DeepSeek is closing in on one of the largest private funding rounds in Chinese technology history, with the artificial intelligence startup set to raise at least $12 billion from investors led by Tencent and battery maker CATL, Bloomberg reported Monday. People familiar with the matter said the total could approach $15 billion as investor demand continues to climb.
The figure marks a dramatic escalation from DeepSeek's original target. The company had aimed to raise roughly $7.5 billion at a valuation near $75 billion before strong interest from investors pushed the round higher, according to The Decoder. CATL and Tencent are contributing the largest shares of the new round. CATL declined to comment, while Tencent and DeepSeek did not respond to requests for comment. For more context on this story, see our ongoing AI trends.
Investor demand follows V4-Flash
The surge in interest is being driven largely by DeepSeek's V4-Flash model, which has set new benchmarks for cost and performance against rivals from Anthropic and OpenAI, The Decoder reported. The company released V4-Pro and V4-Flash as preview versions in April, and the efficiency of the smaller Flash variant has become a talking point among developers weighing inference costs.
For DeepSeek, the round would more than double the capital it gathered just a few months ago. In June, the company closed its first external funding round at about $7.4 billion with a valuation above $50 billion, a striking result for a startup that had long been funded primarily by its founder's quant fund.
Restructuring for an early 2027 IPO
Once the round closes, DeepSeek plans to restructure ahead of an initial public offering targeted for early 2027, according to The Decoder. A listing would make the Hangzhou-based lab one of the first frontier AI companies to go public, giving retail investors direct exposure to a firm that has repeatedly outperformed far larger Western rivals on cost.
Reuters separately reported Monday that DeepSeek is set to net over $12 billion in the new fundraising, citing a source, which corroborates the Bloomberg figures. A report by Calcalist suggested the round could approach $15 billion given the level of demand.
Huawei chips and cutting Nvidia dependence
The fundraising coincides with an infrastructure push designed to reduce DeepSeek's reliance on US hardware. The company is building a data center equipped with at least 160,000 AI chips from Huawei, The Decoder reported, and is also working on its own inference chip to lessen its dependence on both Nvidia and Huawei silicon.
The Nvidia relationship has been a delicate subject. DeepSeek paused the funding round earlier this year after comments by founder Liang Wenfeng about the company's reliance on Nvidia chips went viral, before reviving it with an expanded target.
Open weights remain the strategy
Despite the influx of capital and the IPO plans, Liang intends to keep DeepSeek developing open models with freely available weights, according to The Decoder. That commitment distinguishes the company from Western labs that have moved toward closed, subscription-based frontier models, and it has helped DeepSeek build goodwill among developers worldwide.
The open-weight approach has also turned DeepSeek into a benchmark for the broader Chinese AI ecosystem, with rivals racing to match its training efficiency and pricing. A successful $12 billion raise would give the company substantial runway to continue that strategy while funding the compute build-out needed to stay competitive at the frontier.
What it means for the global AI race
A round of this size would place DeepSeek among the most valuable AI startups in the world and intensify the competition between Chinese and American labs for both capital and talent. It also signals that Chinese investors, including industrial giants like CATL, are willing to write frontier-scale checks despite geopolitical tensions over AI chips and export controls.
The IPO plans also reflect how quickly the center of gravity in Chinese AI is shifting toward public markets. Rival lab Moonshot AI is separately eyeing a Hong Kong IPO after reaching a $50 billion valuation, according to Yahoo Finance UK, suggesting a wave of Chinese AI listings could be forming for 2027.
For enterprises choosing AI providers, DeepSeek's growing war chest and infrastructure investment suggest its aggressively priced models will remain a force in the market, with the added possibility of a public listing that could reshape how AI companies are valued. A company that trains and serves frontier-class models on domestic Huawei silicon while undercutting Western providers on price would alter procurement math for cost-sensitive buyers far beyond China.
The open questions are scale and sustainability. V4-Flash's cost advantages have been demonstrated in benchmarks, but a 160,000-chip data center represents an enormous capital commitment with its own power and cooling demands. Whether DeepSeek can convert its new $12 billion — or perhaps $15 billion — into a durable technological lead, rather than a longer race, will define the next phase of the US-China AI competition.
---
Stay Ahead of AIGet the latest AI news, analysis, and breakthroughs — all in one place.
Read more AI news →
