DeepSeek, the Chinese artificial intelligence startup that has captured global attention with its advanced language models, is on the verge of securing its first-ever round of external funding at a valuation of $45 billion (approximately £33 billion). According to a report by the Financial Times citing unnamed sources familiar with the matter, the China Integrated Circuit Industry Investment Fund, commonly known as the "Big Fund" — the country's largest state-backed semiconductor investment vehicle — is in advanced talks to lead the round. This development underscores the strategic importance Beijing places on domestic AI capabilities and the broader push for technological self-sufficiency.
The valuation marks a dramatic leap from DeepSeek's initial expectations. When talks began just a few weeks ago, the startup sought a valuation of at least $10 billion. By late April, that figure had already climbed to $20 billion. The latest $45 billion figure reflects intense investor confidence in DeepSeek's potential, even as the company focuses primarily on developing frontier AI models rather than broad commercialisation — a strategy that distinguishes it from rivals such as Moonshot and Zhipu, the latter of which is valued at around $52 billion and listed in Hong Kong.
Key Players in the Funding Round
Beyond the Big Fund, other major Chinese tech conglomerates are reportedly circling the deal. Tencent Holdings and Alibaba Group — two of China's largest internet companies — are also said to be in discussions to participate in the funding round. The involvement of these industry giants highlights the competitive landscape of China's AI sector, where companies are racing to secure stakes in promising startups that could shape the future of artificial intelligence. For Tencent and Alibaba, investing in DeepSeek aligns with their broader strategies to embed advanced AI into their ecosystems, from cloud computing to consumer applications.
The Big Fund's participation is particularly noteworthy. Established in 2014, the China Integrated Circuit Industry Investment Fund has been a cornerstone of the government's effort to reduce reliance on foreign semiconductor technology. The fund has backed significant domestic players such as SMIC (Semiconductor Manufacturing International Corporation) and YMTC (Yangtze Memory Technologies Co.), both of which are critical to China's chipmaking ambitions. By investing in DeepSeek, the Big Fund is extending its reach into the software and AI modelling layer, recognizing that hardware and software must evolve together to compete with US-based giants like OpenAI and Google.
State Support and National Strategy
The investment comes at a time when the Chinese government is urging domestic tech companies to collaborate and create an integrated ecosystem around AI. This push is part of a broader national strategy to counter US dominance in technology, particularly in areas like semiconductors, cloud infrastructure, and advanced algorithms. Official statements from Beijing have repeatedly emphasized the need for "self-reliance" in critical technologies, and DeepSeek's rapid rise fits neatly into that narrative.
Earlier this year, Chinese Premier Li Qiang called for accelerated development of the digital economy, with a specific focus on AI and its applications across industries. The government has also rolled out various subsidies and incentives for companies working on AI chips, training models, and data centers. DeepSeek, with its groundbreaking models like the recently launched V4, is seen as a flagship example of what China can achieve in the AI race.
DeepSeek V4 and Huawei Optimization
At the launch of its latest V4 model, DeepSeek announced that the software was optimized to run inference on Huawei's Ascend 950PR chips. This is a significant milestone for China's tech independence efforts. Huawei, once heavily reliant on Western suppliers, has emerged as a leading domestic chip designer despite facing severe US export controls. The Ascend 950PR is designed specifically for AI workloads, offering high-performance computing capabilities that rival Nvidia's products in certain tasks.
The optimization means that Chinese companies and government agencies can use DeepSeek's advanced models without depending on Nvidia's hardware, which remains largely unavailable in China due to US export restrictions. These restrictions, imposed by the Biden administration and continued under subsequent policies, have blocked the sale of advanced Nvidia chips like the A100, H100, and the latest Blackwell series to Chinese entities. In response, Nvidia has developed downgraded versions for the Chinese market, but these chips still face regulatory hurdles from both US and Chinese authorities.
As a result, Huawei has overtaken Nvidia in market share within China for AI accelerators. Official figures from Chinese market research firms indicate that Huawei's Ascend series now accounts for over 40% of domestic AI chip sales, up from less than 10% two years ago. This shift is reshaping the entire AI supply chain in China, with software developers increasingly building for Huawei's architecture. DeepSeek's decision to optimize its V4 model for Ascend chips is a clear endorsement of this trend.
Background on DeepSeek
DeepSeek was founded in 2021 by a group of AI researchers who had previously worked at major Chinese internet companies and academic institutions. The company quickly gained recognition for its large language models, which have demonstrated performance comparable to OpenAI's GPT series on several benchmarks. DeepSeek's models are used in a variety of applications, including natural language processing, code generation, and scientific research. The company has maintained a relatively low profile compared to startups like Moonshot, which have aggressively pursued consumer-facing products.
Despite its focus on frontier research, DeepSeek has attracted attention from venture capitalists and strategic investors. The current funding round would be the first time the company has taken external capital, having previously been funded by its founders and small private investments. The decision to open up to outside investors is seen as a sign that DeepSeek is preparing for a large-scale expansion, potentially including plans for commercialization and global outreach.
Comparison with Peers
DeepSeek's valuation of $45 billion places it among the top-tier AI startups globally, though still behind leaders like OpenAI (valued at over $80 billion) and Anthropic. In China, it competes with Zhipu (valued at $52 billion), Moonshot, and Baidu's AI division. Zhipu, which is listed in Hong Kong, has a broader product portfolio but weaker technical benchmarks. Moonshot has focused on conversational AI and has secured funding from Alibaba and other investors. DeepSeek, by contrast, has carved out a niche in foundational model research, often releasing papers and open-source components that earn credibility in the academic community.
Investors are betting that DeepSeek's technology will translate into commercial success over time, especially as the Chinese government mandates the use of domestic AI solutions in critical sectors like finance, healthcare, and infrastructure. The $45 billion valuation implies that DeepSeek's current revenue, while likely small, is expected to grow exponentially as the market matures.
Challenges Ahead
Nevertheless, DeepSeek faces several challenges. The US export controls on advanced chips remain a bottleneck, even with Huawei's alternatives. While Huawei's Ascend chips are competitive, they are not yet on par with Nvidia's top-tier products in terms of raw performance for training large models. DeepSeek will need to continue investing in software optimization to make the most of available hardware. Additionally, the company must navigate a complex regulatory environment in China, where data privacy laws and AI governance frameworks are still evolving.
Another challenge is the increasing competition from both domestic and international players. OpenAI's models are still the gold standard, and US companies continue to innovate rapidly. Chinese rivals like Baidu's ERNIE Bot and Alibaba's Tongyi Qianwen have also made strides. DeepSeek's ability to maintain its technical edge will be crucial for justifying its lofty valuation.
The involvement of the Big Fund adds a layer of political consideration. State-backed investments often come with expectations about strategic alignment with national priorities, which could affect DeepSeek's operational independence. However, given the government's desire to see a homegrown AI champion, such support is likely to be more beneficial than burdensome in the short term.
As the funding round progresses, all eyes will be on the final terms and the investor roster. If completed at $45 billion, DeepSeek will become one of the most valuable private AI companies globally, and a symbol of China's ambitions in the technology race against the United States.
Source: Silicon UK News