Chinese artificial intelligence startup DeepSeek is nearing the close of its first external fundraising round, a landmark deal that values the company at just under $60 billion. The round is expected to raise more than 50 billion yuan ($7.4 billion; £5.5 billion), according to people familiar with the matter. Only two months ago, DeepSeek was valued at around $10 billion, making the new valuation a more than six-fold jump in a remarkably short span.
The funding round signals a strategic shift for DeepSeek, which had long resisted outside capital. The company was founded by Liang Wenfeng, a billionaire entrepreneur who also leads the highly profitable quantitative hedge fund High-Flyer Quant. With backing from High-Flyer, DeepSeek had sufficient resources to develop its early models. But as competition in the global AI sector intensifies, DeepSeek has decided to open its cap table to a select group of strategic and financial investors.
Investor Lineup and Expected Contributions
The anticipated investor group includes some of the biggest names in Chinese technology and manufacturing. According to reports citing unnamed individuals close to the deal, Tencent Holdings is expected to invest around 10 billion yuan (about $1.48 billion). NetEase and JD.com, both major internet players, are each expected to contribute roughly 3 billion yuan. CATL, the world’s largest electric vehicle battery manufacturer, is expected to put in about 5 billion yuan. The participation of CATL is notable; it underscores how DeepSeek’s AI advancements could intersect with industrial applications, including smart manufacturing and energy management.
A number of venture capital firms are also reported to be joining the round, including IDG Capital, Monolith, Loyal Valley Capital, and Shixiang Tech. These firms are likely drawn by DeepSeek’s rapid growth and its potential to reshape the AI landscape with highly efficient model architectures. In addition, state-backed funds are expected to take part, although the final terms—including which specific vehicles will provide the capital—have not yet been settled.
Liang Wenfeng himself is expected to invest about 20 billion yuan of his own capital into the round. This significant personal commitment demonstrates his confidence in DeepSeek’s future and helps align his interests with those of new external shareholders. It also reduces the need for outside investors to shoulder the entire funding amount, making the round less dilutive for the company.
A Strategic Pivot to External Capital
DeepSeek’s decision to seek external funding marks a notable shift from its earlier stance. In its early days, the startup preferred to rely on internal resources, particularly from High-Flyer Quant, which had accumulated substantial profits during strong years in China’s quantitative trading markets. Liang often emphasized the importance of independent research and long-term thinking, free from the pressure of external shareholders.
However, local media reports suggest that DeepSeek has come to recognize the need for a well-defined market valuation, especially to attract and retain top AI talent. In China’s fiercely competitive tech industry, equity incentives are a vital tool for keeping senior researchers and engineers from being poached by rivals or by global giants offering lucrative compensation packages. A transparent valuation provides employees with confidence in the worth of their stock options. It also allows the company to offer more attractive pay packages without exhausting its cash reserves.
The decision also comes at a time when deep-pocketed competitors, both in China and abroad, are spending enormous sums on talent and infrastructure. By bringing in strategic partners like Tencent, NetEase, JD.com, and CATL, DeepSeek gains not only capital but also potential commercial synergies. Tencent, for instance, has extensive cloud and gaming businesses that could integrate DeepSeek’s AI models. JD.com is one of China’s largest e-commerce and logistics companies, where AI-driven optimization can yield substantial cost savings and customer experience improvements. NetEase, with its focus on gaming and education, could apply DeepSeek’s language models to interactive content. CATL’s involvement might open doors in industrial AI, battery technology simulation, and supply chain forecasting.
DeepSeek’s Disruptive AI Models
DeepSeek burst onto the global AI scene last year with the release of a series of highly capable models that were surprisingly cost-efficient to develop and operate. Its models achieved performance levels comparable to those of leading Western systems while reportedly using fewer computational resources and less training data. This efficiency generated widespread attention because it suggested that cutting-edge AI could be developed without relying on massive clusters of the most advanced—and hardest-to-get—GPUs.
In April, DeepSeek released its V4 model, which represented another leap forward in capability. According to the company, V4 is designed to deliver strong performance across a range of tasks, including natural language understanding, code generation, mathematical reasoning, and multimodal processing. The release reinforced DeepSeek’s reputation as an innovator in model architecture and training methodology.
The company has also been aggressive in lowering prices for access to its models. At the time of V4’s release, DeepSeek said that prices would “drop significantly” in the second half of this year. That prediction was tied closely to expectations that Huawei’s Ascend 950PR supernodes would begin shipping at scale. These supernodes are central to China’s effort to create a domestic high-performance computing ecosystem that can reduce reliance on advanced chips from Nvidia, which are subject to strict US export controls.
DeepSeek’s efficiency-first approach aligns well with China’s broader push to develop a self-sufficient AI supply chain. The company is already working with Chinese chipmakers to ensure its models run optimally on domestic silicon. Huawei, Cambricon, and MetaX have all announced that they are optimizing their AI processing chips for DeepSeek’s V4 architecture. This collaboration is not merely technical; it is part of a national strategy to ensure that Chinese AI companies can continue to innovate even as access to foreign advanced semiconductors becomes more constrained.
Implications for the AI Market
The new funding round will likely have ripple effects across the Chinese AI sector. DeepSeek’s valuation jump from $10 billion to nearly $60 billion in just two months is extraordinary, even in an industry known for rapid value creation. It sends a powerful signal to both domestic and international investors that China’s AI ecosystem still holds enormous potential despite geopolitical headwinds.
The round also illustrates how strategic investors in China are moving beyond the traditional consumer internet businesses and betting heavily on foundational AI technologies. Tencent and JD.com, together with battery giant CATL, represent a cross-section of China’s economic pillars: social media and gaming, e-commerce and logistics, and advanced manufacturing. Their participation suggests that AI is no longer seen as a tool that merely improves advertising systems or product recommendations; it is becoming the core infrastructure for the next generation of industrial transformation.
For DeepSeek, external capital is not an end in itself. The funds will be used to scale computing infrastructure, expand research teams, commercialize its models across industries, and possibly explore advanced robotics or embodied AI. The company has also indicated that lower prices for its models could accelerate adoption among small and medium-sized businesses in China, which have historically been underserved by expensive AI services from U.S. providers.
Yet there are also questions about whether valuations in the AI sector have reached levels that cannot be sustained. DeepSeek’s rapid rise is based on real technological achievements, but it will need to generate substantial revenues to justify a $60 billion price tag. The company has yet to disclose detailed revenue figures. Its business model so far has centered on open-source or low-cost AI models, which raises the question of how it plans to monetize its technology beyond providing API access.
Liang Wenfeng has given no public indication that he plans to take DeepSeek public in the near term. In the past, he has emphasized the importance of basic research and open collaboration. His decision to allow external investors to take a minority stake is a pragmatic concession to the competitive realities of the AI talent market. It does not mean that he has abandoned his vision of a research-driven company.
China’s AI chip sector is also closely watching DeepSeek’s trajectory. Huawei’s Ascend series has become the flagship platform for domestic AI models, and large-scale shipments of the 950PR supernodes could give Chinese developers a reliable alternative to Nvidia’s GPUs. Cambricon and MetaX are also vying for a share of the fast-growing demand from AI startups and traditional enterprises alike. DeepSeek’s willingness to optimize its models for these chips could accelerate the entire domestic supply chain, creating virtuous cycles of improvement in hardware and software.
The recent hikes in DeepSeek’s valuation also come as U.S. regulators continue to restrict the export of high-end AI chips and related technology to China. Many analysts argue that these export controls have inadvertently pushed China to become more innovative in hardware efficiency and model optimization. DeepSeek is often cited as a prime example: by using fewer chips and smarter algorithms, the company has achieved results that would have been difficult for teams relying solely on brute-force compute.
International observers have taken note. The fact that DeepSeek’s models have been reproduced or imitated by researchers in other countries highlights their technical significance. Some global companies have begun to integrate DeepSeek’s open-source models into their own products, despite geopolitical sensitivities. This international appetite could give DeepSeek an additional revenue stream through licensing or usage fees
As the funding round moves toward completion, the exact allocation of shares and the valuation cap remain to be finalized. State-backed participation is still under discussion, and nothing will be official until the contracts are signed. But the broad shape of the deal is clear: DeepSeek is about to become one of the most highly valued AI companies in China, with a cap table that includes sovereign-influenced funds and the largest names in tech and manufacturing.
This is a testament to how far the startup has come since its early days as an offshoot of High-Flyer Quant. It also underscores the global significance of the Chinese AI industry, which continues to produce groundbreaking models and attract massive amounts of capital. Whether DeepSeek can maintain its momentum remains to be seen, but its first external funding round is already shaping up to be one of the defining financial events of the year in the technology sector.
The next few months will be critical, as DeepSeek’s price cuts and the broader rollout of Huawei’s Ascend 950PR supernodes are expected to reshape the competitive dynamics of AI deployment in China. If chipmakers and model creators can deliver on their promises, the result could be a new wave of affordable, high-performance AI applications across manufacturing, finance, healthcare, and consumer services. DeepSeek, with its newfound financial firepower and strategic partnerships, is well positioned to lead that wave.
Source: Silicon UK News