Kingsoft Cloud delivered impressive financial results, driven by a surge in artificial intelligence demand. On Wednesday, the company reported its fourth quarter and full-year 2025 performance.
Fourth-quarter revenue reached a historic high of 2.761 billion yuan, marking a 23.7% year-over-year increase, with growth accelerating from previous quarters. Public cloud revenue was a core growth driver, surging 34.9% year-over-year to 1.902 billion yuan. The key factor behind this strong public cloud growth was AI-related demand. The company disclosed that fourth-quarter "AI-optimized cloud services billings" reached 926 million yuan, a substantial 95% increase compared to the same period last year. Management stated they expect robust demand for intelligent cloud computing to continue into 2026.
Signs of structural improvement emerged on the profitability front. Non-GAAP operating profit was positive for the second consecutive quarter, reaching 54.6 million yuan. Non-GAAP EBITDA doubled year-over-year to 785 million yuan, with the EBITDA margin rising to 28.4%. However, due to the heavy-asset expansion phase, the company still reported a GAAP net loss of 163 million yuan, and the gross margin declined year-over-year, primarily impacted by increased depreciation from investments in AI computing infrastructure.
For the full 2025 fiscal year, total revenue was 9.559 billion yuan, a 22.8% increase year-over-year. The GAAP net loss narrowed significantly to 944 million yuan from 1.979 billion yuan in 2024. Non-GAAP EBITDA reached 2.336 billion yuan, with the EBITDA margin rising to 24.4%.
**Revenue: Public Cloud Accelerates, AI Computing Drives Largest Increment**
The financial report showed fourth-quarter total revenue of 2.761 billion yuan, an 11.4% increase sequentially. Breaking down the structure:
* **Public Cloud:** Revenue was 1.902 billion yuan, up 34.9% year-over-year and 8.6% sequentially. The company attributed growth primarily to AI demand, disclosing quarterly AI-optimized cloud billings of 926 million yuan, a 95% year-over-year increase. * **Industry Cloud:** Revenue was 859 million yuan, increasing 4.5% year-over-year and 18.4% sequentially. The company explained that the sequential improvement in Industry Cloud was related to "centralized delivery," and enhanced profit contributions from Industry Cloud projects also supported the sequential recovery in gross margin.
By proportion, Public Cloud revenue accounted for approximately 69% of total Q4 revenue, while Industry Cloud accounted for about 31%. Driven by AI computing demand, Kingsoft Cloud's growth is increasingly fueled by "AI computing" on the Public Cloud side, while Industry Cloud acts more as a stable revenue base providing delivery flexibility.
**Costs and Gross Margin: Rising Depreciation Weighs on Margin, Sequential Improvement from Industry Cloud**
The company's gross profit for the fourth quarter was 465 million yuan, up 9.2% year-over-year and 22.2% sequentially. However, the gross margin was 16.9%, down from 19.1% a year ago, reflecting faster growth on the cost side, particularly pressure from depreciation related to AI computing resources. Key changes in cost of revenues (2.296 billion yuan in the quarter, up 27.1% year-over-year) were:
* **IDC Costs:** 812 million yuan, up 12.5% year-over-year, primarily due to leasing additional cabinets to support AI computing business expansion. * **Depreciation and Amortization:** 741 million yuan, significantly higher than the prior year (roughly doubling), mainly due to depreciation on newly acquired/leased servers and network equipment "primarily related to the AI-optimized cloud business." * **Solution Development and Service Costs:** 642 million yuan, up 15.3% year-over-year, attributed to an expansion of solution personnel.
Therefore, Kingsoft Cloud presented a typical financial profile for an "AI computing expansion phase": high revenue growth, simultaneous cash consumption and asset expansion, and rapidly rising depreciation pressuring gross margins. On the other hand, the gross margin recovered sequentially from Q3's 15.4%, which the report attributed to improved profit contributions from Industry Cloud projects. This suggests that while heavily investing in AI computing assets, the company is attempting to mitigate some cost pressures through project structure and delivery efficiency. (On a Non-GAAP basis, excluding share-based compensation costs included in cost of revenues, the Q4 Non-GAAP gross margin was 17.1%, showing little difference from the GAAP figure, indicating that depreciation and other "hard costs" are the primary cause of margin pressure.)
**Expenses: Sales Expenses Fall Sequentially, but Share-Based Compensation Lifts G&A**
Total operating expenses for the fourth quarter were 532 million yuan, up 13.3% year-over-year and essentially flat sequentially (+1.1%). Breakdown:
* **Sales and Marketing Expenses:** 123 million yuan, up 6.2% year-over-year but down 19.2% sequentially. * **General and Administrative Expenses:** 219 million yuan, up 21.8% year-over-year and increasing 25.4% sequentially, primarily due to increased share-based compensation expenses. * **Research and Development Expenses:** 190 million yuan, up 9.2% year-over-year but down 4.7% sequentially.
Overall, the expense ratio did not show runaway increases, but the impact of share-based compensation on the GAAP income statement is growing, which is part of the context for the company's emphasis on Non-GAAP metrics.
**Profitability: Non-GAAP Operating Profit Stays Positive, But Watch for One-Time Items and Finance Costs**
The income statement showed improvement at the operating level, but net income was still impacted by financial items:
* **GAAP Operating Loss:** The Q4 operating loss was 66.45 million yuan, wider than the prior year's loss of 43.51 million yuan, but a significant improvement from Q3's operating loss of 145 million yuan. * **Non-GAAP Operating Profit:** Profit for the quarter was 54.6 million yuan (operating margin ~2.0%), marking the second consecutive quarter of positive Non-GAAP operating profit.
Notably, the company also disclosed a "normalized" figure: excluding the impact of a 72.66 million yuan gain from disposal of property and equipment, the normalized Non-GAAP operating profit was -18.08 million yuan (operating margin ~-0.7%). This indicates the company is very close to achieving comprehensive, stable profitability at the operating level, but the quarter was still influenced by certain one-time factors.
* **GAAP Net Loss:** 163 million yuan (net loss margin -5.9%), narrower than the prior year's loss of 201 million yuan, but significantly wider than Q3's net loss of 7.847 million yuan. * Rising interest expense was a significant pressure point on net profit: Q4 interest expense was 153 million yuan, substantially higher than the 61.82 million yuan a year ago and also above Q3's 137 million yuan. During the phase of AI-related capital expenditure and financing expansion, financial costs have a more direct erosive effect on profitability. * **Non-GAAP EBITDA:** 785 million yuan, up 118.3% year-over-year, with an EBITDA margin of 28.4%. Excluding the gain on asset disposal, normalized Non-GAAP EBITDA was 713 million yuan, corresponding to a margin of 25.8%.
**Cash Flow and Balance Sheet: Operating Cash Flow Strengthens, Financing Offsets High Capex**
On the cash flow side, the company achieved in Q4:
* **Net cash provided by operating activities** was 1.043 billion yuan (compared to 570 million yuan a year ago), indicating significantly enhanced cash generation capability from revenue expansion, collections, and cost control. * **Net cash used in investing activities** was 428 million yuan, reflecting the ongoing high capital expenditure phase. * **Net cash provided by financing activities** was 1.514 billion yuan, corresponding to the company's described funding logic of "equity financing inflows being offset by capital expenditures."
As of the end of 2025:
* Cash and cash equivalents stood at 6.018 billion yuan (a significant increase from 3.955 billion yuan at the end of Q3). * The net book value of property and equipment rose to 10.095 billion yuan (compared to 4.630 billion yuan at the end of the previous year), visually reflecting the scale of investment in AI computing-related servers and network equipment. * Borrowings also expanded: short-term borrowings were 3.348 billion yuan, and long-term borrowings were 3.024 billion yuan. In the "supply-first" phase of AI computing capacity build-out, asset expansion, rising depreciation, and increasing interest expenses often occur simultaneously. Subsequent margin recovery will depend more on whether utilization rates, pricing, and customer structure for AI computing resources can be continuously optimized.
Comments