Kingsoft Cloud (NASDAQ:KC) highlighted record quarterly revenue growth and improving profitability on its fourth quarter and full-year 2025 earnings call, driven by surging demand for AI-related cloud services and expanding adoption across both ecosystem and external customers.
Fourth-quarter results set new revenue high
Chairman and CEO Tao Zhou said the company delivered “a historical high” quarterly revenue of RMB 2.76 billion, up 24% year over year. Public cloud revenue rose 35% to RMB 1.9 billion, with management pointing to strong momentum in intelligent computing services.
CFO Yi Li said total fourth-quarter revenue was RMB 2,761 million, consisting of:
- Public cloud services: RMB 1,902 million (up 35% year over year)
- Enterprise cloud services: RMB 859 million (a seasonally strong quarter with a high volume of project completion)
Profitability improves; operating profit positive for second straight quarter
Management emphasized continued progress on margins. Zhou said adjusted gross margin rose sequentially to 17.1% and adjusted operating margin reached 2.0%, marking operating-level profitability for two consecutive quarters.
Li reported adjusted operating profit of RMB 55 million in the quarter (a 2% margin), and non-GAAP EBITDA of RMB 785 million, up from RMB 360 million a year earlier. She attributed the improvement to a larger contribution from AI-related business, scale expansion, and cost and expense discipline, while noting that depreciation is a primary cost component given infrastructure investment.
On costs, Li said fourth-quarter cost of revenues rose 27% year over year to RMB 2,296 million, driven mainly by infrastructure investment to support intelligent cloud growth. She cited higher IDC costs (up to RMB 812 million) and a rise in depreciation and amortization to RMB 741 million due to newly acquired and leased servers and network equipment largely allocated to AI business.
Adjusted operating expenses (excluding share-based compensation) were RMB 459 million, up 3% year over year.
Business execution: AI clusters, industry expansion, and product platforms
Zhou said customer demand broadened beyond leading AI enterprises and internet companies to include automotive manufacturing, autonomous driving, embodied AI, and fintech, among others. He also described continued efforts to “push the limits of cluster scale” to support large-model training and “explosive” inference demand.
Operational examples included delivery of an inference cluster for a “top video streaming platform,” which management said has supported the customer’s AI business serving over 100 million users. The company also said it signed a major fintech customer for token-based inference services, citing customer recognition of stable model and computing power services.
In enterprise cloud, the company said revenue rose to RMB 859 million in the quarter, up 18% quarter over quarter. Zhou pointed to AI policy support and demand for verticalized models, real-world usability, and stricter data compliance requirements as factors making cloud services increasingly essential in industrial intelligence.
On products, management said it is building a next-generation computing service system spanning training, inference, and industrial intelligence, and shifting from traditional cloud computing toward an “AI-first, AI-native” architecture. Zhou said the StarFlow platform continued to upgrade, adding MCP (Model Context Protocol Hub), prompt optimization, and AI search features to support enterprise AI agent development and deployment. For private deployments, the company highlighted “Galaxy Stack” capabilities including heterogeneous GPU management, RoCE networking, and intelligent container scheduling.
Ecosystem relationships, pricing discipline, and 2026 investment plans
Management discussed both its Xiaomi/Kingsoft ecosystem ties and its non-ecosystem growth. Zhou said quarterly ecosystem revenue (Xiaomi and Kingsoft ecosystem) was RMB 804 million, up 63% year over year and representing 29% of total revenue. He added that for full-year 2025, related-party transactions with Xiaomi and Kingsoft ecosystem partners reached 94% of the company’s net annual cap. He also said external customers accounted for around 70% of total revenue, and revenue from the top five non-ecosystem customers rose 44% year over year.
In Q&A, Zhou said the company’s ecosystem AI strategy remains anchored on a “1+N” approach, where the “1” refers to Xiaomi’s model (MiMo), and “N” includes other models the company can host as a neutral cloud platform. He reiterated that the Kingsoft ecosystem does not plan to build its own large language model, while Xiaomi develops the model.
On pricing, management said it had anticipated supply-chain price increases beginning in the third quarter of the prior year and stocked some key components. The company outlined two pricing principles: existing contracts and existing resource commitments generally would not see price increases, while new contracts, new customers, and significant incremental usage could involve “significant” price hikes. Management said pricing changes may reflect both upstream cost pass-through and additional increases tied to demand and profitability objectives.
Addressing competitive pricing, Zhou said he did not believe a low-price strategy would be broadly implementable given explosive demand and resource tightness, adding that headline catalog price changes may not reflect actual transaction pricing.
Looking ahead, management discussed 2026 capital spending. Li said the company expects 2026 CapEx and tangible assets to exceed RMB 10 billion, expanding from 2025 levels, with roughly half targeted to be covered by customer prepayment arrangements. She said the company plans to access more assets through short- and long-term leasing to reduce upfront capital needs, and stated it currently has no equity financing plans. She outlined funding sources as proceeds from 2025 financing, customer operating receipts, strategic customer prepayments, and committed credit facilities.
Li also said management expects growth to accelerate and EBITDA to improve in 2026, and added that the mix of incoming demand is skewing toward inference: the company is seeing “more than half” of potential demand coming from inference versus training, with StarFlow’s MaaS-related business growing quickly and offering better margin, in management’s view.
About Kingsoft Cloud (NASDAQ:KC)
Kingsoft Cloud Holdings Limited (NASDAQ: KC) is a leading provider of cloud computing services in China, offering a comprehensive suite of infrastructure and platform solutions to enterprise customers. Established in 2012 as a subsidiary of Kingsoft Corporation, the company has grown into an independent public entity with dual listings, serving as a critical backbone for digital transformation across multiple industries. Headquartered in Beijing, Kingsoft Cloud leverages advanced technologies to optimize cloud operations and deliver scalable, reliable services.
The company’s core offerings span Infrastructure-as-a-Service (IaaS), Platform-as-a-Service (PaaS) and Software-as-a-Service (SaaS), encompassing compute, storage, database, content delivery networks (CDN) and security solutions.
