DeepSeek's first-ever external funding round could value the Chinese artificial intelligence startup at approximately $45 billion, according to people familiar with the matter. The talks are being led by the China Integrated Circuit Industry Investment Fund, widely known as the "Big Fund," which is the country's largest state-backed semiconductor investment vehicle. The figure marks a dramatic escalation from earlier discussions, when DeepSeek initially sought a valuation of at least $10 billion just a few weeks ago.
State Funding and Strategic Backing
The Big Fund has historically focused on significant semiconductor enterprises, including contract chipmaker SMIC and memory manufacturer YMTC. Its reported interest in DeepSeek signals a broader strategic push by Beijing to strengthen China's AI ecosystem. By late April, reports indicated that the valuation had already risen to $20 billion, and it has continued to climb as more investors express interest.
Earlier reports also suggested that Tencent Holdings and Alibaba Group are in talks to take part in the round. The involvement of these major Chinese technology companies would provide DeepSeek with not only capital but also potential distribution channels, cloud infrastructure, and enterprise partnerships. For Tencent and Alibaba, investing in DeepSeek offers a way to align with a rapidly advancing AI model developer without having to build all frontier technology in-house.
DeepSeek's Positioning in the Chinese AI Race
DeepSeek is widely recognized for its focus on developing frontier models rather than pursuing broad commercialisation at an early stage. This distinguishes it from Chinese rivals such as Moonshot and the Hong Kong-listed Zhipu, which is currently valued at around $52 billion. Zhipu has moved quickly to deploy its models across consumer and enterprise applications, while Moonshot has made significant strides in long-context reasoning and conversational AI. DeepSeek, by contrast, has built a reputation for fundamental research, model efficiency, and a willingness to push architectural boundaries.
The company's rise has captured global attention, particularly because it demonstrated that advanced AI models could be trained with comparatively modest computing budgets. Earlier DeepSeek releases challenged prevailing assumptions about the scale of investment required to compete with leading American labs. That efficiency narrative has made DeepSeek especially attractive to investors who see a path to sustainable differentiation in a crowded field.
Government Push for an Integrated AI Ecosystem
The reported investment comes as the Chinese government is urging domestic technology players to collaborate more closely on creating a unified AI ecosystem that can compete with American products. Chinese regulators and state-linked entities have increasingly encouraged cooperation among chip designers, cloud providers, model developers, and application builders. The goal is to reduce dependence on foreign technology and build a resilient domestic supply chain.
DeepSeek's potential tie-up with the Big Fund would fit directly into that strategy. The Big Fund's mandate is to accelerate the development of China's semiconductor industry, and AI accelerators are a critical component of that mission. By investing in an AI model developer, the fund could help create a closer loop between chip production, model training, and real-world deployment. This is especially important as the United States has imposed export controls on advanced semiconductors, making it harder for Chinese firms to access the most powerful Nvidia GPUs.
Huawei Optimisation and the Ascend 950PR
DeepSeek said at the launch of its latest V4 model that the software was optimised to run inference on Huawei's Ascend 950PR chips. This announcement was widely seen as a significant boost for China's technology independence efforts. Huawei has developed its own family of Ascend AI processors to challenge Nvidia's dominance in the Chinese market. The Ascend 950PR is among the most capable chips Huawei has produced, and DeepSeek's optimisation work demonstrates that domestic models can run efficiently on domestic hardware.
The partnership with Huawei is strategically important for both companies. Huawei gains a high-profile AI model developer that can showcase the performance of its chips. DeepSeek gains access to a plentiful and legally available source of computing power at a time when Nvidia's advanced products remain largely unavailable in China due to regulatory restrictions from both the United States and China. Nvidia has been pushed to develop lower-performance versions of its chips specifically for the Chinese market, but these have often failed to fully meet the needs of frontier AI research.
Huawei's Market Share Gains
Huawei has overtaken Nvidia in market share in China, according to industry analysts and supply chain reports. The shift is partly due to U.S. export controls that prevent Nvidia from selling its most advanced GPUs to Chinese customers. Chinese companies have consequently turned to domestic alternatives, and Huawei has been the biggest beneficiary. The Ascend series has become the preferred choice for many Chinese AI labs and cloud providers seeking reliable long-term supply.
DeepSeek's optimisation for Ascend 950PR could accelerate this trend. If a frontier lab like DeepSeek can deliver strong inference performance on Huawei hardware, it may convince other model developers and enterprises to adopt domestic chips. That would help China build a more self-reliant AI infrastructure and reduce the strategic vulnerability created by dependence on imported chips.
Implications for the Global AI Landscape
The reported valuation of $45 billion underscores how quickly the global AI landscape is shifting. DeepSeek has emerged as a serious contender in the race to build advanced reasoning models, despite being founded only in 2023. Its research team has published papers that have influenced both academic and industrial AI communities. The company has also demonstrated a willingness to experiment with model architectures that differ from the dominant transformer paradigm, which could yield long-term advantages in efficiency and capability.
Investors are reportedly betting on DeepSeek's potential even though it has not yet pursued aggressive commercialisation. The company's focus on frontier models suggests that it may be aiming to establish technical leadership before turning to revenue generation. If DeepSeek can maintain its research momentum, the valuation could be justified by future product applications in areas such as enterprise software, coding assistants, scientific discovery, and autonomous systems.
Comparison With Other Chinese AI Startups
Moonshot AI, another prominent Chinese startup, has concentrated on large-scale language models with advanced reasoning abilities. It has attracted its own substantial funding and has built a loyal user base. Zhipu, which is listed in Hong Kong, has pursued a more diversified strategy, offering a range of models for different industries and integrating AI into business workflows. Both companies have higher valuations than DeepSeek's current $45bn figure, but the gap is narrowing.
The Chinese AI market is characterised by intense competition and rapid technological change. Startups must constantly release new models, improve efficiency, and demonstrate practical value. DeepSeek's decision to prioritise research depth over speed to market is a calculated risk. It could allow the company to leapfrog competitors with more advanced thinking capabilities, but it also leaves room for rivals to capture market share in the meantime.
The Big Fund's Evolving Role
The China Integrated Circuit Industry Investment Fund has been a cornerstone of Beijing's efforts to build a world-class semiconductor industry. The fund has invested heavily in manufacturing capacity, packaging, chip design, and equipment. Its move into AI model funding would mark a notable expansion of its scope. This is not entirely surprising, however, because AI hardware and AI software are becoming increasingly intertwined. A model developer that can influence chip architecture and software frameworks can help shape the entire ecosystem.
If the Big Fund leads the DeepSeek round, it would send a strong signal that the Chinese government views frontier AI as a strategic industry worthy of direct state support. That could invite further regulatory scrutiny from the United States and other Western governments, which are already concerned about China's progress in both AI and semiconductors. It might also trigger more investment into other Chinese AI labs, as private investors seek to match the government's confidence.
What the Future Holds
DeepSeek's valuation growth from $10 billion to $45 billion in a matter of weeks reflects the extraordinary demand for AI assets and the perceived importance of Chinese models in the global market. While the final valuation has not been confirmed, the trajectory suggests that investors believe DeepSeek is on the cusp of major breakthroughs. The company's ability to attract funding from both state-backed funds and major private technology firms would give it substantial resources to expand its research team, acquire computing hardware, and pursue ambitious projects.
The success of DeepSeek's V4 model and its optimisation for Huawei's Ascend 950PR is a significant milestone for China's technology independence. As the country continues to face export controls and other restrictions from the United States, domestic collaboration between AI developers and chipmakers will be essential. DeepSeek and Huawei are showing that such collaboration can produce competitive results, and this may reshape the global balance of power in artificial intelligence over the coming years.
Source: Silicon UK News