DigitalOcean (NYSE:DOCN) executives used the company’s fourth-quarter earnings call to outline a broader investor update centered on accelerating growth, expanding AI-related demand, and planned data center capacity additions that management said underpin a higher long-term outlook.
Q4 and full-year results show reacceleration and profitability
CEO Paddy Srinivasan said the company “had a fantastic quarter and a very strong finish to the year,” ending Q4 with 18% year-over-year revenue growth. CFO Matt Steinfort reported fourth-quarter revenue of $242 million, up 18% year-over-year, and full-year 2025 revenue of $901 million.
DigitalOcean reported adjusted EBITDA of $99 million in Q4 (41% margin) and $375 million for the year (42% margin). Trailing 12-month adjusted free cash flow was $168 million in Q4, which Steinfort described as 19% of revenue. Srinivasan cited full-year adjusted free cash flow margins of 19%.
Top customers highlighted as “growth engine”
Management repeatedly emphasized that scaling larger customers—previously described as a constraint—has become a driver of the business. Srinivasan said the company delivered record incremental organic ARR of $51 million in Q4 and $150 million on a trailing 12-month basis.
DigitalOcean’s Digital Native Enterprise (DNE) cohort—cloud- and AI-native companies—reached $604 million in ARR in Q4, which Srinivasan said is 62% of total ARR and grew 30% year-over-year. DNE net dollar retention (NDR) was 102% in Q4, which he said continued to outperform developer NDR.
Growth was strongest among the largest customers. Srinivasan said customers at $100,000 in ARR grew 58% year-over-year, $500,000 customers grew 97%, and $1 million customers grew 123%. He said $1 million customers reached $133 million in ARR and had 0% churn in Q4 and averaged 0% churn over the last 12 months. NDR increased with scale: 102% for $100,000 customers, 106% for $500,000, and 115% for $1 million customers.
AI and “Agentic Inference Cloud” strategy, plus new AI metric
Srinivasan framed the market opportunity around a shift from model training to inference and from “seats to tokens,” describing a future dominated by “agentic systems” that run continuously and require full-stack capabilities beyond GPU rentals. He positioned DigitalOcean’s “Agentic Inference Cloud” as an integrated platform combining inference infrastructure with core cloud services such as storage, databases, networking, observability, and security.
To provide more visibility, the company introduced a new metric: AI customer revenue, defined as all revenue from customers leveraging DigitalOcean AI products, including inference and core cloud services. Srinivasan said Q4 AI customer ARR reached $120 million, up 150% year-over-year and representing 12% of total ARR. He also noted that 70% of AI customer ARR in Q4 came from inference services or general-purpose cloud products rather than bare metal GPU rentals.
Management cited several examples of traction with AI-native customers, including Character.ai, Ocado, and Hippocratic AI. Srinivasan said DigitalOcean delivered a “100% throughput increase” and “roughly 50% lower cost per token” for Character.ai on production inference workloads using AMD Instinct GPUs. He said Hippocratic AI selected DigitalOcean to run HIPAA-compliant clinical AI workloads, highlighting “enterprise-grade security and compliance” and optimization on NVIDIA hardware.
Executives also discussed a recent open-source agent framework, OpenClaw, as an illustration of agentic workloads needing an integrated cloud. Srinivasan said that within days of launching OpenClaw, nearly 30,000 one-click OpenClaw GPU Droplets were created, alongside “thousands” of other deployments activated by customers.
Capacity expansion, financing approach, and updated outlook
Steinfort said DigitalOcean has expanded its “financial toolkit” to include equipment financing to better align infrastructure investment timing with the revenue it supports. He said the company will use a mix of upfront asset purchases and equipment leasing as it invests for growth.
He also highlighted stock-based compensation and repurchases. SBC declined to 9% of revenue in 2025 from 12% in the prior year, according to Steinfort. The company repurchased 2.4 million shares in 2025 for $82 million at an average price of about $35, and Steinfort said the full $100 million buyback authorization remains in place through July 31, 2027.
On the balance sheet, Steinfort said the company addressed the upcoming maturity of its 2026 convertible notes through establishing an $800 million bank facility, issuing $625 million of 2030 convertible notes, and repurchasing the majority of the then-outstanding 2026 notes. He said DigitalOcean ended 2025 with sufficient liquidity and projected cash generation to address the remaining $312 million balance of 2026 notes, and that it plans to repurchase or redeem the remainder for cash before or at maturity in December 2026.
As a portfolio action, Steinfort said DigitalOcean is sunsetting a “small legacy dedicated bare metal CPU offering” and expects approximately $13 million of ARR to roll off by the end of Q1 2026. He said this non-core revenue was excluded from customer-specific year-over-year growth metrics.
Looking ahead, management tied its growth outlook to data center capacity coming online. Steinfort said the company is bringing 31 megawatts (MW) of new data center capacity online across three facilities in 2026, with a 6 MW site beginning to ramp revenue in Q2 and two additional facilities ramping in the second half. He said costs such as lease expense and depreciation will hit financials “several months before” revenue begins in new facilities, pressuring gross margin and near-term profitability metrics.
Guidance provided on the call included:
- Q1 2026 revenue: $249 million to $250 million (about 18% to 19% year-over-year growth)
- Q1 2026 adjusted EBITDA margin: 36% to 37%
- Q1 2026 non-GAAP EPS: $0.22 to $0.27 (on ~111 million to 112 million fully diluted shares)
- Full-year 2026 revenue growth: 19% to 23% (21% at midpoint), or 21% to 24% excluding the legacy bare metal CPU roll-off
- Full-year 2026 adjusted EBITDA margin: 36% to 38%
- Full-year 2026 unlevered adjusted free cash flow margin: 18% to 20% (Steinfort cited $207 million at the midpoint)
- Full-year 2026 non-GAAP EPS: $0.75 to $1.00 (on ~111 million to 112 million fully diluted shares)
Management said it expects revenue growth to remain around 18% to 19% in Q2 2026, ramp in Q3, and exit 2026 at 25%+ growth in Q4. Steinfort and Srinivasan said the existing committed capacity alone supports a path to 30% revenue growth in 2027, with Steinfort adding that the 2027 profitability framework discussed was reflective of the 31 MW expansion already committed.
Q&A: open-source models, pricing, and metrics
In the Q&A, Srinivasan argued that open-source models will be increasingly important for AI-native companies managing unit economics, claiming open-source cost per token can be “about 90% cheaper” while maintaining comparable accuracy as models mature. He cited OpenRouter data indicating 30% of traffic is already served by open source.
Steinfort explained the company’s “weighted Rule of 50” approach—multiplying revenue growth by 1.5 and adding 0.5 times free cash flow margin—while also noting management expects to be a “regular Rule of 50” in 2027 based on 30% growth and 20% unlevered free cash flow margins.
On pricing, Srinivasan said pricing is “holding” and “in some cases it has gone up,” while noting it varies based on GPU generation and customer requirements. He also said customers can enter DigitalOcean’s stack at different layers, from GPU-hour pricing to token-based pricing higher up the stack.
Executives also addressed how AI affects traditional metrics. Steinfort said AI revenue is not yet included in NDR and suggested it may be “still 12 months away” from being rolled in, citing early AI customer lumpiness and the company’s intention to use AI customer revenue and RPO as leading indicators. He said RPO increased in Q4 to $134 million, up 121% sequentially and close to 500% year-over-year.
About DigitalOcean (NYSE:DOCN)
DigitalOcean Holdings, Inc is a cloud infrastructure provider that focuses on simplicity, performance and developer experience. The company offers a range of cloud services designed to help software developers, startups and small- to medium-sized businesses deploy, manage and scale applications. Its flagship offering, Droplets, provides virtual private servers that can be configured with various CPU, memory and storage options. In addition to compute instances, DigitalOcean’s platform includes managed Kubernetes, scalable object and block storage, managed databases, load balancers and networking capabilities such as Virtual Private Cloud (VPC) and Floating IPs.
Founded in 2011 and headquartered in New York City, DigitalOcean was created with the goal of making cloud computing more accessible to individual developers and smaller teams.
