DeepSeek Raises $15 Billion With Tencent and CATL Backing, Report Says
DeepSeek reportedly raises $15 billion from investors including Tencent and CATL to expand data centers and open models, while preparing for an IPO.
DeepSeek develops artificial intelligence models and services, with work spanning language, reasoning and software-related tasks.
DeepSeek develops artificial intelligence models and services, with work spanning language, reasoning and software-related tasks. The company provides model technology that developers and other users can access through hosted products and, for selected releases, downloadable model weights under the applicable licensing terms.
Its research includes methods for improving model efficiency and the capabilities needed to handle complex instructions. Published technical material and model releases give the wider developer community ways to evaluate and build on parts of that work. Access to model weights is distinct from a hosted service: users running a model themselves remain responsible for computing resources, integration and operational controls.
DeepSeek also offers consumer-facing conversational products and interfaces for developers. Customers can use its models in applications that involve generating text, assisting with code or working through multistep problems. As with other generative AI systems, the usefulness of an output depends on the task, the surrounding application and the review applied before the result is relied upon.
Based in China and founded by Liang Wenfeng, DeepSeek combines model research with product distribution. Its commercial services sit alongside the broader reach of its openly distributed technology. The business is shaped by the cost of computing, continuing research and the ability to turn model capability into dependable tools for users and developers across different technical environments.
DeepSeek API; Model weights under applicable licenses
Commercial subscriptions; usage-based and enterprise plans
DeepSeek reportedly raises $15 billion from investors including Tencent and CATL to expand data centers and open models, while preparing for an IPO.
DeepSeek is nearing at least $12 billion in funding backed by Tencent and CATL, after targeting a valuation of roughly $75 billion ahead of a planned IPO.
Nvidia has struck a $6 billion agreement with AI startup Poolside to develop a powerful open-source AI model capable of competing with leading Chinese and U.S. systems.
A new AI-themed bar in Beijing is giving customers free DeepSeek tokens instead of Wi-Fi, reflecting China's rapidly growing enthusiasm for artificial intelligence.
DeepSeek has reportedly paused its latest fundraising round after comments by founder Liang Wenfeng on the U.S.-China AI rivalry attracted widespread attention.
Tencent and Alibaba are in talks to invest in DeepSeek at a valuation exceeding $20 billion.
The hedge fund behind DeepSeek delivered over 50% returns last year, boosting capital reserves for one of the world’s fastest-rising AI developers.
Experimental Grok 4.20 was the only AI model to earn a profit in a trading tournament, gaining 12.11% as rivals posted heavy losses.