EPS · actual vs est
Revenue · actual vs est
Summary
Generated 2026-11-16
Management highlights
- AI Cloud growth and economics
- AI Cloud Infra was the principal driver of General Business returning to growth. GPU Cloud demand remained strong across training and inference, with inference expected to expand as AI applications and agents are deployed more broadly.
- The customer base broadened across industries and company sizes; customers are expanding workloads from GPU compute into storage, networking and other cloud services. Longer contracts and prepayments are improving demand visibility and cash collection.
- Higher-margin product mix, better capacity utilization and full-stack optimization improved segment profitability and shortened infrastructure investment payback periods.
- External token usage on Qianfan MaaS grew more than 12x year over year. Inference optimization improved throughput, latency and utilization; management expects MaaS to become a more meaningful profit contributor as it scales.
- Models, applications and search
- ERNIE 5.5 Flash Preview was introduced with 280 billion total parameters and 14 billion activated per token; Baidu plans faster model iteration, guided by application needs rather than broad benchmark chasing.
- DuMate MAU rose more than 30-fold quarter over quarter in September, without large-scale marketing or super-app distribution. The company also expanded specialized agents including KuKu, Yijing and Famou; monetization remains early, with product value and retention the immediate focus.
- ERNIE Assistant’s daily average conversation rounds grew 126% year over year in September. Baidu is prioritizing AI search user experience over near-term monetization, accepting near-term revenue and profit pressure.
- Other operations and capital allocation
- Yijing production cost has fallen approximately 80% since Q4 2025, while production time has dropped by more than an order of magnitude; customer adoption expanded across industries and internationally.
- Apollo Go is concentrating on priority markets. Fully driverless rides in Shenzhen increased approximately 170% quarter over quarter; its Dubai fleet doubled. Safety remains the top priority.
- Operating cash flow was RMB 2.8 billion, the fifth consecutive positive quarter. Total cash and investments were RMB 276.6 billion. Management emphasized disciplined investment, higher utilization and balancing AI growth spending with shareholder returns.
Segment performance
- Baidu General Business: RMB 25.0 billion in revenue, up 1% year over year and down 1% quarter over quarter. AI Cloud Infra growth offset pressure in advertising. Baidu Core AI-powered Business generated RMB 13.2 billion, or 53% of General Business revenue; management did not provide a separate growth rate for this total.
- AI Cloud Infra: Revenue rose 94% year over year. GPU Cloud revenue grew 191% year over year and approximately 43% quarter over quarter. GPU Cloud and MaaS together represented more than half of AI Cloud Infra revenue; higher-margin offerings including GPU Cloud, Kunlunxin and MaaS collectively made up a majority. AI Cloud Infra profit and margin improved year over year and sequentially, but no absolute segment revenue or profit was disclosed.
- Baidu overall: Revenue was RMB 30.8 billion, down 1% year over year and 2% quarter over quarter. Operating income was RMB 626 million (2% margin); non-GAAP operating income was RMB 1.7 billion (6% margin). Net income attributable to Baidu was RMB 1.4 billion; non-GAAP net income was RMB 2.4 billion.
- iQIYI: Revenue was RMB 6.0 billion, down 10% year over year and 4% quarter over quarter. This was approximately 19.5% of Baidu’s reported total revenue (using reported segment figures, which may not sum exactly due to rounding).
Management Statement and Operational Highlights
- AI Cloud growth and economics
- AI Cloud Infra was the principal driver of General Business returning to growth. GPU Cloud demand remained strong across training and inference, with inference expected to expand as AI applications and agents are deployed more broadly.
- The customer base broadened across industries and company sizes; customers are expanding workloads from GPU compute into storage, networking and other cloud services. Longer contracts and prepayments are improving demand visibility and cash collection.
- Higher-margin product mix, better capacity utilization and full-stack optimization improved segment profitability and shortened infrastructure investment payback periods.
- External token usage on Qianfan MaaS grew more than 12x year over year. Inference optimization improved throughput, latency and utilization; management expects MaaS to become a more meaningful profit contributor as it scales.
- Models, applications and search
- ERNIE 5.5 Flash Preview was introduced with 280 billion total parameters and 14 billion activated per token; Baidu plans faster model iteration, guided by application needs rather than broad benchmark chasing.
- DuMate MAU rose more than 30-fold quarter over quarter in September, without large-scale marketing or super-app distribution. The company also expanded specialized agents including KuKu, Yijing and Famou; monetization remains early, with product value and retention the immediate focus.
- ERNIE Assistant’s daily average conversation rounds grew 126% year over year in September. Baidu is prioritizing AI search user experience over near-term monetization, accepting near-term revenue and profit pressure.
- Other operations and capital allocation
- Yijing production cost has fallen approximately 80% since Q4 2025, while production time has dropped by more than an order of magnitude; customer adoption expanded across industries and internationally.
- Apollo Go is concentrating on priority markets. Fully driverless rides in Shenzhen increased approximately 170% quarter over quarter; its Dubai fleet doubled. Safety remains the top priority.
- Operating cash flow was RMB 2.8 billion, the fifth consecutive positive quarter. Total cash and investments were RMB 276.6 billion. Management emphasized disciplined investment, higher utilization and balancing AI growth spending with shareholder returns.
Guidance
- Management expects AI Cloud Infra to sustain strong growth in coming quarters, supported by GPU Cloud demand, expanding inference and agents, broader customer relationships, recurring revenue, longer contracts and prepayments. No numerical growth target was provided.
- AI Cloud Infra margins and profitability are expected to improve over time through a richer mix, scale and utilization, full-stack cost optimization, and higher-value services. Management said segment margins improved in Q3 both year over year and sequentially.
- Baidu expects continued near-term pressure on advertising and overall profit from macro conditions, competition, AI search transformation and investment in cloud, chips and models. Management expects AI Cloud growth to increasingly offset advertising weakness and support medium- to long-term revenue and profit improvement.
- Q4 and next year’s compute investment will draw on existing capacity and be paced to customer demand and cash collection. No overall CapEx target was disclosed; Q3 CapEx was RMB 4 billion lower than Q2.
- Baidu expects to declare its first dividend before year-end, subject to Board approval and market conditions, and intends to continue share repurchases at an appropriate scale. Year-to-date repurchases were approximately $373 million.
- Kunlunxin’s M100 is expected to reach volume production soon; the M300 is planned for next year. Baidu continues to advance Kunlunxin’s Hong Kong listing process and will update the market as appropriate.
Risks
- Advertising remains pressured by macroeconomic conditions, intense competition and the AI search transition; prioritizing user experience delays monetization and weighs on near-term revenue and profitability.
- Continued spending on AI cloud infrastructure, chips and foundation models may constrain near-term profit improvement. Infrastructure returns depend on demand, utilization, customer payment terms and cash collection; new capacity must be paced carefully.
- Cloud growth entails execution demands around reliable, scalable capacity and efficient resource scheduling. Longer-term customer commitments improve visibility but do not eliminate demand or delivery risks.
- Apollo Go’s expansion across markets carries operational and safety considerations; management explicitly said safety remains its top priority. The call did not report a specific safety incident or operational failure.
- Forward-looking statements are subject to risks and uncertainties that could cause actual results to differ materially; Baidu cited its regulatory filings for fuller discussion.
Question and Answer
Q: What supports AI Cloud Infra’s growth and how durable is it? / A: Management cited accelerating training and inference demand, wider customer adoption, expansion into adjacent cloud services, and more recurring revenue. Longer contracts and prepayments improve visibility and help Baidu plan capacity; GPU Cloud and MaaS together exceeded half of segment revenue.
Q: How will ERNIE investment generate returns, and what is its strategic position? / A: Baidu will continue disciplined, application-driven investment, focusing on capabilities most valuable to its products. Better in-house models should improve user experience and reduce costs in Baidu applications, while Qianfan provides external monetization and diversifies cloud revenue.
Q: What drives AI Cloud margins, and how does Baidu assess infrastructure returns? / A: Management expects improved mix, utilization, full-stack optimization and higher-value services to support margin expansion. Contracts are assessed for revenue, margins, payment terms and payback; prepayments and faster recovery of upfront costs can fund further growth.
Q: How will AI agent competition, monetization and DuMate’s prospects evolve? / A: Baidu sees application value—reliable task completion and measurable productivity—as the basis for long-term competitiveness. DuMate’s rapid, largely organic user growth is encouraging, but monetization is still nascent; the immediate priority is building products users return to, with meaningful longer-term revenue potential.
Q: How will Baidu balance AI investment, CapEx and shareholder returns? / A: It will use existing compute and raise utilization, then pace new investment to demand and customer cash collection. Management reiterated plans for a first dividend before year-end subject to approvals and conditions, alongside continued repurchases under the existing authorization.
Guidance
- Management expects AI Cloud Infra to sustain strong growth in coming quarters, supported by GPU Cloud demand, expanding inference and agents, broader customer relationships, recurring revenue, longer contracts and prepayments. No numerical growth target was provided.
- AI Cloud Infra margins and profitability are expected to improve over time through a richer mix, scale and utilization, full-stack cost optimization, and higher-value services. Management said segment margins improved in Q3 both year over year and sequentially.
- Baidu expects continued near-term pressure on advertising and overall profit from macro conditions, competition, AI search transformation and investment in cloud, chips and models. Management expects AI Cloud growth to increasingly offset advertising weakness and support medium- to long-term revenue and profit improvement.
- Q4 and next year’s compute investment will draw on existing capacity and be paced to customer demand and cash collection. No overall CapEx target was disclosed; Q3 CapEx was RMB 4 billion lower than Q2.
- Baidu expects to declare its first dividend before year-end, subject to Board approval and market conditions, and intends to continue share repurchases at an appropriate scale. Year-to-date repurchases were approximately $373 million.
- Kunlunxin’s M100 is expected to reach volume production soon; the M300 is planned for next year. Baidu continues to advance Kunlunxin’s Hong Kong listing process and will update the market as appropriate.
Risks
- Advertising remains pressured by macroeconomic conditions, intense competition and the AI search transition; prioritizing user experience delays monetization and weighs on near-term revenue and profitability.
- Continued spending on AI cloud infrastructure, chips and foundation models may constrain near-term profit improvement. Infrastructure returns depend on demand, utilization, customer payment terms and cash collection; new capacity must be paced carefully.
- Cloud growth entails execution demands around reliable, scalable capacity and efficient resource scheduling. Longer-term customer commitments improve visibility but do not eliminate demand or delivery risks.
- Apollo Go’s expansion across markets carries operational and safety considerations; management explicitly said safety remains its top priority. The call did not report a specific safety incident or operational failure.
- Forward-looking statements are subject to risks and uncertainties that could cause actual results to differ materially; Baidu cited its regulatory filings for fuller discussion.
Q&A highlights
Q: What supports AI Cloud Infra’s growth and how durable is it? / A: Management cited accelerating training and inference demand, wider customer adoption, expansion into adjacent cloud services, and more recurring revenue. Longer contracts and prepayments improve visibility and help Baidu plan capacity; GPU Cloud and MaaS together exceeded half of segment revenue.
Q: How will ERNIE investment generate returns, and what is its strategic position? / A: Baidu will continue disciplined, application-driven investment, focusing on capabilities most valuable to its products. Better in-house models should improve user experience and reduce costs in Baidu applications, while Qianfan provides external monetization and diversifies cloud revenue.
Q: What drives AI Cloud margins, and how does Baidu assess infrastructure returns? / A: Management expects improved mix, utilization, full-stack optimization and higher-value services to support margin expansion. Contracts are assessed for revenue, margins, payment terms and payback; prepayments and faster recovery of upfront costs can fund further growth.
Q: How will AI agent competition, monetization and DuMate’s prospects evolve? / A: Baidu sees application value—reliable task completion and measurable productivity—as the basis for long-term competitiveness. DuMate’s rapid, largely organic user growth is encouraging, but monetization is still nascent; the immediate priority is building products users return to, with meaningful longer-term revenue potential.
Q: How will Baidu balance AI investment, CapEx and shareholder returns? / A: It will use existing compute and raise utilization, then pace new investment to demand and customer cash collection. Management reiterated plans for a first dividend before year-end subject to approvals and conditions, alongside continued repurchases under the existing authorization.
Key numbers
Reported versus consensus
Earnings calendar feed
| Metric | Reported | Consensus | Delta | Prior year |
|---|---|---|---|---|
| EPS | $1.01 | $0.85 | +18.8% | $1.56 |
| Revenue | $4.59B | $4.60B | -0.3% | $4.38B |
Transcript
November 16, 2026Full transcript unavailable for redistribution
The structured summary above covers the available call sections. Full transcript text is not included on this page.
Continue exploring
Prior quarters
This page presents the stored structured earnings-call summary and deterministic earnings calendar values. How this is generated. For informational purposes only; not investment advice.