1.65 billion impairment pressure, 4000P intelligent computing paving the way: the stock dilemma and growth opportunity for Aurora Cloud Network

Image source: TuChong Creative

Source | Times Business Research Institute

Author | Hao Wenran

Editor | Han Xun

In January 2026, Huanhuan New Network (300383.SZ) released a highly contradictory 2025 performance forecast: 26,000 newly commissioned server racks for the full year, hitting a historical high; revenue of RMB 7.17 billion to RMB 7.22 billion; slightly down year over year, but the net profit attributable to shareholders is expected to be a loss of RMB 730 million to RMB 780 million.

This leading player in China’s third-party data center market is facing a mix of ice and fire: on one side, the market views it as actively positioning itself for ByteDance’s AI computing demand; on the other, it has been dragged into loss territory by a goodwill impairment of RMB 865 million. Huanhuan New Network is standing at a critical gear-shift period—its strategic leap from a traditional IDC landlord to an AIDC ecosystem service provider is a comeback battle it must win.

Three layers of pressure on strategic transformation: from “selling server racks” to “selling services”

Huanhuan New Network has two main businesses: first, the Internet data center business (IDC). It builds data centers across the country, rents server racks to internet giants, financial institutions, and cloud computing service providers, and provides server hosting services; second, the cloud computing business, as an important operator of Amazon Web Services (AWS) in China, providing cloud solutions.

In terms of revenue contribution, cloud computing and IDC are roughly a “30/70 split”; on the profit side, the situation is completely reversed—IDC’s gross margin is about 32%-35%, while cloud computing is only 7%-9%. IDC has long contributed 60%-70% of the company’s profits.

This stems from the different business models of the two. Cloud computing business is made up of AWS operations and the Wushuang Technology acquired in 2015; both follow an agency/service model, earning service fees or revenue sharing. By contrast, the IDC business builds data centers on its own land and earns rental price differentials by renting out server racks—this is a capital-intensive, high-gross-margin profit engine.

However, this “source of profit” is now facing three layers of pressure.

The first layer comes from industry competition. In parts of the IDC industry, supply and demand are imbalanced and price wars have become the norm, weakening the company’s bargaining power as a supplier. As a result, IDC business gross margin fell from 37.30% in 2022 to 32.71% in the first half of 2025.

The second layer comes from the cost side. In 2025, the company rolled out 26,000 new server racks, setting a historical high, but the overall rack utilization rate is only about 60%. After the new racks are commissioned, depreciation costs increase significantly. During the ramp-up period, coupled with some customers canceling their leases, revenue fails to keep pace, causing unit costs to rise and putting pressure on the profit side.

The third layer comes from asset impairments. In 2025, the company recorded a goodwill impairment of RMB 865 million. Of this, RMB 838 million was recorded for Zhongjin Yunwang acquired in 2016; RMB 27.1296 million was fully recorded for Huanhuan Zanzup acquired in 2022. After the impairment, Zhongjin Yunwang’s remaining goodwill is about RMB 200 million.

Behind it lies three layers of pressure; ahead is the AI wave—Huanhuan New Network has proactively started its strategic transformation.

On March 11, 2026, the company’s controlling subsidiary Huanhuan Cloud announced the acquisition of an AI engineering delivery team, Baiyang Intelligent, and released the “Panacea Intelligent Agent Training and Push Platform” co-built by both sides. A research report by Kaiyuan Securities said that this acquisition marks the company’s “full-scale acceleration of AI intelligent agent development and end-to-end operations deployment.” Essentially, it is a precise strategic alignment between the underlying capabilities of new AI computing infrastructure and the engineering delivery capabilities of AI Agents.

It is worth noting that Huanhuan Cloud, the company’s controlling subsidiary, is the core authorized cooperation partner for Volcano Engine. Volcano Engine is an enterprise-level technical services platform under ByteDance. Although the company did not disclose specific cooperation details on the grounds that it has signed confidentiality agreements with customers, the market generally speculates that the two sides have in-depth cooperation.

More practically, as of March 2026, the company has deployed high-performance computing hardware in the Beijing KeXIn ShengCai and Tianjin ZANPU cloud computing centers, forming an intelligent computing scale of over 4,000P. Annual contract value has exceeded RMB 100 million. Although this figure is less than 2% of total revenue, its strategic significance is far greater than its financial contribution—it indicates that the company has officially extended its business to “providing computing power services.”

In addition, in December 2025, Huanhuan New Network signed a strategic cooperation agreement with Inner Mongolia Unicom. It plans to build a large-scale same-city dual-active green intelligent computing center cluster in Hohhot; after the two projects are completed, the two same-city dual centers can serve as backups to each other. In February 2026, Huanhuan New Network also jointly established Ganzan Data Information Technology (Kunshan) Co., Ltd. with Gree Electric, focusing on liquid cooling technology, green computing equipment, and end-to-end solutions for intelligent computing centers.

This series of moves clearly points to Huanhuan New Network’s intention to transform its business strategy toward AI, and also reflects the differences in business models between its IDC and AIDC businesses.

The core of traditional IDC is “selling server racks”: providing the data center environment and rack resources, while customers configure servers themselves. The company earns the rental price differential. AIDC requires the service provider not only to provide high-power server racks, but also to have capabilities such as liquid cooling heat dissipation, intelligent scheduling, computing power operations, and even AI application development. Its profit model evolves from “rental” to “computing power service fees + platform revenue sharing.”

The annual rent for a single-rack unit in traditional IDC is about RMB 72,000, while a training-type rack equipped with GPUs can reach RMB 368,000 to several million per year. Gains of 5 times or more may, to a large extent, open up Huanhuan New Network’s performance upside ceiling.

However, behind the high returns there are also risks: AIDC’s upfront capital expenditures far exceed those of traditional IDC, and it is extremely sensitive to technology iteration. Equipment updates are frequent and assets depreciate quickly. This requires the company to remain highly prudent in both selecting its technology roadmap and pacing its capital investment.

Impairment burden and pains of transformation

The future prospects of the transformation are certainly enticing, but what Huanhuan New Network must face right now is the reality of a weak fundamental picture.

If the AIDC transformation represents the company’s “future,” then the RMB 865 million goodwill impairment is the heavy burden left by its “past.” The goodwill impairment recorded this time mainly comes from two acquisitions: the RMB 1.952 billion goodwill formed from acquiring 100% of Zhongjin Yunwang in 2016, and the RMB 27.1296 million goodwill formed from acquiring 65% of Huanhuan Zanzup in 2022.

The direct reasons for goodwill impairment are that intensified industry competition leads to sustained price declines, and that the rise of AI technology brings structural changes in customer demand. Early-built data centers such as Zhongjin Yunwang suffer from equipment aging and rising retrofit costs, while operating costs continue to increase. Meanwhile, the trend of customers concentrating on leading players further weakens their bargaining power. The company has stated that the current customer structure of its IDC business shows the characteristics of “large order volume, few customer groups, and high stickiness,” meaning that customers have strong bargaining power and acquiring customers is difficult.

This also reflects Huanhuan New Network’s recent operating snapshot: over the past three years, the fundamentals have shown the traits of “both revenue and profit declining.” In 2023, revenue was RMB 7.855 billion and net profit attributable to shareholders was RMB 388 million. In 2024, revenue was RMB 7.281 billion and net profit attributable to shareholders was RMB 381 million. In 2025, revenue is expected to be RMB 7.17 billion to RMB 7.22 billion, and net profit attributable to shareholders is expected to plummet to a loss of RMB 730 million to RMB 780 million.

Even after excluding the impact of goodwill impairment, the company’s 2025 net profit attributable to shareholders is about RMB 85 million to RMB 135 million, down 64.61%-77.72% year over year—still showing a cliff-like decline. The drop in net sales profit margin is even more pronounced: from 10.58% in 2021 down to 2.83% in the first three quarters of 2025. This reflects the continued erosion of the company’s core business profitability.

Looking ahead, the intensely competitive landscape in the IDC industry is unlikely to change in the short term, and Huanhuan New Network’s traditional main business may continue to face pressure. On the supply side, the company has a rich project pipeline, with more than 230,000 racks planned nationwide. In 2026, it will continue to推进 (advance) the construction progress of the Changsha and Hangzhou projects, but the issue of insufficient rack utilization rate in the short term is still hard to alleviate. On the demand side, the incremental demand brought by AI is mainly concentrated in high-power racks, while renovating existing low-power racks requires time and capital investment.

The company states that whether it will continue investing and laying out the computing power business in the future will be decided cautiously based on customer order situations. This means that, given that the company’s traditional business cash generation capacity has weakened significantly while new business requires substantial investment, Huanhuan New Network must carefully balance resources.

Core viewpoint: 2026 is a critical period for the transformation gear shift

Huanhuan New Network’s investment value is shifting from “IDC landlord” to “AIDC ecosystem value realization.” The confirmation of RMB 865 million goodwill impairment reflects the aging of assets in its traditional business and the bottoming out of the cycle. The linkage of 4,000P intelligent computing and Volcano Engine opens up new narrative space.

The key current contradiction is this: transformation requires time and investment, while losses are eroding patience. In 2026, the speed of customer rollout for the Panacea platform and the gross margin level of its intelligent computing business will be the key to validating whether the “gear shift” succeeds or fails. If it can successfully build a dual-wheel drive model of “traditional business cash flow as a backstop + new business valuation uplift,” this long-established IDC leader may see a value re-rating amid the AI wave. If the new business scales up below expectations, the pressure from existing assets will continue to suppress valuation repair.

(Full text 2,558 Chinese characters)

Disclaimer: This report is intended for use only by clients of the Times Business Research Institute. This company does not consider the recipient as its client merely because the recipient receives this report. This report is prepared based on information that this company deems reliable and has been made public, but this company provides no guarantee regarding the accuracy or completeness of such information. The opinions, assessments, and forecasts contained in this report only reflect the viewpoints and judgments made on the date the report is issued. This company does not guarantee that the information contained in this report remains up to date. The company may modify the information contained in this report without issuing notice; investors should monitor relevant updates or modifications on their own. This company strives to present the contents objectively and fairly, but the viewpoints, conclusions, and recommendations in this report are for reference only and do not constitute a solicitation for the purchase or selling price of the securities mentioned. These viewpoints and recommendations do not take into account the specific investment purposes, financial situations, and specific needs of individual investors, and do not constitute private investment advice to clients at any time. Investors should fully consider their own specific circumstances and fully understand and use the contents of this report; they should not treat this report as the sole factor for making an investment decision. For any consequences arising from or related to reliance on or use of this report, this company and the author shall not bear any legal responsibility. Within the scope of what the company and the author know, there are no legally prohibited interests or relationships with the securities or investment targets referred to in this report. Where permitted by law, this company and its affiliated institutions may hold securities positions issued by the companies mentioned in this report and conduct trades, and may also provide or seek to provide related services such as investment banking, financial advisory, or financial products. The copyright of this report belongs only to this company. Without the company’s written permission, no institution or individual may in any form infringe the company’s copyright by reproducing, copying, publishing, quoting, or redistributing others’ materials. If approval is obtained from this company to quote and publish, the use must be within the permitted scope and include the source as “Times Business Research Institute,” and no quoting, abridgment, or modification that is contrary to the original intent of this report is allowed. This company reserves the right to pursue relevant liabilities. All trademarks, service marks, and labels used in this report are the trademarks, service marks, and labels of this company.

A massive amount of information and precise interpretation—available on the Sina Finance app