CATL intends to participate in DeepSeek’s ongoing initial funding round, The Information reported on Friday, citing two people familiar with the matter, News.Az reports, citing Cnevpost.
***
The funding round targets raising about 50 billion yuan ($7.36 billion) and could close as early as June, according to the report.
Once this massive capital increase is successfully completed, DeepSeek’s overall valuation could exceed 350 billion yuan, the report said.
The potential investment by CATL comes as the company expands into new businesses. The battery maker is seeking to sell power supply equipment to massive AI data centers to meet the energy demands brought on by the surge in computing power.
Chinese tech giants JD.com and NetEase are also currently in talks to acquire stakes. However, these investment discussions have not been finalized, and details such as the investment amount and the final lineup of participants remain subject to change, according to the report.
DeepSeek rose to prominence after releasing its powerful open-source large language model R1 in January 2025, which was trained at a fraction of the computing cost of its US rivals.
In its latest V4 model release, the company said that the model has been optimized for inference on Huawei’s Ascend 950PR chips. This trend of coordinated software and hardware development is drawing close attention from the global tech community.
On May 6, the Financial Times reported that China’s largest state-backed semiconductor investment fund was in talks to lead DeepSeek’s initial funding round.
The potential deal could value the frontier AI lab at about $45 billion, according to the report.
Other investors still in talks to acquire stakes include Chinese tech giants Tencent and Alibaba, the report said.
On May 8, The Information said in a separate report that DeepSeek was seeking to complete its first funding round, targeting up to 50 billion yuan.
If this massive financing plan ultimately materializes, it will become the largest funding round for a Chinese AI company to date.
DeepSeek founder and CEO Liang Wenfeng plans to provide the largest single chunk of funding for this round, according to the May 8 report.
At the same time, the massive financing plan has prompted the Chinese AI lab to accelerate its overall pace of generating revenue and achieving commercialization.
DeepSeek has informed some investors of its plan to increase the release frequency of its AI models, The Information reported, citing two people involved in the discussions.
To support its growing computing power needs, the startup is currently actively expanding into the asset-heavy physical infrastructure sector.
DeepSeek plans to build a large-scale data center in Ulanqab, Inner Mongolia, in northern China, according to job postings from last month.
Ulanqab is considered one of the most suitable cities in the country for establishing data centers. The local abundance of electricity and extremely low average annual temperatures can significantly reduce server cooling costs, thereby helping to improve the power usage effectiveness of the new facility.
Building its own data center will help DeepSeek gain more initiative in future model iterations. Late last month, the company suffered a system crash lasting nearly 12 hours, further highlighting the importance of owning large-scale physical server facilities.
This asset-heavy strategic move comes as DeepSeek prepares to release its next-generation model update. The company plans to launch the V4.1 version in June to bring its release cadence closer to current industry standards.
