger irreparable risks. This exposes a fundamental flaw of large models: they are essentially pattern matchers rather than logical reasoners. Transitioning from “being able to talk” to “speaking the truth,” and from “guessing answers” to “understanding causality” is a threshold that the industry must cross for deeper development.

High-quality datasets still struggle to meet model development needs. Research has found that a common issue is that while there is an abundance of “raw oil” data, there is insufficient “refining” capability. The scale of available private data globally far exceeds that of public data, but due to institutional barriers such as non-unified data standards, inadequate authorization mechanisms, and unclear compliance boundaries, a large amount of high-value data remains trapped in “islands.” Although China possesses vast data resources, the data truly usable for large model training is severely lacking. In globally universal datasets of 5 billion scale, the proportion of Chinese language data is only 1.3%. Furthermore, the bottlenecks in data circulation hinder the full transformation of China’s data scale advantage into core competitiveness. Additionally, copyright and legal risks are on the rise. An overseas company informed us that its video generation model was accused of unauthorized scraping of videos from overseas platforms for training, facing collective lawsuits abroad. If data sovereignty and copyright barriers evolve into new trade weapons, they could cut off domestic companies’ legal access to high-quality international data resources.

The commercial closed loop of AI applications has yet to be established. The AI industry application is at a crossroads of transitioning from policy-driven to market-driven, with sustainable business models still under exploration. First, there is a misalignment in the “gears” of the industry chain. The computing power layer is expensive and insufficiently compatible with models, the model layer is general but lacks industry customization capabilities, and the application layer consists mainly of single-point tool-type products that do not communicate with each other, resulting in a lack of effective engagement mechanisms among the three segments of computing power, models, and applications. Second, the profit model for enterprises is unclear. Domestic users have not yet formed a payment habit, leading many application companies to rely on project-based contracts or government subsidies for sustenance. The transition from “policy blood transfusion” to “market blood production” is crucial for the industry to move out of the nurturing phase. Third, scaling product replication is challenging. An industrial AI founder admitted, “Three factories have successful pilot projects, but when clients say to change a production line, the plan becomes obsolete. Without standardization, there can be no scalability; without scalability, we are forever burning money.” The difference between a “model room” and a “commercial property” is not a single technology but a standardized product system that is configurable, replicable, and maintainable, which requires standardized interfaces across all segments of the industry chain.

Accelerating the Development of China’s AI Industry through Systematic Collaboration

AI is a very special general-purpose technology, distinctly different from any frontier technology of historical technological revolutions. First, it has a strong path dependence and ecological lock-in effect. The underlying chips define the form of computing power, the intermediate frameworks determine the development paradigm, and the upper-level applications heavily rely on the interface standards of the first two layers—this highly coupled technological architecture means that once a first mover achieves dominance at any layer, it can penetrate upwards and downwards, locking the entire industry chain into its ecological system. Second, competition has evolved into a systematic game with interlinked dimensions. Traditional technological competition focuses on single technologies, where overcoming a challenge can lead to breakthroughs; however, AI competition encompasses a full-dimensional contest covering chips, frameworks, data, applications, and rules, where any shortcoming in one dimension could become the “Achilles’ heel” of the entire system. Third, the diffusion cycle has been extremely compressed. The electrical revolution took a century to achieve full societal penetration, and information technology took half a century to reshape business forms; however, AI is rewriting the underlying logic of industries at an unprecedented speed, with the conversion of first-mover advantages into lock-in advantages occurring much more rapidly, leaving little time for latecomers to react. In this global competition that will determine the future, we are not facing a “bottleneck” in a specific technology point but a full-stack competition from underlying hardware to upper-level ecosystems, from technical standards to governance rules. To break the deadlock and seize the initiative, it is futile to pursue single breakthroughs; we must engage in a systematic collaborative battle that encompasses “all elements + all ecosystems.” This requires the full flow of various elements such as computing power, data, algorithms, and scenarios, while also stimulating the innovative vitality of diverse entities such as enterprises, universities, research institutions, and developer communities. Moreover, national strategies must lead the way, uniting all forces into a cohesive effort.

In recent years, Hebei has emerged as a key node in the national computing power industry layout, accelerating the construction of a leading computing power industry ecosystem through policy guidance, infrastructure as a foundation, and integrated development as a goal. The “2025 Comprehensive Computing Power Index” shows that Hebei’s comprehensive computing power index remains first in the country. The image shows the Qinhuai Big Data Industrial Park in Huailai County, Zhangjiakou City, Hebei Province, taken on September 7, 2025.

Strengthening Core Technology Breakthroughs to Build a Self-Controlled Development Foundation

Core technology breakthroughs must upgrade the goal from chasing single indicators to a systematic battle driven by ecosystem building. First, it is essential to root in fundamental principles. If source innovation only focuses on the application and engineering layers, it will forever be limited to patching within others’ theoretical frameworks. More resources must be directed towards foundational research in areas such as algorithm interpretability, causal reasoning, and brain-like computing to master the underlying logic that defines technological routes, thereby fundamentally breaking free from path dependence. Second, targeted breakthroughs and large-scale iterations must be balanced. Focus on core links in the AI chip, development framework, and foundational software industry chain, implementing mechanisms like “reveal the list and take the lead” and “horse racing” to concentrate efforts on overcoming key bottlenecks. More importantly, technological breakthroughs must form a closed loop with market applications; only by massively investing domestic software and hardware into real training scenarios and continuously iterating and optimizing through large-scale trial and error can market feedback enhance technological maturity, gradually forming an ecological attractiveness that can compete with first movers.

Optimizing Data Element Supply to Unblock High-Quality Supply Bottlenecks

China has significant advantages in data resources, but it must address the two bottlenecks of “refinable” and “circulable.” First, it is necessary to build high-quality “data oilfields.” Relying on national-level data annotation bases, standardized dataset systems should be established in mature fields such as industry, healthcare, and finance, while increasing investment in data synthesis and intelligent enhancement technologies. Only by processing raw data into high-quality data that can be directly used for model training can data elements truly enter the production function. Second, institutional innovations must be implemented to unblock circulation bottlenecks. Accelerate the supply of foundational systems around property rights definition, revenue distribution, and safety compliance, promoting innovative models like “data sandboxes” and “regulatory sandboxes” to achieve multi-source data fusion training while ensuring ownership remains unchanged and safety is controllable, allowing data to realize value multiplication through flow.

Accelerating Large-Scale Application Promotion to Build a Sustainable Commercial Closed Loop

Application scenarios are the ultimate battlefield for testing the quality of AI. The core challenge facing the development of the AI industry today is not the lack of good pilot projects, but rather the inability to replicate successful pilots in bulk. It is essential to deeply implement the “AI +” initiative. First, deeply embed AI into core business processes, promoting its transition from auxiliary scenarios to high-value links such as R&D design, production scheduling, and risk control, thereby significantly reducing costs and increasing efficiency to stimulate enterprises’ willingness to pay. Second, establish a collaborative engagement mechanism across the industry chain. Promote deep coupling among computing power providers, model vendors, and industry users to form a collaborative network where computing power is supplied on demand, models are adapted to needs, and scenarios are rapidly implemented, breaking the “each manages their own” situation through standardized interfaces. Third, firmly advance productization transformation. Transition from customized project-based solutions to standardized solutions that are configurable, replicable, and maintainable, allowing for the dilution of R&D and computing costs through scalability, driving the industry from a money-burning cycle into a profit-generating cycle.

Enhancing Safety Governance Capabilities to Establish a Secure Development Bottom Line

The black-box nature, self-evolving capabilities, and generalization abilities of AI extend risk sources from external attacks to the “genetic defects” of the models themselves. Safety governance must transition from static compliance checks to dynamic protection throughout the entire lifecycle. First, establish an agile governance framework that is layered and categorized. Emphasize transparency and traceability for general foundational models, while implementing differentiated regulations based on risk levels for vertical application scenarios, such as strict certification and robustness assessments for high-risk areas like healthcare and finance, and lighter regulation for lower-risk scenarios, achieving a precise balance between safety and development. Second, strengthen internal safety barriers through technology. Increase R&D investment in safety technologies such as algorithm interpretability, privacy computing, and adversarial training, establishing a regular model safety inspection mechanism that preemptively addresses risks with a “technical firewall,” making safety capabilities a model’s “factory setting” rather than an afterthought. Third, proactively lead the construction of global rules. Promote the transformation of China’s practical experience in data classification, algorithm filing, and safety assessment into international governance solutions, seizing the initiative in rule-making within multilateral frameworks to avoid being locked in from behind.

Strengthening Multi-Dimensional Collaborative Guarantees to Build a Support System for All Elements and All Ecosystems

Systematic breakthroughs require matching institutional supply and element support. In terms of funding, it is essential to cultivate truly patient capital that adapts to innovation. Leverage national funds to lead and form a matrix of patient capital with local governments, ensuring long-term investments in foundational breakthroughs and infrastructure construction. Simultaneously, promote inclusive tools like “computing power vouchers” to lower the barriers for small and medium-sized enterprises to participate in innovation. In terms of talent, focus on cultivating “dual-skilled talents” who understand both algorithm logic and industry pain points. Such composite talents cannot be mass-produced in classrooms; they must be nurtured through partnerships between leading enterprises and universities to create platforms for industry-education integration, allowing them to “immerse” in real industrial scenarios over the long term. Accelerate the establishment of a composite talent training system with scale effects, forming a tiered supply from top scientists to large-scale application talents. In terms of open cooperation, it is essential to root in China while connecting with the world. Relying on mechanisms like the “Belt and Road Initiative,” support enterprises in deeply embedding themselves in the global innovation network through open-source collaboration and joint R&D, breaking non-commercial barriers under compliance, and enhancing competitiveness in open competition, thus seizing strategic initiatives in the new round of technological revolution and industrial transformation.

Research Notes

From the perspective of the grand historical coordinates of human civilization, the profound significance of AI may far exceed our current cognitive boundaries. It is not only a technology iteration or an industrial upgrade but also a systemic reshaping of human cognition and social organization. As machines begin to learn, reason, and create, we face not just a technological competition but a re-examination of humanity’s own position. Throughout the research journey, from the computing arteries woven by “Eastern Data, Western Computing” to the data vitality activated by “regulatory sandboxes,” from the ecological wave sparked by open-source large models among global developers to humanoid robots working alongside humans on production lines, the sights and sounds have instilled in us a sense of vigorous upward momentum. This suggests that in this wave of technological innovation, we are no longer latecomers, followers, or catch-up players, but competitors on the same stage, and in certain fields, we are even leaders. As global AI development and governance remain in a state of chaos, our path choices are opening up a new possibility—replacing closure with openness, replacing monopoly with collaboration, and replacing control with empowerment in a new paradigm of intelligent civilization. Years from now, when people look back at the starting point of this AI transformation, they may evaluate it this way: at the historical juncture of the new era, China did not hesitate or miss the opportunity but instead stepped forward boldly.