Key Points
- Sichuan Province’s Comprehensive AI Strategy: Sichuan 四川 is implementing a multi-faceted approach to become an AI hub, focusing on infrastructure, data, financial incentives, and real-world application.
- Innovative Voucher Programs: The province is expanding “Computing Power Vouchers” (Suanli Quan 算力券) to subsidize GPU access and exploring new “Scenario Vouchers” (Changjing Quan 场景券) for real-world AI deployment and “Model Vouchers” (Moxing Quan 模型券) for AI model development and fine-tuning.
- Data Labeling and Infrastructure Investment: Sichuan recognizes the importance of behind-the-scenes AI development by building national and provincial-level data labeling bases, with top projects eligible for up to ¥10,000,000 RMB (approx. $1,385,000 USD) in support.
- “Data Elements ×” Initiative: This program (Shuju Yaosu × 数据要素×) aims to unlock economic value from data by encouraging trusted data spaces and demonstrating real-world data monetization scenarios.
- Digital Innovation Labs & Open-Source Incentives: Sichuan is funding digital application scenario laboratories (up to ¥10,000,000 RMB for top projects) and incentivizing the development of first-edition software products and the industrial application of open-source projects to foster a robust AI ecosystem.
- Computing Power Vouchers (Suanli Quan): Subsidies for GPU/compute access
- Scenario Vouchers (Changjing Quan): Funding for real-world AI deployment
- Model Vouchers (Moxing Quan): Support for fine-tuning specialized models
- Data Labeling Bases: Infrastructure for training-data creation
- Digital Innovation Labs: Funding for application-focused R&D labs
China’s western regions are making serious moves in artificial intelligence.
Sichuan Province (Sichuan 四川) just dropped a major policy initiative that signals a shift in how regional governments are backing AI development.
The General Office of the Sichuan Provincial People’s Government (Sichuan Sheng Renmin Zhengfu Bangongting 四川省人民政府办公厅) recently released the “Implementation Opinions on Further Promoting Innovative Breakthroughs in the Service Industry,” and the details reveal a government thinking differently about how to fuel AI innovation at scale.
This isn’t just bureaucratic noise—it’s a roadmap for how regions are competing to become AI hubs.
The AI Infrastructure Race: Why Sichuan Is Making Its Move Now
The stakes for AI infrastructure dominance are higher than ever.
Provinces across China are realizing that having raw computing power isn’t enough anymore.
You need:
- High-quality datasets to train models
- Linguistic corpora (language training data) that actually work in real applications
- Smart financial incentives to get companies to build here instead of elsewhere
- Data infrastructure that’s trustworthy and accessible
Sichuan is tackling all four at once.
The province’s new guidelines propose solidifying artificial intelligence infrastructure to secure a leading position in data and information services.
Translation: They want to be the place where AI companies come to build and train their models.
Computing Power Vouchers: Making AI Accessible Through Financial Engineering
Here’s where it gets interesting.
Sichuan is expanding its provincial-level “Computing Power Vouchers” (Suanli Quan 算力券)—essentially subsidies that make GPU and computing resources cheaper for startups and enterprises.
Think of it like AWS credits, except issued by the government to level the playing field.
This is smart policy because:
- It lowers the barrier to entry for AI startups that can’t afford massive compute bills upfront
- It attracts talent and companies to the region (if you can train your LLM cheaper in Sichuan, why wouldn’t you?)
- It creates a virtuous cycle where more companies come, more data gets generated, and the ecosystem gets stronger
The expanded computing power vouchers program is exactly the kind of economic leverage that’s been working in other regions across China.
But Sichuan isn’t stopping there.
The New Frontier: Scenario Vouchers and Model Vouchers Explained
Beyond computing power, Sichuan is exploring two totally new policy tools that haven’t seen widespread adoption yet.
Scenario Vouchers (Changjing Quan 场景券)
This one’s innovative.
“Scenario Vouchers” would subsidize companies to deploy AI models in real-world use cases within the province.
The idea is straightforward but powerful:
- A logistics company wants to use computer vision to optimize warehouse operations
- A healthcare provider wants to deploy an AI diagnostic tool
- A manufacturer needs AI-powered quality control
- The government says: “We’ll help fund the implementation if you do it here”
This drives real-world AI adoption instead of just theoretical development.
Model Vouchers (Moxing Quan 模型券)
“Model Vouchers” appear to be subsidies for developing and fine-tuning AI models specifically.
A company building a vertical-specific LLM (say, for legal analysis or financial forecasting) could get government backing to create specialized models.
This fills a gap in the market—there’s huge demand for domain-specific AI models, but the development costs are high.
Model vouchers could unlock a whole new generation of niche AI tools built in Sichuan.
These aren’t official yet—the province is still in the “exploration” phase.
But the fact that they’re being seriously considered shows how creative regional competition is getting around AI funding.
Data Labeling Infrastructure: The Unglamorous Foundation of Modern AI
If computing power and models are the flashy parts of AI, data labeling is the grinding, essential foundation that nobody talks about at conferences.
But Sichuan gets it.
The province will focus on building national and provincial-level data labeling bases.
Here’s why this matters:
- Every AI model needs labeled training data (images marked with what they contain, texts labeled with sentiment, audio transcribed, etc.)
- This is labor-intensive and expensive but absolutely necessary
- Building centralized labeling infrastructure can make it more efficient and lower costs for the entire ecosystem
- It creates jobs in data annotation and quality assurance
Sichuan will provide support for data labeling public service platform projects.
Top-tier projects are eligible for up to ¥10,000,000 RMB (approx. $1,385,000 USD) in financial support.
That’s real money for building out the infrastructure most regions overlook.
The “Data Elements ×” Initiative: Connecting Data to Real Economic Value
Alongside the physical infrastructure, Sichuan is pushing something called the “Data Elements ×” (Shuju Yaosu × 数据要素×) initiative.
This is less about computing hardware and more about unlocking economic value from data itself.
The government will encourage:
- Enterprises and industry alliances to build trusted data spaces (controlled environments where data can be shared safely)
- Typical application scenarios under this framework (real use cases that demonstrate how data creates value)
- Data monetization while protecting privacy and security
This connects back to one of the most important insights: raw data is worthless without a use case.
Sichuan’s approach recognizes that the real competitive advantage isn’t hoarding data—it’s building ecosystems where data flows between trusted parties to create products and services.
Digital Innovation Labs: Where Theory Meets Practice
To make this work in practice, Sichuan is creating digital application scenario laboratories.
Here’s the funding structure:
- Approved digital application scenario labs can receive funding
- Top-performing projects get up to ¥10,000,000 RMB (approx. $1,385,000 USD) for digital innovation capability enhancement
- These funds support everything from infrastructure upgrades to talent recruitment to R&D
The logic here is clear: Let a thousand experiments run, and fund the ones that work.
This is competitive innovation funding, not picking winners in advance.
Open-Source and First-Edition Software: Building the Ecosystem
Sichuan is also offering incentives for:
- Eligible first-edition software products (new tools being built for the first time)
- Industrial application of open-source projects (taking open-source code and building real products with it)
This dual approach is smart because it supports both proprietary innovation and open-source collaboration.
It’s acknowledging that AI development doesn’t happen in isolation—companies build on open-source foundations, then add proprietary layers on top.
What This Means for the Broader AI Landscape
Sichuan’s policy package isn’t revolutionary on any single dimension.
But taken together, it represents a comprehensive approach to regional AI competitiveness.
The key insight: AI isn’t just about algorithms anymore—it’s about infrastructure, data, computing power, talent, and use cases all working together.
Other Chinese regions will likely study this playbook closely:
- Provinces like Guangdong (Guangdong 广东) and Beijing (Beijing 北京) are already heavy in AI, but Sichuan’s approach shows how second-tier regions can compete
- The computing power voucher model is replicable and could spread across China
- Scenario and model vouchers might become the new standard for regional AI subsidies
- The emphasis on data labeling infrastructure signals that the unsexy parts of AI are getting government attention too
From an investor’s perspective, this matters because policy shapes where AI innovation happens next.
Sichuan is signaling that it’s serious about becoming a destination for AI development—not through slogans, but through concrete financial mechanisms and infrastructure investments.
The Bottom Line on Sichuan’s AI Strategy
Sichuan Province is making a calculated bet on artificial intelligence infrastructure.
By expanding computing power vouchers, exploring new scenario and model voucher policies, building data labeling bases, and creating digital innovation labs, the province is addressing the full stack of what AI companies need to succeed.
This isn’t flashy.
It’s methodical, well-designed policy aimed at making Sichuan a more competitive place to build AI products.
For founders and investors tracking China’s regional AI race, Sichuan’s moves are worth watching.
Regional competition drives innovation, and provinces that get their fundamentals right—computing power, data infrastructure, financial incentives, and real-world deployment opportunities—tend to win in the long run.
References
- Sichuan: Promoting AI Datasets and Corpora Construction, Expanding Provincial “Computing Power Vouchers” – People’s Daily Finance (Renmin Caixun 人民财讯)
- Official Website of the People’s Government of Sichuan Province
- Digital Economy and AI Infrastructure Developments in China – People’s Daily IT
- China Academy of Information and Communications Technology (CAICT) – Research on Computing Power Vouchers
