Aug 19, 2026 earnings call — AI cloud growth and profitability
Company reported record revenue of RMB 3.07 billion, up 31% year-over-year.
Revenue $0.5B (est $0.4B) · EPS $-0.03 (est $-0.07)
Key quotes
"The rapid growth of the open source model ecosystem is creating significant opportunities for neutral cloud providers."
"Operating profit turned positive for the first time with adjusted operating margin reaching a record high of 4.0%."
"Accordingly, we maintain our full-year CapEx base case unchanged at RMB 15 billion."
Analyst Q&A 5
MaaS impact and resource allocation
Q (Liping Zhao · CICC): The analyst asked how open-source model advances affect the company's MaaS business and which use cases drive growth. They also asked if the shorter payback for MaaS means the company will allocate more resources to it.
A: Management said open-source large language models have three main impacts: more demand from China users adopting GLM and K3, growth in agentic scenarios (they launched an agentic product), and routine tasks favor price-for-value Chinese models. They said computing power and MaaS use different business models: computing power sales have 100% utilization and long-term contracts, while MaaS is affected by token price, new models and operating efficiency. The company balances the two businesses and dynamically evaluates resource allocation.
"So we basically have different business models for these 2 business." — Tao Liu
Chip procurement and full-year capex
Q (Wenting Yu · CLSA): The analyst asked how chip procurement has progressed since June and for the latest full-year CapEx guidance. They also asked about the enterprise cloud revenue slowdown and AI transformation positioning.
A: Management said supply constraints are a long-term norm and they cope by increasing supplier and partner numbers and supporting Made in China chips, which are improving for model inference. They noted monthly CapEx is volatile because purchases are large lump sums, but their view for the year remains unchanged. They reported CapEx reached RMB 6.2 billion in H1 2026 and they maintain full-year CapEx base case at RMB 15 billion.
"Accordingly, we maintain our full-year CapEx base case unchanged at RMB 15 billion." — Yi Li (CFO)
MaaS positioning and pricing trends
Q (Timothy Zhao · Goldman Sachs): The analyst asked how Kingsoft Cloud's MaaS offering compares with peers on use cases and competitiveness, and about revenue recognition and profitability. They also asked about recent pricing trends and customer reactions after price increases.
A: Management said their MaaS positioning is neutral: they do not have proprietary models, so they encourage sales to offer models customers prefer, such as GLM. They said they raised prices for storage and computing: storage price hikes were broadly accepted by customers and could increase profitability, while computing power price increases could be passed through thanks to their PaaS and operating capabilities and sometimes raise profitability. They also said they are doing more managed services projects, which are asset-light.
"Mainly, we're different from some of the full-stack cloud providers, which they have their in-house proprietary models." — Unknown Executive
Long-term cloud positioning and margins
Q (Wei Xiong · UBS): The analyst asked how to view the company's long-term position in the cloud market given rivals' proprietary models and ecosystems, and what sustainable margin levels might be.
A: Management said being a neutral cloud player lets them partner with all top model providers and offer customers the best models while delivering stable, highly available services under SLAs. They emphasized working closely with LLM labs to optimize model inference efficiency, sometimes reaching or matching the inference efficiency of the LLM companies themselves. This collaboration supports service quality and profitability.
"So as mentioned, as a neutral cloud player, we are able to be in good relations with all of the top model providers, large language model labs and be able to provide the best model according to our customers' demands." — Unknown Executive
ROIC for computing vs MaaS
Q (Yang Liu · Morgan Stanley): The analyst asked for ROIC for computing power leasing versus MaaS and how ROIC is changing. They wanted the marginal change for ROIC.
A: Management said they do not disclose separate ROIC for MaaS and AI computing power because ROIC varies across projects driven by CapEx cycles, gross margin, fixed assets and depreciation policies. They stated overall MaaS delivers much better profitability than AI computing power services at this stage and that operating leverage is improving, trailing 12 months adjusted operating profit has turned positive, and overall ROIC should keep improving steadily.
"Overall, MaaS delivers much better profitability than AI computing power services at this stage." — Yi Li (CFO)