DeepSeek Chases $20B+ Valuation as Tencent, Alibaba Talk

Chinese AI startup DeepSeek is seeking a valuation above $20 billion as Tencent and Alibaba hold talks about possible investments.

DeepSeek is seeking more than $20 billion in valuation as talks with Tencent Holdings and Alibaba Group progress, people familiar with the discussions said. The company is negotiating with potential investors and has increased its asking price as interest from major tech groups has grown.

The startup is owned by hedge fund High-Flyer Capital Management and only recently began pitching to outside investors. Initial targets for the round were roughly $300 million at a valuation near $10 billion; those figures have risen as Tencent and Alibaba explored commitments. Negotiations remain ongoing and the amount DeepSeek ultimately raises, and the final valuation, could change.

Some U.S. venture firms are approaching DeepSeek cautiously because it is a Chinese company working on large AI models. Earlier this year DeepSeek did not allow certain U.S. chipmakers to tune its flagship model for performance. Company insiders also say a newer model was trained on an Nvidia processor that is restricted from export, complicating technical collaboration with some overseas partners.

Nvidia CEO Jensen Huang warned that if major AI models are optimized to run primarily on non-U.S. hardware, the global balance of technology could shift. He described such an outcome as “a horrible outcome” for the United States if models favor another hardware stack over American chips.

On hardware performance, the Huawei Ascend 910C accelerator delivers roughly 60% of the inference performance of Nvidia’s H100, according to benchmark estimates. The H100 is two generations behind Nvidia’s current top chip. Industry benchmarks indicate U.S. accelerators are about five times more powerful than Chinese rivals today, with some estimates projecting a gap widening to about 17 times by 2027. Huawei has set a target of shipping 750,000 AI chips in 2026, while its current production capacity represents roughly 3% to 5% of Nvidia’s combined computing power.

DeepSeek’s first public model release in January 2025 prompted broad market attention and pushed domestic competitors in China to accelerate upgrades to their own models. Company engineers are working on a new model internally referred to as V4; if V4 runs efficiently on Huawei Ascend chips, it would provide an additional hardware route for deploying large models.

The fundraising around DeepSeek comes as investors continue to back infrastructure providers. Storage and data platform Vast Data announced a $1 billion financing at about a $30 billion valuation, with Nvidia among the backers. Drive Capital and Access Industries led that round, which included Fidelity and NEA and combined primary and secondary capital. Industry tallies show AI companies have raised roughly $280.5 billion this year, with more than $170 billion going to a few of the largest developers.

Drive Capital partner Chris Olsen called Vast’s architecture “capable of supporting the world’s most demanding AI environments,” reflecting investor interest in companies that supply compute and storage for large model training and inference.

For now, DeepSeek’s funding talks are fluid. Potential backers are weighing the startup’s technical progress and market position against export controls, chip supply limits and geopolitical concerns. Any final deal terms could shift as those factors evolve.

Content on BlockPort is provided for informational purposes only and does not constitute financial guidance.
We strive to ensure the accuracy and relevance of the information we share, but we do not guarantee that all content is complete, error-free, or up to date. BlockPort disclaims any liability for losses, mistakes, or actions taken based on the material found on this site.
Always conduct your own research before making financial decisions and consider consulting with a licensed advisor.
For further details, please review our Terms of Use, Privacy Policy, and Disclaimer.

Articles by this author

This site is registered on wpml.org as a development site. Switch to a production site key to remove this banner.