On the evening of July 20, Jiechuang Intelligence (301248), a company listed on the Growth Enterprise Market, released a procurement announcement that stirred the entire capital market. The announcement revealed that the company's board of directors unanimously approved a proposal to procure IT equipment and components worth no more than 10 billion yuan, all earmarked for AI cloud computing power cluster construction and the implementation of self-developed technology research and development. The day after the announcement, the company's stock price opened significantly lower, closing down 5.71% for the day, with its market capitalization evaporating by over 460 million yuan. A wave of skepticism erupted on stock forums, institutional research groups, and financial media outlets.

New Finance News learned that Jiechuang Intelligence's total annual revenue in 2025 was merely 852 million yuan, making the 10 billion yuan procurement cap 11.7 times its annual revenue. As of the end of 2025, the company's audited total assets stood at 3.261 billion yuan, with net assets of 1.521 billion yuan. The billion-yuan procurement scale was 3.06 times and 6.57 times its total assets and net assets, respectively. In the first quarter of 2026, the company's asset-liability ratio soared to 73.77%, with only 1.28 billion yuan in monetary funds on hand. Its first-half earnings forecast further showed a shift from profit to loss.

What baffles investors even more is that this billion-yuan procurement was not a last-minute plan. In January 2026, the company had just disclosed a 4 billion yuan equipment procurement plan, only to increase the procurement limit by 150% within half a year. Simultaneously, the announcement cited "business secrets and preventing peer competition" as reasons to fully exempt the disclosure of all supplier information, leaving the counterparties of this billion-yuan transaction completely hidden.

On one side are deteriorating profitability, high debt levels, and meager internal funds; on the other is an exponentially expanding equipment procurement budget and large-scale omissions in information disclosure. Is Jiechuang Intelligence's billion-yuan expansion into AI computing power a forward-looking strategic move aligned with the computing power trend, or is it a reckless gamble that overextends the listed company and harms the interests of minority shareholders?

Procurement Budget Doubles in Six Months, Key Information Missing from Announcement

According to Jiechuang Intelligence's Announcement on Equipment Purchase and board resolution announcement released on July 20, 2026, the core terms of this procurement are as follows: First, procurement targets: Various IT servers, GPU computing power equipment, storage devices, network switches, and supporting spare parts and other hardware facilities, excluding land, data centers, personnel, or software procurement. Delivery is uniformly scheduled for completion in batches by the end of December 2026.

Second, funding cap: The total transaction amount does not exceed 10 billion yuan, adopting an "upper-limit framework procurement" model. Independent procurement contracts will be signed with multiple suppliers in batches, with payment methods including wire transfers, bank acceptance bills, and domestic letters of credit. Third, review process: All nine directors unanimously approved the proposal without any objections, abstentions, or recusals. The proposal still requires approval at the 2026 extraordinary general meeting before it can be officially implemented.

Fourth, information disclosure exemption: All supplier names, registered addresses, cooperation durations, pricing bases, unit procurement prices, and equipment details for this transaction will not be disclosed. The company justified this compliance decision by stating that "premature disclosure of supplier information would violate business confidentiality agreements, trigger malicious competitive bidding from peers, and undermine the company's strategic competitiveness in AI business."

Fifth, transaction classification: The company determined that this equipment procurement does not constitute a major asset restructuring or a related-party transaction, eliminating the need for a supporting asset appraisal report. It will only be disclosed as a routine operational equipment procurement matter. Sixth, risk warnings: The announcement voluntarily highlighted two major risks: First, funding pressure, as the company will rely on a combination of internal funds, bank credit lines, and external financing to pay for the procurement, with interest rate fluctuations potentially driving up financial expenses. Second, performance default risk—if financing funds fail to arrive in full, the company may be unable to make timely payments and could face substantial penalties.

Reviewing Jiechuang Intelligence's equipment procurement plans over the past year reveals an extremely aggressive expansion pace: June 2025: The board approved a 600 million yuan equipment procurement quota for early small-scale intelligent computing node setup, marking the company's first foray into computing power hardware procurement.

January 2026: The procurement cap was raised to 4 billion yuan, accompanied by an application for a 12 billion yuan comprehensive bank credit line. The controlling shareholder simultaneously provided no more than 1.5 billion yuan in interest-free loans to supplement liquidity. July 2026: Half a year later, the procurement limit was directly increased to 10 billion yuan, with the credit line expanded to 20 billion yuan, while the controlling shareholder's loan quota remained unchanged.

In just 13 months, the equipment procurement budget surged from 600 million yuan to 10 billion yuan, an expansion of over 15-fold. Compared to industry peers, annual capital expenditures for small and medium-sized computing power service providers in China typically do not exceed 1 billion yuan, while leading government and enterprise cloud vendors usually procure computing power equipment worth 3-5 billion yuan annually. Jiechuang Intelligence's billion-yuan procurement plan is extremely rare among small and medium-sized Growth Enterprise Market companies.

On the evening of the announcement, thousands of skeptical messages flooded major platforms and investor interaction forums for listed companies. Investors on stock forums pointed directly to core contradictions: "With annual revenue of 850 million yuan, how will the company repay billion-yuan equipment loans?" "By not disclosing suppliers, is there an off-balance-sheet funding cycle?" "Why did the order demand jump from 4 billion yuan to 10 billion yuan in just six months?"

On July 21, Jiechuang Intelligence's stock opened over 5% lower, oscillating downward throughout the day to close at 52.39 yuan, down 5.71%. Daily trading volume shrank to 213 million yuan, with northbound capital and public funds slightly reducing their holdings.

Financial Capacity Completely Out of Balance: Billion-Yuan Procurement to Push Company into High-Debt, Loss-Making Cycle

The first and most fatal layer of market skepticism centers on Jiechuang Intelligence's weak financial fundamentals, which are entirely unable to support 10 billion yuan in equipment investment. Leveraged expansion will continuously erode profits, creating a vicious cycle of "depreciation and interest double-eating into earnings."

As of March 31, 2026, Jiechuang Intelligence had 1.28 billion yuan in monetary funds on hand—all of its usable internal liquidity, covering only 12.8% of the 10 billion yuan procurement cap. The remaining nearly 8.7 billion yuan funding gap must be filled through debt instruments such as bank loans, bill financing, letters of credit, and shareholder loans.

The company's liability structure was already under severe pressure during this period: short-term borrowings of 926 million yuan, non-current liabilities due within one year of 627 million yuan, and long-term borrowings of 1.444 billion yuan, bringing total interest-bearing liabilities to 2.997 billion yuan. With net assets of just 1.508 billion yuan, interest-bearing liabilities were twice its net assets, and the asset-liability ratio stood at a lofty 73.77%, far exceeding the Growth Enterprise Market computer industry's average of 42%.

At its May earnings briefing, the company disclosed that it had applied for a combined 20 billion yuan in comprehensive credit lines from multiple banks, with nearly 10 billion yuan in approved credit currently available. This means the billion-yuan procurement plan will almost entirely rely on new bank credit for execution.

Relying entirely on debt financing for capital expenditures directly creates two irreversible financial pressures: First, explosive growth in interest expenses. With the current 1-year LPR rate at 3.45% and mid-to-long-term equipment loan rates floating 20-40% higher, if the 8.7 billion yuan funding gap is financed entirely at a 4.5% annualized rate, annual new interest expenses will reach approximately 392 million yuan.

Compared to the company's 2025 full-year net profit attributable to shareholders of just 13.6735 million yuan, the new annual interest is 28.5 times its full-year net profit. In the first quarter of 2026, the company's financial expenses already reached 14.5222 million yuan, soaring 2,831.79% year-on-year. The first-half loss announcement explicitly stated that the primary cause of losses was the surge in interest and depreciation from computing power expansion. If the 10 billion yuan in equipment is fully deployed, annual financial expenses will exceed 400 million yuan, plunging the company into long-term losses.

Second, fixed asset depreciation will sharply compress gross margins. AI servers and GPU equipment typically depreciate over 3-5 years, with annual depreciation for 10 billion yuan in equipment reaching 2-3.3 billion yuan. In 2025, the company's total revenue was only 852 million yuan. Even if AI cloud computing business doubles in 2026, annual computing power-related revenue is unlikely to exceed 500 million yuan, with annual depreciation costs far exceeding total business revenue. The computing power segment will continue to incur losses, dragging down the company's overall gross margin.

The core survival logic in the capital-intensive computing power sector is sustained positive operating cash flow to cover depreciation and financial expenses. However, Jiechuang Intelligence's cash flow indicators continue to deteriorate, leaving it entirely without the cash flow foundation to support 10 billion yuan in capital expenditures: In 2025, full-year operating cash flow net of 29.2026 million yuan, with the core business bleeding cash throughout the year. In 2025, investing cash flow net of 852 million yuan, all allocated to early computing power equipment procurement, reflecting sustained large-scale capital investments. All cash growth came from financing activities, with 2025 financing cash flow net of 1.105 billion yuan. The company relies solely on borrowing and financing to sustain operations, lacking internal cash generation capacity.

In the first half of 2026, the company disclosed AI + cloud computing business revenue of 204 million yuan, up 550% year-on-year. While the growth rate appears impressive, critical data blind spots remain in the market: The company did not disclose this segment's gross margin, the contract value of long-term locked-in customer agreements, or the scale of stable annual cash collections.

With 204 million yuan in computing power revenue for the first half of the year, against annual depreciation and interest costs of 3 billion yuan from the 10 billion yuan in equipment, a 50-fold gap exists between revenue scale and capital expenditure costs. Even if computing power business maintains 100% annual growth, it will take at least 8-10 years to cover the annual fixed costs from this billion-yuan procurement. With equipment update cycles of just 3-5 years, the mismatch between equipment lifespan and payback period creates extremely high risks of equipment idleness and asset impairments.

In 2025, the company appeared to return to profitability with net profit attributable to shareholders of 13.6735 million yuan. However, dissecting the profit structure raises questions about earnings authenticity: Full-year net profit excluding non-recurring items was -58.536 million yuan, with core operations still incurring losses. Accounting profits relied entirely on one-time gains such as government subsidies and changes in fair value of assets.

The first-half 2026 earnings preview directly returned to losses, with net profit attributable to shareholders ranging from -2.2 million to -4.4 million yuan, shifting from profit to loss year-on-year. The company attributed losses to increased depreciation and interest from computing power expansion, indirectly confirming that the computing power business currently generates no positive profits.

Business Demand Cannot Match Billion-Yuan Equipment Capacity: High Risk of Overcapacity in Computing Power Sector

The company's only plausible explanation for the billion-yuan procurement is to implement its AI cloud computing strategy and meet government and enterprise customers' demands for computing power leasing and private cloud deployment. However, analyzing industry landscape, customer base, and business execution capacity reveals that the computing power capacity corresponding to 10 billion yuan in equipment has no matching downstream order support. Massive equipment idleness and significant asset impairments loom as the greatest risks.

In 2025, China's total cloud computing market size was approximately 420 billion yuan, with Alibaba Cloud, Tencent Cloud, Huawei Cloud, and China Telecom Cloud collectively holding 72% market share. These leading vendors possess self-developed chips, self-built large-scale data centers, nationwide networks, and low-cost scaling advantages. As a regional small-to-medium service provider, Jiechuang Intelligence's core market focuses on government and enterprise clients in South China, with industry barriers, customer resources, and funding costs all far inferior to those of leading vendors.

Leading cloud vendors finance their annual billion-yuan capital expenditures through ample group cash flows to amortize costs, while Jiechuang Intelligence—a single listed company—independently shoulders 10 billion yuan in debt, leaving its computing power leasing pricing power inherently disadvantaged. Should the industry face computing power oversupply, leading vendors could seize orders through price cuts, forcing smaller players to concede margins further and even leave equipment idle without revenue.

Currently, governments and enterprises across China are massively investing in intelligent computing centers. In 2026, the total scale of computing power clusters under construction or planned nationwide will exceed 200 billion yuan, with industry overcapacity expectations already forming. Multiple regional computing power service providers saw data center vacancy rates exceed 40% in 2025. Jiechuang Intelligence's countercyclical bet on 10 billion yuan in computing power equipment precisely coincides with the industry's capacity expansion peak, substantially raising the difficulty of demand fulfillment.

According to 2025 annual report customer data, the company's top five customers accounted for 358 million yuan in combined annual sales, or 41.97% of total revenue. Customers were highly fragmented, with the largest single customer contributing just 134 million yuan in annual purchases. Traditional businesses (system integration, security equipment) accounted for over 70% of customer orders, with AI computing power-related orders making up less than 30%.

First-half 2026 computing power business revenue was 204 million yuan, representing only short-term revenue for the period. No announcements disclosed long-term framework cooperation agreements or annual locked-in computing power leasing orders. The market's core question remains: If all 10 billion yuan in equipment is delivered, the company's annual computing power service capacity will reach 3-5 billion yuan in revenue scale. Without long-term large orders to support this, how will the new capacity be absorbed?

Government and enterprise computing power projects feature long bidding cycles, slow cash collections, and fierce competition. A single large intelligent computing project takes 1-2 years to land, with payments received in 3-5 annual installments. Without confirmed large orders as collateral, Jiechuang Intelligence's advance lock-in of 10 billion yuan in equipment procurement represents a classic heavy-asset operating model of "building factories before securing customers." Order fulfillment falling short of expectations will directly result in billions of yuan in idle fixed assets, triggering substantial asset impairments that could wipe out the company's entire net assets.

Jiechuang Intelligence initially focused on government and public security system integration and communication security equipment sales, only formally entering the AI cloud computing sector in 2024. With just two years of transformation, it remains a newcomer in the computing power industry. Compared to professional computing power service providers with over a decade of experience, the company suffers from multiple operational shortcomings:",

However, in this announcement, Jiechuang Intelligent simply stated 'trade secrets and to prevent competition between peers (horizontal competition)' without providing any supporting materials. It did not specify which domestic chip/server manufacturer the supplier belongs to, did not disclose the number of cooperative suppliers, did not provide a comparison of fair equipment procurement prices, and did not demonstrate what specific commercial losses would result from disclosing the suppliers. The counterparty in this transaction worth tens of billions has completely vanished, representing an extremely rare case of minimal information disclosure.

Previous inquiry cases by regulatory authorities show that when multiple listed companies were exempted from disclosing suppliers for large-scale equipment procurement, they all received inquiry letters from the exchange, requiring them to supplement details about the suppliers, pricing basis, and whether there were any hidden related-party relationships.

In addition, a complete announcement for major equipment procurement must list the equipment categories, procurement quantities, unit prices, and a comparison with the average market prices for the same model of equipment, in order to prove that the procurement prices are fair and that there is no situation of overpriced asset procurement. However, in this announcement by Jiechuang Intelligent, there are no details about any equipment, making it impossible for investors to verify whether the hardware quantities and unit prices corresponding to the procurement worth tens of billions deviate from fair market prices.

Currently, the market prices for AI servers and GPU chips are transparent, and bulk procurements by leading manufacturers come with a 10%-20% discount. If the company procures equipment at high prices from concealed suppliers, it will directly lead to an overstatement of the listed company's assets, with future equipment depreciation and impairment losses further expanding. In the absence of three key sets of data—pricing, equipment details, and supplier information—minority shareholders cannot determine whether this transaction is fair, and their right to information is severely compromised.

The procurement budget has skyrocketed in a stepwise manner, from 600 million to 4 billion and then to 10 billion, with no supporting evidence of new orders or an explosion in downstream demand, making the expansion pace highly abnormal. For the 4 billion equipment procurement plan disclosed in January 2026, as of the announcement in July six months later, the company has never disclosed in earnings briefings or interim announcements the actual contracted amount, equipment delivery progress, payments made, or the scale of fixed assets formed for the 4 billion procurement.

Investors have repeatedly inquired about the progress of the procurement on interactive platforms, but the company has only vaguely replied that 'procurement is progressing steadily and construction in progress continues to increase,' without providing specific quantitative data. The company should first implement the 4 billion procurement, then evaluate whether to increase equipment investment based on order fulfillment and the revenue from computing power operations. However, without clarifying the progress of the initial large-scale procurement and without validating the profit model, the company has directly raised the procurement cap to 10 billion, which is a radical expansion that completely violates the logic of gradual corporate investment.

The company repeatedly emphasizes that its AI cloud computing revenue in the first half of 2026 will be 204 million, a year-on-year increase of 550%, as evidence of the rationality of the procurement worth tens of billions. However, this growth rate is significantly influenced by a low base: the computing power business in the first half of 2025 had a base of only 31 million, and the low base has led to high growth, but the absolute revenue scale remains small.

Breaking down the industry demand logic, government and enterprise computing power procurement is cyclical, with local government smart computing center tenders concentrated in the fourth quarter of each year, resulting in naturally fewer computing power orders in the first half of the year. The revenue of 204 million in the first half of the year is not sustainable. Moreover, the gross profit margin of the computing power business is only 33.86%, lower than the 42.94% gross profit margin of the company's traditional security business, indicating weak growth quality. Even if the annual computing power revenue exceeds 500 million, it will still be unable to cover the annual fixed costs brought by the equipment worth tens of billions, and the short-term high growth rate cannot support a capital investment of tens of billions.

The long-term development space of the AI computing power industry is undeniable, as the digital transformation of domestic government and enterprise sectors generates a massive demand for computing power, and the layout of smart computing infrastructure itself has industrial logic. However, Jiechuang Intelligent's 10 billion equipment procurement plan exposes the core contradiction of small-cap listed companies venturing into capital-intensive sectors: the prospect of the industrial sector is severely disconnected from the company's own financial and business carrying capacity. Radical expansion, compounded by a lack of information disclosure, amplifies the operational risks throughout the entire chain.

Currently, the controversy over Jiechuang Intelligent's procurement worth tens of billions has not yet concluded. The shareholder meeting vote, the exchange's inquiry response, and the subsequent progress of the procurement will continue to verify whether this tens-of-billions computing power expansion is a forward-looking strategic move or a blind gamble that overextends the listed company. Xin Cai Wen Network will continue to follow the story.