DigitalOcean's Q3 2025: Contradictions Emerge on AI Revenue Predictability, Net Dollar Retention, and CAPEX Strategy

Business Commentary:

* Revenue Growth and Strategic Investments: - DigitalOcean reported revenue of $230 million for Q3, up 16% year-over-year. - Growth was driven by increasing traction with AI-native customers and strategic investments in data centers and GPU capacity to support growing demand.

• Customer Growth and Retention:
• Revenue from customers with $100,000 or more in annual run rate grew 41% year-over-year, to 26% of overall revenue.

• This growth is attributed to the company's ability to attract, retain, and grow its largest customers through product innovation and a focused direct sales motion.

• AI Platform Adoption:

• AI revenue more than doubled year-over-year for the fifth consecutive quarter, contributing significantly to the company's growth.

• The adoption of AI is driven by DigitalOcean's comprehensive AI infrastructure and platform capabilities, supporting scaling AI and digital native enterprise customers.

  

• Expansion of Data Center Capacity:

• DigitalOcean announced securing 30 megawatts of incremental data center capacity to support growth in 2026 and beyond.
• This expansion is in response to increased demand from AI-native customers and the need to maintain attractive margins while scaling operations.
  

Contradiction Point 1

AI Revenue Growth and Predictability

It involves changes in financial forecasts, specifically regarding AI revenue growth and predictability, which are critical indicators for investors.

Can you provide details about the 8-figure contracts signed after Q3, which customers are involved, and how they relate to AI revenue? - Gabriela Borges(Goldman Sachs Group)

2025Q3: Our AI infrastructure capability has grown significantly. We now have over 1,000 AI nodes in Amsterdam, New York, and Singapore, and we plan to double that in 2026. AI workloads are predominantly inferencing, providing predictable growth. - Padmanabhan Srinivasan(CEO)

Can you break down the components driving the over 100% AI/ML revenue growth and recent ARR trends? - Michael Joseph Cikos(Needham & Company)

2025Q2: The AI/ML revenue grew more than 100% year-over-year, driven by the introduction of NVIDIA H100 GPUs last year. Our strategy focuses on customer acquisition, AI-infrastructure scaling, and AI-platform adoption. - W. Matthew Steinfort(CFO)

Contradiction Point 2

Impact of AI on Net Dollar Retention

It involves the impact of AI revenue on net dollar retention, which is a critical metric for understanding customer retention and growth.

Should AI be included in net dollar retention as predictability increases? What is the NDR outlook? - Matthew Calitri(Needham & Company)

2025Q3: We'll look to incorporate AI revenue growth into our metrics in the future. The AI workloads are behaving more predictably, and we'll likely revisit this in early 2026. The focus remains on attracting and growing larger customers. - Matt Steinfort(CFO)

Does the net dollar retention rate include AI revenue? How will large deals affect future performance? - Joshua Phillip Baer(Morgan Stanley)

2025Q2: AI revenue isn't included in NDR yet. We're conservative in forecasting large deals, booking revenue as it's earned, and we'll be cautious in reflecting projected revenue from large deals until we're confident customers are growing. - W. Matthew Steinfort(CFO)

Contradiction Point 3

AI Revenue and Customer Traction

It involves differing perspectives on the predictability and growth trajectory of AI revenues, which are crucial for understanding the company's strategic focus and financial outlook.

What details can you share about the 8-figure contracts signed post-Q3, which customers are involved, and their connection to AI revenue? - Gabriela Borges (Goldman Sachs Group)

2025Q3: AI workloads are predominantly inferencing, providing predictable growth, and the company is expanding data center capacity to accommodate this growth. - Padmanabhan Srinivasan(CEO)

Can you discuss the $20 million multiyear deal and how customer conversations are adapting to AI evolution? - Gabriela Borges (Goldman Sachs)

2025Q1: I think the most important thing is just there are a lot of things that are coming together to enable AI to be a much more predictable revenue stream. - Paddy Srinivasan(CEO)

Contradiction Point 4

CAPEX Investment Strategy

It addresses the company's strategy for capital expenditures, which are critical for growth and financial health, revealing differing approaches and priorities.

Can DigitalOcean immediately fulfill the 8-figure contracts, or is the capacity expansion for 2026? What is the ramp-up timeline for these contracts? - Josh Baer (Morgan Stanley)

2025Q3: We have some capacity in existing data centers. New data center capacity will come online progressively through 2026, with most coming online in the first half. Our guidance includes paying for new data center capacity, and we expect a smooth revenue ramp. - Padmanabhan Srinivasan(CEO)

How are you planning CapEx for the year and addressing new capacity needs? - James Fish (Piper Sandler)

2025Q1: We've already front-loaded Q1 with Atlanta data center capacity. This capacity supports our growth projections. We're open to leasing some incremental gear if it accelerates growth beyond our current outlook, maintaining strong free cash flow margins. - Matt Steinfort(CFO)

Contradiction Point 5

AI Revenue and Customer Base

It highlights a change in the composition and predictability of AI revenue, which is crucial for investor projections and strategic planning.

Can you provide details on the multiple eight-figure committed contracts signed after Q3? What are the customer types and their relation to AI revenue for these contracts? - Gabriela Borges (Goldman Sachs Group)

2025Q3: We've always said that our Gen AI revenue is structured as a pull-through revenue into our core cloud services. The AI workloads are predominantly inferencing, providing predictable growth. - Padmanabhan Srinivasan(CEO)

What is the current AI/ML ARR base and its composition? How does it compare with the scalers cohort? - Michael Cikos (Needam & Company)

2024Q4: AI ARR is intermingled with other services, making precise disclosure unfeasible. AI is adopted by a mix of existing and new customers, with a significant chunk from the scalers cohort, but not entirely within it. - Matt Steinfort(CFO)