KC — Kingsoft Cloud Holdings Limited

Kingsoft Cloud Q3 2025 Earnings Call Summary

1. Key Financial Results and Metrics

• Total Revenue: RMB 2,478 million, a year-over-year increase of 31%, up from 24% in the previous quarter.
• Public Cloud Revenue: RMB 1,752 million, a significant increase of 49% year-over-year.
• Enterprise Cloud Revenue: RMB 726 million, compared to RMB 110 million in the same quarter last year.
• Intelligent Computing Cloud Gross Billings: RMB 782 million, up 122% year-over-year, accounting for 45% of public cloud revenue.
• Adjusted Gross Profit: RMB 393 million, a 28% increase year-over-year.
• Adjusted Operating Profit: RMB 15.4 million, a turnaround from a loss of RMB 140 million in the same period last year.
• Adjusted Net Profit: RMB 28.73 million, marking the first positive net profit in the company’s history.
• Adjusted EBITDA Margin: 33%, up from 10% in the same quarter last year.

2. Strategic Updates and Business Highlights

• Kingsoft Cloud is focusing on integrating artificial intelligence into its cloud services, particularly in model training and inference.
• The company has launched a model API service and upgraded its online model services, enhancing its technological capabilities.
• Revenue from the Xiaomi and Kingsoft ecosystem reached RMB 691 million, up 84% year-over-year, contributing to 28% of total revenue.
• The company is investing in infrastructure to support its intelligent cloud business and has expanded its workforce significantly, particularly in its Wuhan R&D center.

3. Forward Guidance and Outlook

• Management expressed confidence in continued revenue growth, particularly driven by increased demand for AI capabilities across various industries.
• The budget process for 2026 is underway, with specific revenue guidance to be shared once finalized.
• The company anticipates that the shift from training to inference demand will lead to higher profit margins in the future.

4. Bad News, Challenges, or Points of Concern

• The significant quarter-on-quarter improvement in margins was partly driven by one-time other income, which may not be sustainable in future quarters.
• There is a risk associated with the reliance on large customers for training demand, as the market transitions to a broader base of inference demand.
• Increased costs due to infrastructure investments and rising depreciation expenses could pressure margins if not managed effectively.

5. Notable Q&A Insights

• Management noted a structural shift in demand from large customers' training needs to a wider adoption of inference services.
• The pricing strategy for AI training and inference is based on resource usage, with inference through API services expected to yield better margins.
• The company is strategically using a mix of procurement and leasing models for computing resources, depending on customer needs and risk profiles, with no rigid allocation targets in place.

Overall, Kingsoft Cloud demonstrated strong financial performance in Q3 2025, driven by robust growth in its public and intelligent cloud services, while also laying the groundwork for future expansion in the AI sector. However, management acknowledged potential challenges related to margin sustainability and customer demand dynamics.