DigitalOcean Holdings, Inc. (DOCN)

Executive Summary / Key Takeaways

• DigitalOcean is executing a two-pronged transformation: AI revenue has doubled year-over-year for five consecutive quarters while core cloud Net Dollar Retention returned to 100% for the first time since June 2023, suggesting both new growth engines and existing business health are accelerating simultaneously.

• The company is successfully migrating upmarket, with Scalers+ customers (>$100K ARR) growing 41% year-over-year to represent 26% of total revenue, and customers exceeding $1M ARR growing 72%, demonstrating that simplicity and approachability resonate with larger enterprises, not just individual developers.

• Capacity constraints have become the primary near-term risk, with management acknowledging "more demand than we can support" and signing multiple 8-figure committed contracts after Q3 2025, creating an execution test for 2026 growth targets that now exceed prior 2027 projections.

• The Gradient AI Agentic Cloud platform has evolved from experimental to essential, with over 19,000 agents created and more than 7,000 in production, while AI-native customers increasingly require the full-stack capabilities beyond raw GPU access that DigitalOcean is uniquely positioned to provide.

• Balance sheet optimization through a $625M 2030 convertible note issuance and $800M secured credit facility provides financial flexibility, but the company must prove it can scale data center capacity—30 megawatts secured for 2026 and beyond—fast enough to capture demand that is already outpacing supply.

Setting the Scene: The Developer Cloud Grows Up

DigitalOcean Holdings, incorporated in 2012, built its foundation as the anti-hyperscaler: a cloud platform that prioritized simplicity, predictable pricing, and developer-friendly workflows over the sprawling complexity of AWS (AMZN), Azure (MSFT), and GCP (GOOGL). For years, this positioning carved out a durable niche serving digital-native enterprises, startups, and individual developers who valued approachability over enterprise feature breadth. The company generated revenue primarily through Infrastructure-as-a-Service (IaaS) offerings—Droplet virtual machines, storage, networking—and expanded into Platform-as-a-Service (PaaS) with Managed Databases, Kubernetes, and Marketplace offerings.

The cloud infrastructure market has long been dominated by three players controlling roughly two-thirds of global spend, leaving smaller providers to fight over the remainder. DigitalOcean's strategy was never to out-feature the hyperscalers, but to out-serve the underserved: companies spending $50 to $8,333 per month that needed reliable infrastructure without the overhead of enterprise sales cycles, complex pricing calculators, or certification requirements that measured in months rather than minutes.

This positioning created a profitable but modest business, generating $780 million in trailing twelve-month revenue with 60% gross margins and 20% operating margins. The company traded at a discount to growth rates, constrained by perceptions of limited TAM and lack of exposure to the AI boom that was driving 30-40% growth at the hyperscalers. That perception is now obsolete.

The arrival of CEO Paddy Srinivasan in 2024 marked an inflection point. Under his leadership, DigitalOcean accelerated product innovation, expanded its executive team, and augmented its product-led sales motion with strategic go-to-market enhancements. More importantly, the company recognized that the AI revolution wasn't just about training large models—it was about inference, the durable and predictable workload of running models in production. This insight, combined with the Paperspace acquisition's GPU capabilities, positioned DigitalOcean to capture a segment of the AI market that hyperscalers were structurally disadvantaged to serve: AI-native companies needing a complete stack, not just compute cycles.

Technology, Products, and Strategic Differentiation: The Full-Stack AI Play

DigitalOcean's competitive moat has evolved from simplicity alone to a comprehensive agentic cloud architecture that addresses the complete needs of AI-native enterprises. The Gradient AI Agentic Cloud comprises three layers: Infrastructure (GPU Droplets and Bare Metal), Platform (LLMs, serverless inferencing, knowledge services, guardrails), and Agents (purpose-built applications like Cloudways Copilot). AI-native companies don't just need GPUs—they need orchestration, storage, databases, authentication, and agentic workflow capabilities that hyperscalers provide through disjointed services requiring significant integration effort.

The platform's traction validates this approach. Over 19,000 agents have been created, with more than 7,000 already in production as of Q3 2025. More than 30% of these customers are new to DigitalOcean, indicating the AI platform is expanding the addressable market rather than merely upselling existing users. The shift toward inference workloads—management notes these are "predominantly inferencing, if not all, on the inferencing side"—creates durable, predictable demand patterns that mirror traditional cloud economics, unlike the project-based, experimental nature of early AI adoption.

Product innovation velocity has accelerated dramatically. In Q1 2025 alone, DigitalOcean released over 50 new products and features, more than five times the prior year's pace, without increasing R&D spend as a percentage of revenue. This efficiency stems from AI-assisted development, where internal use of generative AI has improved coding output by up to 40%. New capabilities like VPC peering, internal load balancers, droplet auto-scale pools, and Per-Bucket Access Keys directly address enterprise scalability requirements, while Spaces Cold Storage and automated database storage scaling remove friction for growing workloads.

The Atlanta data center, brought online in Q2 2025, exemplifies the strategic shift. Purpose-built for high-density GPU infrastructure optimized for AI inferencing, it hosts the full core cloud stack alongside GPU capacity. AI-native customers run full-stack applications, not just model inference. A company deploying a customer service agent needs databases for customer records, storage for conversation history, networking for secure access, and orchestration for deployment—exactly what DigitalOcean provides in a unified, simple interface that hyperscalers cannot match without significant integration overhead.

Financial Performance: Evidence of Business Model Evolution

DigitalOcean's Q3 2025 results provide compelling evidence that the strategic transformation is working. Revenue grew 16% year-over-year to $229.6 million, marking the highest growth rate since Q3 2023. More importantly, the composition and quality of that growth signal a fundamental business model evolution.

The $44 million in incremental organic ARR was the highest in company history, driven by two distinct engines: AI revenue more than doubled year-over-year for the fifth consecutive quarter, while general-purpose cloud products generated their highest incremental ARR since Q2 2022. This dual acceleration reinforces each other, with AI-native customers consuming core cloud services and traditional customers adopting AI capabilities.

Customer tier dynamics reveal the upmarket migration thesis in action. Scalers+ customers (>$100K ARR) grew revenue 41% year-over-year, increasing to 26% of total revenue—200 basis points higher than the prior year. Customers exceeding $500K ARR grew 55%, while those above $1M ARR grew 72%. This demonstrates DigitalOcean's ability to attract, retain, and expand its largest customers, a capability that was questionable when the company was perceived as merely a developer tool. The fact that over 35% of customers with more than $100K ARR have adopted at least one new feature released in the past year, and these customers saw a several-hundred-basis-point increase in their growth rate after adoption, proves that product innovation directly translates to revenue expansion.

Net Dollar Retention reached 99% in Q3 2025, up 200 basis points from 97% a year prior. While this remains below the 100%+ benchmark of best-in-class SaaS companies, the trajectory is critical: traditional cloud services NDR hit 100% in Q4 2024 for the first time since June 2023. Management explicitly notes that NDR does not yet include AI/ML revenue, meaning the reported figure understates true expansion potential as AI customers mature beyond their first year and begin expanding their footprints.

Profitability metrics validate the scalability of the model. Gross margin held steady at 60% in Q3 2025, while adjusted EBITDA margin reached 43% on $100 million of EBITDA.

The company generated $85 million in adjusted free cash flow, representing 37% of revenue—up from $19 million or 10% of revenue in Q3 2024. This 27-percentage-point improvement in free cash flow margin demonstrates that the growth investments are generating returns quickly, with growth capital for core cloud paying back in less than two years and AI growth capital paying back in around three years.

Outlook and Guidance: Pulling Forward 2027 Targets

Management's guidance narrative reveals extraordinary confidence in the trajectory. For 2025, DigitalOcean projects revenue of $896-897 million, representing 15% year-over-year growth—100 basis points higher than prior guidance. More significantly, the company expects to achieve its 2027 revenue growth target of 18-20% in 2026, a full year earlier than previously projected.

This acceleration is predicated on three levers: continued rapid growth of the largest customers (both AI and core cloud), AI revenue continuing its doubling trend to become a mid-teens or even high-teens percentage of total revenue, and sustained contributions from the product-led growth engine. The CFO explicitly states that AI revenue from committed contracts is "behaving more like traditional cloud, with scaled production workloads and more predictable demand," addressing a key investor concern about AI revenue quality.

The capacity expansion plan supports this ambition. DigitalOcean secured approximately 30 megawatts of incremental data center capacity for 2026 and beyond, driven by signing multiple 8-figure committed contracts after Q3 closed. The company began ordering more GPU capacity and entered into leases for co-location space with total estimated undiscounted payments of $284.8 million, with ten-year terms commencing in March 2026. This front-loaded investment will pressure margins in early 2026 as COGS and operating expenses ramp, but management anticipates maintaining high-30s to 40% adjusted EBITDA margins and mid-to-high-teens adjusted free cash flow margins.

The guidance assumptions are notably conservative on macroeconomic factors. Management explicitly states they have not modeled major variations in end-user demand, expecting stability despite acknowledging "unfavorable conditions in the economy both in the United States and abroad." This creates potential upside if macro conditions remain stable or improve, but also risk if economic deterioration reduces SMB cloud spending.

Risks and Asymmetries: What Could Break the Thesis

The most material risk is execution on capacity expansion. Management's admission that "we are now seeing more demand than we can support" is both a validation of product-market fit and a warning sign. The company must deliver 30 megawatts of new data center capacity in the first half of 2026 to meet committed contracts and support projected growth. Any delay in data center commissioning, supply chain disruption for GPUs (though management notes they are "in a real healthy position from a supply standpoint"), or cost overruns could constrain revenue and damage customer relationships.

Competitive pressure from hyperscalers remains a persistent threat. AWS, Azure, and GCP are all aggressively targeting SMB and developer segments with simplified pricing and managed services. Recent high-profile outages at AWS and Azure could catalyze migrations to DigitalOcean's more reliable, simpler infrastructure, but hyperscalers have deep resources to match features and undercut pricing. The risk is asymmetric: DigitalOcean's differentiation is based on simplicity and approachability, which is harder to maintain as competitors improve their developer experience.

AI demand volatility presents a nuanced risk. While management emphasizes that inference workloads are "durable" and "predictable," the AI market remains nascent. A shift in AI architecture toward more efficient models (like DeepSeek) or a slowdown in AI application deployment could reduce GPU demand. However, management frames such innovations as reinforcing their conviction that "value creation will occur at the platform and application layers, where we are highly differentiated," suggesting they view efficiency gains as expanding rather than contracting their market.

Customer concentration in SMBs creates macro sensitivity. While Higher Spend Customers now represent 89% of revenue, the underlying customer base remains weighted toward digital-native enterprises and startups that are more vulnerable to funding crunches and economic downturns than enterprise behemoths. A severe recession could compress NDR and increase churn, particularly among Builders and Scalers tiers.

The balance sheet transformation, while strategically sound, introduces new financial complexity. The $625 million of 0% convertible notes due 2030, combined with $380 million drawn on the Term Loan Facility, creates potential dilution and interest expense headwinds. Management notes the net impact of these activities reduced non-GAAP net income per share by $0.05 in Q3 2025, and the company anticipates net leverage will end 2026 in the mid-3s range. This is manageable but represents a shift from the company's historically conservative capital structure.

Valuation Context: Pricing a Business Model Inflection

At $48.30 per share, DigitalOcean trades at an enterprise value of $5.80 billion, representing 6.71 times trailing twelve-month revenue and 21.54 times adjusted EBITDA. The price-to-free-cash-flow ratio stands at 32.12, while price-to-operating-cash-flow is more attractive at 13.65. These multiples place DigitalOcean in a middle ground between hyperscaler valuations (Amazon trades at 3.65x EV/Revenue, Microsoft at 12.29x) and high-growth SaaS companies.

The valuation must be assessed in the context of accelerating growth and margin expansion. With revenue growth re-accelerating to 16% and management guiding for 18-20% in 2026, the forward revenue multiple compresses meaningfully. The 43% adjusted EBITDA margin and 37% free cash flow margin are substantially higher than the hyperscalers' cloud margins, justifying a premium multiple for capital efficiency.

Balance sheet strength provides downside protection. As of September 30, 2025, DigitalOcean held $236.6 million in cash and had $420 million available under its credit facilities, against $1.32 billion in total debt. The company generated $282.7 million in operating cash flow over the trailing twelve months, giving it ample liquidity to fund capacity expansion without dilutive equity raises.

The key valuation question is whether DigitalOcean deserves a multiple closer to infrastructure providers or platform companies. The evolution toward a full-stack AI platform with 19,000+ agents and mid-teens revenue mix potential suggests platform-like characteristics. However, the underlying infrastructure business and capital intensity (capex paybacks of 2-3 years) keep it tethered to asset-heavy valuations. The current multiple appears fair if the company executes on its 2026 growth targets, but offers limited margin of safety if capacity constraints or competitive pressure derail the trajectory.

Conclusion: A Cloud at the Crossroads of Simplicity and Scale

DigitalOcean stands at an inflection point where its founding principles of simplicity and approachability are becoming competitive advantages in the AI era. The company's ability to double AI revenue for five consecutive quarters while simultaneously re-accelerating core cloud growth demonstrates a business model that is evolving, not being replaced. The upmarket migration, evidenced by 41% growth in Scalers+ customers and 72% growth in $1M+ ARR accounts, proves that simplicity resonates with larger enterprises overwhelmed by hyperscaler complexity.

The investment thesis hinges on two variables: execution on capacity expansion and AI revenue scaling to mid-teens percentage of total revenue. The 30 megawatts of secured capacity and multiple 8-figure contracts signed post-quarter provide visibility, but any slippage in data center commissioning could constrain the 18-20% growth target for 2026. Similarly, while inference workloads appear durable, the AI market's nascence creates uncertainty around long-term demand patterns.

DigitalOcean's competitive positioning is strengthening. The full-stack Gradient AI Agentic Cloud addresses needs that pure GPU providers cannot, while the core cloud's return to 100% NDR and highest incremental ARR since 2022 shows the traditional business is not being neglected. The balance sheet optimization provides financial flexibility, though the shift to leveraged capital introduces new complexity.

For investors, the story is one of a company successfully navigating a strategic transformation while maintaining capital discipline and profitability. The stock's valuation at 6.7x revenue and 21.5x EBITDA appears reasonable for a business re-accelerating to 18-20% growth with 40%+ EBITDA margins, but offers limited cushion for execution missteps. The key monitorables are capacity delivery timelines, AI revenue mix progression, and competitive response from hyperscalers. If DigitalOcean executes, it will have pulled forward its 2027 growth targets and established itself as a permanent fixture in the AI infrastructure landscape. If not, the company risks squandering a rare moment of product-market fit in a market that waits for no one.