Chinese artificial intelligence startup DeepSeek is planning to build a massive AI data center campus in Inner Mongolia, marking one of the country’s largest AI infrastructure projects as it ramps up competition with leading U.S. AI companies. According to people familiar with the matter, the Hangzhou-based company aims to develop around one gigawatt (GW) of computing capacity in Ulanqab, a city that has emerged as one of China’s major data center hubs due to its abundant renewable energy resources, cool climate, and relatively low electricity costs.
The project underscores DeepSeek’s ambitions to expand beyond its highly efficient AI models by investing heavily in proprietary computing infrastructure. The company reportedly plans to build its own facilities while also leasing additional capacity from third-party operators, with at least part of the campus expected to become operational by late 2027 or early 2028.
DeepSeek Plans One-Gigawatt AI Data Center
The proposed project represents a major increase in DeepSeek’s computing resources.
According to reports, the company plans to:
- Build an AI data center campus in Ulanqab, Inner Mongolia.
- Develop approximately 1 GW of computing capacity.
- Construct company-owned facilities.
- Lease additional computing capacity from external providers.
- Bring an initial phase online by the end of 2027 or early 2028.
Project Snapshot
| Item | Details |
|---|---|
| Company | DeepSeek |
| Location | Ulanqab, Inner Mongolia, China |
| Planned Capacity | Approximately 1 gigawatt (GW) |
| Project Type | AI data center campus |
| Expected Initial Operations | Late 2027 or Early 2028 |
Why Inner Mongolia?
Inner Mongolia has become one of China’s fastest-growing data center regions thanks to several structural advantages.
These include:
- Lower electricity costs.
- Significant wind and solar power generation.
- Cooler temperatures that reduce cooling expenses.
- Availability of large industrial land parcels.
- Existing cloud computing infrastructure.
Ulanqab has already attracted investments from several major Chinese technology companies and has earned a reputation as one of China’s leading cloud and AI infrastructure clusters.
Why Ulanqab Is Attractive
| Advantage | Benefit for AI Data Centers |
|---|---|
| Renewable energy | Lower operating costs and improved sustainability |
| Cool climate | Reduced cooling requirements |
| Existing infrastructure | Faster deployment of large-scale facilities |
| Large land availability | Supports expansion of AI campuses |
Expanding AI Infrastructure
Data center investment is surging industry-wide, with Microsoft reporting over $130 billion in new data center lease commitments.
DeepSeek rose to prominence after launching highly capable open-weight AI models that challenged leading systems from U.S. developers while reportedly using significantly fewer computing resources.
The planned campus signals that the company is now investing aggressively in long-term infrastructure to support:
- Training larger foundation models.
- AI inference services.
- Enterprise AI deployments.
- Future multimodal AI systems.
- Large-scale research and development.
While the type of processors to be deployed has not been disclosed, the project is expected to rely on a combination of company-owned infrastructure and leased capacity. Reports note that it remains unclear whether the facilities will primarily use domestic Chinese AI chips or imported hardware where available.
Intensifying Competition in Global AI
The investment reflects the escalating race among AI developers to secure computing capacity.
Leading AI companies worldwide are investing billions of dollars in:
- Hyperscale AI data centers.
- Advanced GPU clusters.
- AI cloud infrastructure.
- Energy-efficient computing facilities.
For Chinese AI firms, building domestic computing infrastructure has become increasingly important as U.S. export restrictions continue to limit access to some of the world’s most advanced AI processors.
Global AI Infrastructure Race
| Trend | Industry Impact |
|---|---|
| Larger AI data centers | Supports increasingly complex AI models |
| Growing energy demand | Drives investment in renewable-powered facilities |
| Computing shortages | Encourages long-term infrastructure commitments |
| Geopolitical restrictions | Accelerates domestic AI ecosystem development |
Strategic Importance for DeepSeek
The buildout comes as DeepSeek also looks to boost revenue elsewhere, having planned a price hike for its AI services.
Building a dedicated AI campus would provide DeepSeek with greater control over one of the industry’s most valuable resources—computing power.
Potential advantages include:
- Reduced dependence on third-party cloud providers.
- Faster AI model training.
- Improved infrastructure for enterprise customers.
- Better long-term cost management.
- Greater flexibility in developing next-generation AI systems.
The move also demonstrates DeepSeek’s transition from an emerging AI startup into a company investing in large-scale physical infrastructure comparable to global AI leaders.
Looking Ahead
DeepSeek’s reported plan to build a 1 GW AI data center campus in Inner Mongolia represents one of the most ambitious infrastructure investments by a Chinese AI company to date. By combining company-owned facilities with leased capacity in Ulanqab, the startup is positioning itself to secure the massive computing resources required for future AI model development and large-scale deployment. The choice of Inner Mongolia highlights the growing importance of regions offering abundant renewable energy, lower electricity costs, and favorable operating conditions for hyperscale AI infrastructure.
Looking ahead, the project could significantly strengthen DeepSeek’s ability to compete with global AI leaders by reducing reliance on external computing providers while supporting the training of increasingly advanced AI models. As AI workloads continue to expand worldwide, investments in dedicated data centers are expected to become a defining competitive advantage, reshaping both the semiconductor industry and the global AI infrastructure landscape.
Frequently Asked Questions
How large is the AI data center DeepSeek is planning?
DeepSeek is planning to build around one gigawatt (GW) of computing capacity in a data center campus in Inner Mongolia.
Why did DeepSeek choose Inner Mongolia for the data center?
The city of Ulanqab has emerged as a major Chinese data center hub due to its abundant renewable energy resources, cool climate, and relatively low electricity costs.
What is the significance of this project for DeepSeek?
It marks one of China’s largest AI infrastructure projects as DeepSeek ramps up competition with leading U.S. AI companies.
Get the day’s top stories in your inbox
One concise email. No spam, unsubscribe anytime.
