Back
DeepSeek resumes fundraising talks at roughly $74 billion valuation

DeepSeek resumes fundraising talks at roughly $74 billion valuation

The Hangzhou-based AI startup is seeking about $8 billion in a second external round after a July pause, with Monolith Management in active discussions to participate.

Fact Check
The central claim — that DeepSeek restarted a second fundraising round targeting 50 billion yuan — is consistently reported across the odaily, PANews, and BlockBeats sources, all attributing the story to Caijing Magazine (财经杂志), with matching figures (¥50 billion raise, ~¥500 billion pre-money valuation, late-August signing target). Independent earlier reporting from Sina Finance and other financial outlets confirms the earlier late-July pause of the same round, lending credibility to the restart narrative. The claim is sourced to reporting rather than an official DeepSeek statement, and some investment institutions reportedly said they had not been notified of the restart, so a small measure of uncertainty remains. The Taiwan server-supplier framing in the headline is peripheral and not the substance of the assessed fundraising claim.
Summary

DeepSeek has restarted its second external funding round, seeking roughly $8 billion at a valuation approaching 500 billion yuan, or about $74 billion. The Hangzhou-based AI startup paused the process in July 2026 after investor comments by founder Liang Wenfeng were leaked, but discussions have resumed and Hong Kong-based Monolith Management is in active talks to join the round. The company closed its first external fundraising in June 2026, raising more than $7.4 billion at a post-money valuation above $50 billion. That round drew investors including Tencent, JD.com and Monolith Management, while Liang Wenfeng made the single largest contribution with a personal commitment of roughly $3 billion. The new valuation represents an increase of about 48% in only a few months. Founded in 2023 and employing around 200 people, DeepSeek develops open-source large language models and multimodal AI technologies. It drew global attention earlier in 2025 when its models were seen as competitive with Western frontier systems at a fraction of the reported training cost, helping trigger a selloff in U.S. tech stocks and intensifying debate over whether AI development must remain as capital-intensive as many leading companies have assumed.

Terms & Concepts
  • post-money valuation: A company's estimated value immediately after new investment capital has been added.
  • large language models: AI systems trained on vast amounts of text to generate, summarize and analyze language.
  • multimodal AI: Artificial intelligence that can process and generate more than one kind of data, such as text, images or audio.