Confluent's Q3 2025 Earnings Call: Contradictions Emerge on Sales Execution, AI Impact, Flink Momentum, and DSP/AI Growth

Date of Call: October 27, 2025

Financials Results

• Revenue: $286.3M subscription revenue, up 19% YOY (represented 96% of total revenue)
• EPS: $0.13 per diluted share (non-GAAP net income per diluted share for Q3)
• Gross Margin: Subscription gross margin 81.8%, above long-term target threshold of 80%
• Operating Margin: Non-GAAP operating margin 9.7%, up 340 basis points YOY (CEO referred to ~10%, an expansion of ~3 percentage points)

Guidance:

• Q4 FY2025 subscription revenue expected $295.5M–$296.5M (~18% growth)
• Q4 non-GAAP operating margin ~7%; Q4 non-GAAP net income per diluted share $0.09–$0.10
• FY2025 subscription revenue expected $1.1135B–$1.1145B (~21% growth); FY non-GAAP operating margin ~7%; FY non-GAAP net income per diluted share $0.39–$0.40; adjusted FCF margin ~6%
• Q4 cloud revenue modeling point ~$165M (~20% growth), ~56% of subscription revenue (midpoint)

Business Commentary:

* Strong Cloud Performance and Consumption Growth: - Confluent's cloud revenue grew 24% to $161 million, representing 56% of subscription revenue. - The growth was driven by higher consumption across core streaming and DSP, particularly with Flink ARR growing more than 70% sequentially.

• Field Alignment and Pipeline Progression:
• The late-stage pipeline progression showed more than 40% sequential growth, reflecting a significant increase in new use cases moving into production.

• This was attributed to improved field alignment and effective execution in driving new use cases into production.

• Increased Adoption of Flink and DSP:

• Over 1,000 customers used Flink, with more than 12 customers spending over $100,000 in ARR and 4 customers with over $1 million in ARR.

• This growth was supported by the build-out of the DSP specialist team and accelerating adoption of Flink across the customer base.

• Stability in Net Retention Rate and RPO Growth:

• Confluent's net retention rate stabilized at 114%, and RPO grew by 43%, reflecting improved visibility into near-term consumption revenue.
• This stability was driven by stronger consumption growth, particularly in cloud and DSP products.

Sentiment Analysis:

Overall Tone: Positive

• Management said they "delivered a strong Q3, exceeding the high end of all guided metrics." Subscription revenue grew 19% to $286.3M; Confluent Cloud grew 24% to $161M; operating margin rose 340 bps to 9.7%. They raised Q4 guidance and highlighted multi-quarter acceleration in late-stage pipeline and Flink consumption growth.

Q&A:

• Question from Brad Zelnick (Deutsche Bank AG): What are the learnings from go-to-market changes (field alignment, coverage ratios) another quarter in, conversion trends on the new pipeline, and capacity to work incremental pipeline?
  Response: The specialization model and tightened field execution are working: late-stage pipeline conversion into production is accelerating (40% sequential growth) and management has high confidence these workloads will drive consumption.

• Question from Brad Zelnick (Deutsche Bank AG): Should we view RPO/CRPO as a reliable leading indicator for Confluent?
  Response: For Confluent Platform, RPO is the single best leading indicator; for Confluent Cloud, short-term focus is new use cases moving to production while long-term RPO-to-cloud-revenue coverage is increasing.

• Question from Sanjit Singh (Morgan Stanley): When do you think growth will start to bottom?
  Response: Management is pleased with current results and raised Q4 guide; cloud growth is stabilizing (normalize for one large customer) and positive trends (Flink, field execution) support continued stability.

• Question from Sanjit Singh (Morgan Stanley): Are new cost-effective offerings (WarpStream, multi-tenant clusters) net accretive or cannibalistic to core streaming?
  Response: They are net accretive — customers are adopting larger workloads and deal sizes have increased, while these offerings are cost-effective and are a tailwind to gross margin.

• Question from Mark Murphy (JPMorgan Chase & Co): What is driving >40% sequential growth in progressing late-stage pipeline and what's a normal level?
  Response: The metric measures dollar value of new use cases moving to production; management uses it internally as a forward indicator — more dollars entering production lead to future consumption ramps, though there's no public historical baseline provided.

• Question from Mark Murphy (JPMorgan Chase & Co): How is early customer response to streaming agents and can agents become material over time?
  Response: Streaming agents are gaining traction; they simplify moving models from batch to real-time and are expected to be a critical, scalable part of the stack as customers iterate models with real-time data.

• Question from Raimo Lenschow (Barclays Bank PLC): What's your take on Flink adoption, how customers use it and pipeline visibility?
  Response: Flink adoption has ramped strongly since GA; it's a serverless, enterprise-capable stream-processing layer addressing large batch-to-stream migrations and AI/real-time workloads — management sees substantial long-term market opportunity.

• Question from Raimo Lenschow (Barclays Bank PLC): What drove the Q3 beat and higher Q4 subscription guide?
  Response: Drivers were accelerated new use cases moving to production, normalized (healthy) optimization levels, and strong Flink/cloud performance; these gave confidence for the raised Q4 cloud guide.

• Question from Ryan MacWilliams (Wells Fargo Securities): Which AI use cases are likely to move to production near term involving Confluent?
  Response: Common near-term production use cases include customer support, anomaly detection/investigations, IoT/operational monitoring and financial services workflows — success depends on data flow, quality and operational reliability.

• Question from Ryan MacWilliams (Wells Fargo Securities): How much did the large AI-native customer contribute in Q3 and what's the Q4 cloud impact as they move to self-managed?
  Response: That customer materially contributed in Q3; their move from cloud to on‑prem is expected to reduce cloud revenue by a low single-digit percent in Q4 and is baked into guidance.

• Question from Robbie Owens (Piper Sandler & Co): How big is the CSP replacement opportunity and why has it inflected recently?
  Response: It's sizable; drivers are improved TCO from Kora-enabled Enterprise/Freight clusters, broader DSP capabilities, and improved migration tooling making large migrations easier and more compelling.

• Question from Robbie Owens (Piper Sandler & Co): Are current optimization levels coming from prior optimizers or new ones — is optimization still a headwind?
  Response: Optimization is present each quarter but is now at normalized, healthy levels versus earlier larger headwinds; cloud momentum is driven by existing customer growth, net new use cases moving to production and new product adoption.

• Question from Jason Ader (William Blair): How much of Q3 improvement is sales execution versus macro/AI tailwinds?
  Response: Both — structural go‑to‑market improvements and new products materially contributed, and macro/AI demand also helps; management cannot precisely apportion but believes execution drove meaningful improvement.

• Question from Jason Ader (William Blair): Did you bake federal shutdown conservatism into Q4 outlook given potential U.S. Federal weakness?
  Response: Q3 federal was in line with expectations; federal revenue is low single digits of total and a couple of federal deals are baked into Q4 guidance.

• Question from Aleksandr Zukin (Wolfe Research): Of the 21 AI-native customers >$100k ARR, common patterns — Kafka or Flink usage?
  Response: AI customers use Confluent for real‑time data flows, feedback/evaluation loops and context for models; both Kafka and Flink are used depending on needs, with streaming central to ongoing evaluation and iteration.

• Question from Aleksandr Zukin (Wolfe Research): How should we think about cloud exit rate modeling and fiscal '26 assumptions?
  Response: Management is finalizing fiscal '26 planning and won't provide FY26 guidance now; note the low single-digit Q4 cloud impact from the large customer and that drivers (late‑stage pipeline, Flink, normalized optimization) inform the outlook.

• Question from Michael Cikos (Needham & Company): How did month-over-month consumption trends play out in Q3?
  Response: Month-over-month consumption growth improved sequentially in Q3; management will avoid granular month-by-month disclosures going forward.

• Question from Michael Cikos (Needham & Company): Is the DSP specialist team built out and are playbooks mature?
  Response: Yes — the DSP specialist team is built out and in full execution mode.

• Question from Howard Ma (Guggenheim Securities): Does Q4 cloud guide assume ongoing optimizations and will NRR decelerate from 114%?
  Response: Optimization is included but is normalized; Q4 guidance, after normalizing the large customer, is roughly flattish vs Q3 growth; NRR stabilized at 114% in Q3 and management expects core streaming, DSP, AI and partners to support long‑term retention, but they are not guiding NRR.

• Question from Howard Ma (Guggenheim Securities): Given Flink disclosures, should we model a sustained multi‑million sequential uplift to cloud from Flink?
  Response: Management declined to provide numeric cadence guidance but emphasized Flink's breadth (1,000+ paying customers) and depth (multiple six- and seven‑figure customers) and said Flink will be a material contributor in fiscal '26.

• Question from Eric Heath (KeyBanc): What easy wins/learnings scale Flink adoption and thoughts on Databricks structured streaming competition?
  Response: Flink wins split between real‑time pipelines to analytics/AI and acting-on-data (fraud, personalization); many customers convert batch jobs to streaming. Databricks overlaps but is largely complementary — different user constituencies and many joint customers; Confluent focuses on real‑time integration and operational apps.

Contradiction Point 1

Sales Execution and Field Alignment

It highlights inconsistencies in the company's approach and effectiveness in aligning field efforts and executing sales strategies, which could impact growth and customer relationships.

What were the key learnings from last quarter's go-to-market changes, especially field alignment and its impact on late-stage pipeline progression? - Brad Zelnick (Deutsche Bank AG, Research Division)

2025Q3: We've made significant improvements in our go-to-market strategy, including field alignment and product specialization. The late-stage pipeline progression metric, indicating dollars of new use cases moving into production, has shown more than 40% sequential growth. - Edward Kreps(CEO)

How does Ryan Mac Ban's focus on coverage ratios and DSP specialist teams address optimization issues for large customers? Is this expected to make a significant difference? - Brad Alan Zelnick (Deutsche Bank)

2025Q2: The sales team is working closely together, and we've made some changes to the sales structure and the sales strategy to try to better align around this. And the early results are showing a 40% increase in late-stage pipeline progression. - Jay Kreps(CEO)

Contradiction Point 2

AI Impact on Subscription Growth

It concerns the company's stance on whether AI workloads directly impact subscription growth, which is crucial for investor expectations and strategic planning.

When do you expect growth to bottom out? - Sanjit Singh (Morgan Stanley, Research Division)

2025Q3: The DSP and AI areas are seeing momentum, and the focus on execution within the field team is positive. - Edward Kreps(CEO)

What trends in consumption optimization are driving growth, and how significant are macro or company-specific factors? How will growth in production AI workloads impact subscription growth? - Matthew George Hedberg (RBC Capital Markets)

2025Q2: The AI workload growth doesn't directly impact subscription growth yet, but it's a promising trend. - Jay Kreps(CEO)

Contradiction Point 3

Flink Performance and Momentum

It shows differing perspectives on the performance and growth trajectory of the Flink product, which could influence investor expectations and strategic decision-making.

How has increased Flink adoption affected the business? - Raimo Lenschow (Barclays Bank PLC, Research Division)

2025Q3: We're excited about Flink's growth, now with over 1,000 customers, and more than 70% sequential growth in Q3. Flink ARR for Confluent Cloud surpassed 70%. - Edward Kreps(CEO)

How has switching sales reps to incremental consumption compensation affected team frictions, and how is Flink building momentum? - Sanjit Kumar Singh (Morgan Stanley)

2025Q2: Flink has built momentum since its introduction mid-last year, with steady ramp-up in cloud and on-prem. It provides a serverless solution that grows with customer queries, driving consistent growth. - Jay Kreps(CEO)

Contradiction Point 4

Sales Execution and Field Performance

It involves differing perspectives on the effectiveness and impact of sales execution and field performance, which directly affects company growth and customer acquisition.

Can you discuss the learnings from last quarter's go-to-market changes, specifically field alignment and their impact on late-stage pipeline progression? - Brad Zelnick(Deutsche Bank AG, Research Division)

2025Q3: We've made significant improvements in our go-to-market strategy, including field alignment and product specialization. The late-stage pipeline progression metric, indicating dollars of new use cases moving into production, has shown more than 40% sequential growth. These improvements are attributed to better field execution, helping use cases move to production more effectively. - Edward Kreps(CEO)

What drove the strong net adds in Q1? - Jason Ader(William Blair)

2025Q1: We saw this slowdown in March, with some stability in April, but no immediate rebound. This cautious approach impacts our cloud guidance. - Jay Kreps(CEO)

Contradiction Point 5

DSP and AI Product Growth

It highlights differing perspectives on the growth trajectory and adoption of DSP and AI-related products, which are critical for the company's future success.

When do you expect growth to bottom out? - Sanjit Singh(Morgan Stanley)

2025Q3: The DSP and AI areas are seeing momentum, and the focus on execution within the field team is positive. - Edward Kreps(CEO)

What are current GenAI use cases in streaming and DSP technology? How do you expect this to evolve over the next year? - Michael Turrin(Wells Fargo)

2024Q4: DSP consumption grew substantially faster than the overall cloud business. Key components are in early stages of growth with potential for continued growth impact. - Rohan Sivaram(CFO)