Author: Gao Linglang Editor: Ai Qingshan

In the late hours of September 28, Feng Assist made an announcement revealing that it had inked computing power server procurement agreements and computing power service agreements with two companies, codenamed C and D, respectively. The total procurement amount stood at 3.96 billion yuan, while the service amount reached 5.717 billion yuan, summing up to nearly 9.7 billion yuan, with a service cooperation period spanning 60 months.

Following this announcement, Feng Assist's stock price initially experienced an uptick, only to reverse course just two days later. On September 30, it plummeted by 9.11%, closing at 36.33 yuan, with its total market capitalization hovering around 10.366 billion yuan.

This marks the second round of significant computing power contracts secured by Feng Assist within a two-month timeframe. Earlier in August this year, its wholly-owned subsidiary, Ya'an Yun Suan, signed procurement and service agreements with two other undisclosed companies, with a combined value of 7.67 billion yuan.

Cumulatively, the locked-in contract value for Feng Assist has surpassed 17.3 billion yuan, a figure more than eight times its projected 2025 annual revenue of 2.09 billion yuan.

However, this impressive contract volume stands in stark contrast to the company's financial health. According to its 2026 mid-year report, Feng Assist had only 329 million yuan in cash on hand, with short-term borrowings and current liabilities due within one year totaling 1.275 billion yuan, and an operating cash flow of -528 million yuan. As large contracts continue to materialize, Feng Assist's liquidity is under considerable strain.

Feng Assist's business model is relatively straightforward. The operational approach for both rounds of contracts follows a similar pattern: the company purchases computing power servers from upstream suppliers, deploys them, and then provides five-year computing power services to downstream clients, earning an operational margin in the process.

By securing downstream demand before procuring upstream equipment, this common industry practice, known as the 'back-to-back' model, theoretically mitigates the risk of idle capacity. However, if any link in the chain falters, the project could quickly shift from profitability to loss.

Financially, both rounds of procurement rely on the same financing scheme, with only 5% to 20% of the funds coming from the company's own capital and the remaining 80% to 95% financed through leasing.

In essence, Feng Assist has leveraged a few billion yuan in equity to drive approximately 7 billion yuan in total procurement. Under the financial leasing model, server ownership remains with the leasing company, and Feng Assist must make timely rental payments to maintain equipment usage rights and its ability to fulfill contracts.

To support this, Feng Assist has expanded its comprehensive credit line to no more than 30 billion yuan, with unsecured joint guarantees provided by its actual controller, Luo Hongpeng.

Under such high leverage, profit margins are quite limited. In its August announcement, Feng Assist estimated that the average annual net profit for the first-round project would range between 60 million and 72 million yuan. Given the procurement cost of 3.062 billion yuan, the static annualized return on assets would be only around 2%.

Bai Wenxi, Deputy Director of the China Enterprise Capital Union, estimated that if both rounds of procurement follow a similar leverage ratio, the scale of new interest-bearing debt could approach 6 billion yuan, significantly raising the overall asset-liability ratio from its current 50.18%.

Misalignments in contract terms also pose significant risks. If upstream suppliers delay deliveries, the maximum compensation for breach of contract is only about 5 million yuan. However, in downstream service agreements, liability for data security incidents or service interruptions can reach 30% to 50% of the total contract value.

If upstream delays occur, Feng Assist would face compensation pressures far exceeding any compensation it could recover from upstream suppliers. Additionally, depreciation, amortization, and financing interest accrue before revenue is realized, while revenue is recognized gradually based on server deployment batches. Profit pressures in the early stages are inevitable, and the company has acknowledged that 'net profit in individual years may decline.'

Feng Assist's journey to its current position has followed its own unique developmental trajectory. Founder Luo Hongpeng, with years of experience in the telecommunications industry, established Feng Assist in 2012. Initially focusing on the aggregated operation of virtual products such as phone credits, data plans, and video memberships, the company went public on the ChiNext board of the Shenzhen Stock Exchange in 2023. Luo currently holds approximately 24.16% of the shares directly and indirectly, making him the company's actual controller.

After more than a decade in this business, growth bottlenecks have become evident. According to its 2025 annual report, revenue from digital commodity operations reached 1.88 billion yuan, accounting for nearly 90% of total revenue. However, the gross profit margin slipped from 18.67% in 2024 to 16.83% and further narrowed to 12% in the first half of this year.

Meanwhile, Feng Assist's 2025 net profit attributable to shareholders was only 154 million yuan, widening the gap with its revenue scale. The ceiling for its traditional core business is clearly visible, and new sources of growth are urgently needed.

The market potential for the computing power sector is indeed substantial. According to data from the China Academy of Information and Communications Technology, the domestic computing power leasing market is expected to surpass 260 billion yuan by 2026. As of the end of June this year, the overall utilization rate of computing power facilities nationwide had reached 71.4%.

Feng Assist stated that it began laying out its computing power operations business as early as 2019, extending from traffic operations to computing power operations. At its core, both involve the aggregation and scheduling of communication resources, following a consistent logic.

However, analysts do not fully agree with this characterization. Jiang Han, a senior researcher at the Pangu Think Tank, pointed out that shifting from a lightweight channel distribution model to a capital-intensive computing power operations model places vastly different demands on the company's capital allocation, technical operations, and risk management capabilities. It more closely resembles embarking on a second entrepreneurial venture rather than a simple business extension.

Currently, this model leans more toward basic computing power leasing, making it difficult to directly leverage capabilities accumulated from its original digital commodity operations.

Hundred-billion-yuan contracts may look impressive on paper, but true success lies in delivering profits. The progress of server deployments, trends in financing costs, and the payment schedules of downstream clients are all critical variables that will determine the success of this transformation. Only time will tell the answers.

Disclaimer: The content of this article is for reference only. The information or opinions expressed herein do not constitute any investment advice. Readers are advised to make investment decisions cautiously.