China's secret AI network: An inconspicuous deal reveals China's 2 trillion yuan (approx. 256 billion euros) master plan for AI
Xpert Pre-Release
Available in 27 languages 📢
Prefer Xpert.Digital on GoogleⓘPublished on: September 9, 2026 / Updated on: September 9, 2026 – Author: Konrad Wolfenstein

China's secret AI network: An inconspicuous deal reveals China's 2 trillion yuan (approx. 256 billion euros) master plan for AI – Creative image on the topic, featuring AI: Xpert.Digital
The new gold rush: Why even niche providers in China are suddenly buying data centers
Race for computing power: China's radical transformation of digital infrastructure
Electricity instead of real estate: What's behind China's gigantic data center fever
A small investment vehicle with a multi-million dollar budget is becoming a symbol of what is arguably the biggest industrial policy transformation of this decade. While the world's attention is primarily focused on tech giants like Huawei, Alibaba, and Tencent, the unassuming entry of niche provider Sanrenxing into a Beijing data center demonstrates just how deeply the state-mandated AI gold rush has already penetrated the Chinese economy. Behind the scenes, Beijing is operating with a gigantic master plan worth over two trillion yuan, designed to completely restructure the country's digital infrastructure. The goal: to catch up with the US in the global race for dominance in artificial intelligence. But the deal reveals something else as well – the true bottleneck of the digital revolution is no longer capital, but access to power grid connections and state-of-the-art memory chips. This analysis explores why there is currently no way around server rooms in China and why even players from outside the industry are suddenly betting everything on this one venture.
When a niche provider suddenly becomes an infrastructure investor: Who is really behind the Fangshan deal?
The announcement initially reads as unremarkable: A company listed on the Shanghai Stock Exchange with the ticker symbol 605168, known as Sanrenxing, is establishing a new investment company together with its wholly-owned subsidiary Shidai Zhisuan and an external partner called Huixinfu. The name of this special purpose vehicle, Beijing Zhonglian Yungang No. 1 Enterprise Management Partnership, sounds like one of the countless management vehicles that appear and disappear daily in the Chinese capital market. The committed capital of 2.33 million yuan, equivalent to a low single-digit million euro figure, seems minuscule at first glance compared to the sums typically circulating in the global data center business. But it is precisely in this apparent insignificance that the true significance of this transaction lies. It serves as a microcosm for one of the largest industrial policy upheavals currently underway in China: the national race for computing power in the age of artificial intelligence.
Sanrenxing is acquiring a 60 percent stake in this new partnership for 1.4 million yuan, thereby securing entrepreneurial control of the vehicle, while Huixinfu, the investment arm of the United Data Center Group, is joining as a specialized industry partner. This constellation is typical of numerous Chinese companies from outside the sector that possess strong balance sheets and access to capital and are now rapidly acquiring a foothold in the high-margin digital infrastructure business. The new company will function as a specialized acquisition and operating platform, sourcing, acquiring, and operationally developing high-quality data center facilities nationwide. The first concrete target is a data center project in Beijing's Fangshan district, where the partnership aims for a complete takeover of the project company.
Why Beijing is becoming the finish line for computing power right now
The geographical choice of Fangshan is no coincidence, but follows a logic deeply rooted in China's power grid architecture and regulatory practices. As a political and economic power center, Beijing traditionally attracts a particularly large number of state-affiliated users of cloud and AI services – from ministries and state-owned enterprises to the major technology companies that prefer to locate their core infrastructure near the capital and its decision-makers. At the same time, as part of its "East Data, West Computing" program, which aims to shift computing loads from the congested coastal regions to the more energy-efficient western provinces, the Chinese central government has significantly tightened the permitting process for new large-scale facilities in the eastern metropolises. Anyone who still manages to gain access to an existing or advanced-plan project in the Beijing area effectively secures a scarce resource that new market entrants can hardly access legally.
This scarcity also explains why companies outside the traditional telecommunications and cloud computing sectors are increasingly investing in such projects. It's not just about building new server halls, but primarily about acquiring already secured site and network access rights, which possess considerable intrinsic value in the current regulatory environment. Sanrenxing's Fangshan deal should therefore be seen less as a classic real estate investment and more as the acquisition of a regulatory-secured access ticket to one of China's most important digital hubs.
The elephant in the room: China's two-trillion-yuan bet on computing power
The Sanrenxing deal only reveals its full significance when viewed within the context of national industrial policy. According to reports circulating in the summer of 2026, China's National Development and Reform Commission and other key agencies are preparing a plan to invest approximately two trillion yuan (roughly US$295 billion) over the next five years in a nationwide network of interconnected data centers. The goal is to consolidate disparate digital infrastructure into a unified national computing network by 2028, thereby closing the technological gap with the United States in the field of artificial intelligence. Including the necessary investments in power grid integration, the total investment could reach at least five trillion yuan.
What's remarkable about this plan is the proposed division of roles. State-owned enterprises like China Mobile and China Telecom are slated to operate the vast majority of the new data centers and provide their network connectivity, while the underlying technology is targeted to be at least 80 percent localized – primarily through domestic chip manufacturers like Huawei. Beijing is thus explicitly positioning computing power as strategic infrastructure on par with energy and transportation networks, financed primarily through long-term government bonds and a national fund for strategic industrial investments, supplemented by bank loans and private capital. The Sanrenxing partnership falls precisely into this third category, the supplementary private capital: a small but symptomatic component of a gigantic, state-orchestrated investment program.
🎯🎯🎯 Data-driven B2B industry hub as a quasi-in-house solution

The quasi-in-house solution: How Xpert.Digital closes operational gaps in B2B marketing and sales – Smart Content-Driven Business - Image: Xpert.Digital
Xpert.Digital is a data-driven B2B industry hub led by Konrad Wolfenstein . The company acts as an external, quasi-in-house solution for industrial partners, closing operational gaps in marketing, content, and sales – without requiring additional resources on the client side.
More information here:
The real bottleneck of AI: electricity and chips instead of capital
Numbers that make the race tangible
The urgency of this national effort is underscored by recent supply and demand data from the Chinese data center industry. According to the China Academy of Information and Communications Technology (CAICT), domestic demand for AI computing power surged by 417 percent in the first quarter of this year compared to the same period last year, while supply increased by only 128 percent. This massive supply gap explains the bidding war currently underway across the country: In the first seven months of 2026 alone, independent surveys indicate that more than 158 large-scale data center and intelligent data center projects, each with a contract value exceeding 100 million yuan, were awarded. This equates to a total volume of over 113 billion yuan, representing an increase of more than 18 percent compared to the same period last year.
The following overview summarizes key market data that illustrates the framework within which sanrenxing engagement operates:
| Key figure | Value | Time period/Source |
|---|---|---|
| Planned government investment volume in national data center infrastructure | approximately 2 trillion yuan (approximately 295 billion US dollars) | 2026 to 2031 |
| Installed data center capacity in China | 32 gigawatts (end of 2025), projected to rise to over 60 gigawatts by 2030 | Rystad Energy |
| Power consumption of Chinese data centers | Increase from 170 to 800 terawatt hours by 2030 is projected | National Energy Authority |
| Growth in AI computing power demand vs. supply | +417 percent demand compared to +128 percent supply | Q1 2026, CAICT |
| Market size China Data Center Market | approximately US$33.3 billion (2026), projected to rise to US$63.8 billion by 2031 | Mordor Intelligence |
| Volume of major projects (over 100 million yuan) | Over 158 projects, total value over 113 billion yuan | Jan.–Jul. 2026 |
These figures reveal a market environment in which even small-scale transactions like Sanrenxing's are part of a much larger structural trend, deliberately fueled by the government to close a physical supply gap that would otherwise threaten to derail the entire national AI strategy.
The real resource shortage is electricity, not capital
Those who view China's data center boom solely as a construction and capital market phenomenon overlook the real Achilles' heel of this growth trajectory: energy supply. Analysts at Rystad Energy expect China's installed data center capacity to more than double from 32 gigawatts at the end of 2025 to over 60 gigawatts in 2030. At the same time, the National Energy Administration forecasts that electricity consumption by data centers will increase at an annual growth rate of 36 percent, from 170 terawatt-hours in 2025 to 800 terawatt-hours in 2030. This would then represent six percent of total national electricity consumption (compared to 1.6 percent today). Other estimates put the figure at around 479 terawatt-hours by 2030 – a value roughly equivalent to France's total current electricity demand.
This dimension explains why Beijing cannot leave the expansion of its data center infrastructure solely to the free market, but rather links it closely to power grid planning. It is therefore no surprise that a significant portion of the aforementioned five trillion yuan total investment is allocated to the necessary grid integration and not to the data centers themselves. For investors like Sanrenxing, this means that acquiring an already grid-connected site, such as the one in Fangshan, represents a considerable strategic competitive advantage over new greenfield projects, because it bypasses the time-consuming and politically sensitive approval processes for new grid connections.
From concrete to chip: The shift in value creation in the data center ecosystem
In the current Chinese market environment, the sheer size and building envelope of a data center are becoming less important compared to the actual computing technology and its cooling. According to industry analyses, China's market for data center infrastructure and AI computing reached a volume of approximately 500 billion yuan in 2025. Of this, around 150 billion yuan was attributable to traditional IDC operating services, 250 billion yuan to AI computing services, and 100 billion yuan to domestically produced AI chips. By 2030, the AI computing market alone (i.e., computing services plus chips) is projected to grow to approximately 1.2 trillion yuan, which would correspond to an average annual growth rate of around 35 percent from 2025 onward.
Another structural change concerns cooling technology. While conventional air cooling still dominates today, liquid cooling is expected to become the standard for all new AI data center projects by 2027, with its share of newly installed racks exceeding 70 percent by 2030. For an acquisition vehicle like the Sanrenxing partnership, this means that simply acquiring an existing property without accompanying technical upgrades will hardly be sufficient to achieve competitive returns in the long term. True value preservation and appreciation will come from the ability to technically upgrade existing facilities to the denser, hotter, and more powerful AI rack generations.
The main players in the boom and where smaller investors fit in
The real driving force behind the nationwide investment boom is China's major cloud and technology companies. Alibaba announced plans to invest more in cloud and AI infrastructure over the next three years than it has spent cumulatively in the previous ten years – an amount initially set at 380 billion yuan, later increased to as much as 480 billion yuan. According to market reports, ByteDance is considering spending between $59 billion and $70 billion this year alone on data centers and AI infrastructure, more than double the previous year's level. Globally, according to the Dell'Oro Group, cumulative investment in data centers is expected to exceed $3 trillion by 2030, nearly double the figure projected in January.
Against this backdrop, Sanrenxing positions itself not as an independent hyperscaler, but as one of countless small and medium-sized players seeking access to this growth market through specialized investment vehicles, often in close cooperation with specialized operator platforms such as the United Data Center Group. This distribution of roles resembles an ecosystem in which a few large corporations assume technological and financial leadership, while a multitude of smaller investors co-finance the operational implementation of individual sites through specialized partnerships, relying on the expertise of established operators rather than developing their own technical know-how.
Beyond the borders of the country: China is exporting its data center boom
Also of interest is the increasingly international dimension of this trend. Chinese cloud companies are increasingly focusing their attention on Southeast Asia, where they are significantly driving up demand for leased data center and AI computing capacity. According to a CAICT report, total data center capacity in key Southeast Asian markets already exceeded 4.4 gigawatts in 2025. In the Malaysian region of Johor, around 850 megawatts of data center capacity were already operational in the second quarter of 2026, with another 1.8 gigawatts under construction and 2.7 gigawatts in development. This internationalization demonstrates that the race for computing power is no longer solely an internal Chinese phenomenon, but is increasingly taking on geopolitical dimensions. Chinese operators are strategically localizing their presence to circumvent regulatory hurdles and geopolitical tensions.
The hidden bottleneck: Memory chips as a global choke point
Another effect of the Chinese investment wave is impacting the global memory chip markets. Because both the United States and China are simultaneously investing heavily in data center capacity, the worldwide shortage of server memory and high-performance memory modules is becoming increasingly acute. According to forecasts by Counterpoint Research, server DRAM and high-bandwidth memory will account for approximately 57 percent of all DRAM units shipped and even 65 percent of total DRAM market revenue in 2026. While the Chinese memory manufacturer CXMT is expanding its capacity, industry estimates indicate that a significant quality and quantity gap remains compared to established suppliers like Samsung Electronics and SK Hynix, preventing it from fully meeting the rapidly growing memory demands of China's AI infrastructure. For every single data center project in China—even smaller ones like the one in Fangshan—this represents a latent procurement risk, as even financially secure projects can fail or be delayed due to global supply bottlenecks for critical components.
Why one inconspicuous news item reveals more than a thousand headlines
The Sanrenxing transaction is almost economically insignificant in its absolute scale, but its very casual nature makes it an extremely revealing case study. It demonstrates how deeply the state-mandated data center boom has now permeated the Chinese corporate landscape, with even publicly listed mid-sized companies from outside the sector establishing their own investment vehicles to secure a share of this growth market. At the same time, it illustrates the structural division of labor between large-scale state investments, the expansion of hyperscalers by their own companies, and private equity, which flows into specific individual projects through specialized partnerships such as those between Huixinfu and the United Data Center Group.
In the long term, the success of such small-scale investments depends less on the capital deployed than on two external factors over which the investors themselves have little control: the availability of sufficient and affordable electricity and unimpeded access to high-performance memory chips and AI accelerators. Should Beijing achieve its ambitious goal of building a nationwide interconnected computing network by 2028, early-positioned players like Sanrenxing could significantly benefit from the resulting increase in the value of strategically located existing facilities. However, if the energy and chip supply does not materialize at the required pace, even well-positioned projects like the one in Fangshan risk becoming mere administrative vehicles whose actual economic benefits will only materialize after a considerable delay.
📈🚀 From visibility to trust 👀🤝 Your scalable path with Xpert.Digital
In industrial B2B, sustainable business relationships rarely emerge overnight. They develop step by step – through visibility, professional relevance, recurring touchpoints, and growing trust. Xpert.Digital's 4-stage model addresses precisely this: It offers a structured path that begins with a manageable entry point and can evolve into deeper collaboration in business development if needed.
Instead of relying on loud marketing promises, this model puts the relationship at the forefront. Companies start with clearly defined, easily calculable measures and then decide, based on their own experience, how far they want to expand the collaboration. A key factor for this undisturbed trust-building process: The platform completely avoids annoying advertising ads, so the editorial focus remains solely on the companies' expertise.
More information here:
Your global marketing and business development partner
☑️ Our business language is English or German
☑️ NEW: Correspondence in your native language!
I and my team are happy to be available to you as your personal advisor.
You can contact me by filling out the contact form here [email protected]:or simply call me at +49 7348 4088 965. My email address is
I'm looking forward to our joint project.


















