Multiple regions have introduced a new round of supportive policies for the artificial intelligence industry.

In recent months, multiple regions across the country have rolled out a new round of industrial support policies for artificial intelligence (AI). For example, the Beijing Municipal Bureau of Economy and Information Technology has officially launched the public solicitation for entities to build high-quality data sets for AI-empowered new industrialization in the third batch in 2026; the Shanghai Zhangjiang Science City Management Committee has opened an application channel for the Zhangjiang AI Innovation Town special voucher policy for the second quarter.

According to a review, since the beginning of this year, AI industry support policies across regions have generally shown three major characteristics. First, they have shifted from technology R&D-driven efforts to scenario implementation-driven efforts. Second, they have moved from broad-based support to more targeted guidance. Third, related policies have continuously strengthened the market-oriented direction of support.

Experts interviewed by Securities Daily said that the evolution of domestic AI industry support policies reflects not only the phased characteristics of China’s AI industry’s development—from technology tackling to value transformation—but also provides clear guidance for enterprises to plan their strategies in different regions and different fields.

Market orientation continues to be strengthened

Since the beginning of this year, focusing on the full life cycle development of the AI industry, many regions have introduced a series of subsidy measures, gradually building an industrial support “toolkit” composed of computing power support to reduce costs, R&D subsidies to overcome technical bottlenecks, scenario-based rewards to promote rollout, and tax incentives to stabilize expectations. With various tools working in coordination, the commercialization achievements of China’s AI industry are accelerating toward real-world implementation.

After reviewing the detailed subsidy rules issued in various places, reporters found that scenario application rewards have become the fastest-growing category. From a nationwide perspective, the support amount for a single project of this type generally ranges from 1 million to 20 million yuan, while the subsidy disbursement ratio is maintained at 15% to 30%. These supports cover multiple vertical application areas such as industrial manufacturing, government services, finance, and healthcare.

It is worth noting that the application scenarios covered by these rewards continue to expand. For example, in the Several Measures of Beijing Municipality on Supporting Industrial Enterprises to Improve Quality and Efficiency issued in June, the policy clearly supports enterprises in purchasing and consuming tokens (Token) to carry out AI applications. For enterprises that meet the requirements, funding support will be provided based on their actual Token usage.

In addition, another notable feature of this year’s AI industry support policies is that policy support is shifting from “broad-spectrum” general technology support to “precise drip irrigation” targeting specific vertical sectors. Taking the AI micro-drama field as an example, the core of related policies is to incentivize large-scale applications of AI technology across content production, virtual human production, and overseas distribution. For example, in May, Shanghai issued the Several Measures for Accelerating the High-Quality Development of AI-Empowered Micro-Dramas in Shanghai, proposing rewards of up to 3 million yuan for outstanding micro-drama projects (including AI micro-dramas) that place equal emphasis on value, content, and quality.

Song Xiangqing, Vice President of the China Society for Business Economics, told Securities Daily that the logic behind this shift is that with AI foundational technologies—especially large-model capabilities—becoming increasingly mature, the core of industrial development has moved from “whether it can be done” to “where to use it and how to use it well.” Therefore, policy resources have begun to tilt toward vertical tracks that have high industrial value, strong positive social externalities, and high compatibility with AI technologies.

It is also worth paying attention to the fact that, this year, AI industry support policies introduced across regions generally strengthen market orientation. Support methods are speeding up the shift from pre-emptive direct allocation of funds to “reward instead of subsidy” and post-event subsidy models that are deeply tied to actual market results. For example, the Notice on Issuing the Implementation Guide for the 2026 Beijing High-End and Precision Industries Development Project Funds and SME Development Funds (First Batch) shows that for high-quality solutions that apply industry vertical large models for the first time to meet the needs of typical industry application scenarios and actually achieve implementation, support will be granted at 15% of the actual amount received for the non-hardware portion in the solution, with the maximum support amount for a single enterprise not exceeding 30 million yuan.

Fu Yifu, a special researcher at Sushang Bank, told Securities Daily that the evolution of AI industry support policies is an important sign that the industry is moving toward maturity. Market-oriented and post-event reward mechanisms overcome the shortcomings of traditional subsidies—“emphasizing project approval while neglecting implementation, and focusing on weak effectiveness.” Using market receipts, scenario implementation, and industrial enablement as the core evaluation standards will make enterprises focus more on real needs and refine high-quality solutions. At the same time, precise vertical track support and diversified factor-based subsidy approaches can accurately match the end-to-end needs of the AI industry from technology R&D to commercialization, effectively activating a positive feedback loop involving large models, scenario applications, and industry enablement.

Enterprises should make good use of relevant policies

At present, China’s AI industry support policies display distinctive characteristics of being scenario-based, fine-tuned, and market-oriented. This is an effective way to support the AI industry’s full transition from the technology tackling stage to the value transformation stage.

Latest invoice data from the State Taxation Administration shows that from January to May, the sales revenue year-on-year of robot and intelligent vehicle equipment manufacturing increased by 27.7% and 46.3%, respectively, indicating that the level of industrial intelligence empowered by AI continues to improve.

Based on the context and logic of policy evolution, experts interviewed by Securities Daily generally believe that the subsequent evolution of AI industry support policies will show three clear directions. First, the focus of subsidies will accelerate the shift from “subsidizing technology R&D” to “supporting implementation scenarios.” Second, policy tools will upgrade from single financial subsidies to diversified, combination-based empowerment measures. Third, cross-regional industrial coordination and supporting mechanisms will continue to be improved and pushed forward more deeply.

Against the backdrop of changes in industrial policy, for enterprises in the AI industry chain to seize advantages in the next round of industry competition, the key lies in thoroughly understanding the underlying logic of policy evolution, accurately aligning enterprise development strategies with national and local policy orientations, and building a multidimensional policy-tool combination application system that fits their own development.

When asked how to make good use of relevant subsidy policies, Fu Yifu said that AI enterprises at different stages of development have significant differences in resource endowments, technological maturity, and commercialization capabilities. Therefore, differentiated application paths are required to maximize the marginal utility of policies. For example, for start-up micro teams, alleviating cash flow and reducing computing power costs should be the core focus. They should prioritize low-threshold universal vouchers for computing power, venues, and taxes issued by various regions, and avoid high-threshold, long-cycle major projects. For growth-stage enterprises that already have mature products and market foundations, their applications can shift toward performance-based scenario rewards and R&D breakthroughs special projects linked to sales and collections.

[Author: Tian Peng] (Edited by: Wen Jing)

Keywords:

                                                            Artificial Intelligence