DOCN
DigitalOcean
Summary
What they do:
DigitalOcean operates a developer-focused cloud platform serving SMB and startup customers with compute, managed databases, Kubernetes, and — since acquiring Paperspace in 2023 — GPU cloud and AI inference services, positioning itself as the "agentic inference cloud" for growing technology companies.
Why they matter:
DOCN is the only public pure-play SMB cloud platform, now pivoting to AI inference at a scale accessible to smaller teams — $120M AI customer ARR growing 150% YoY proves there is demand for AI infrastructure below the hyperscaler threshold.
Recent performance:
FY2025 revenue $901M (+18% YoY); Q4 2025 revenue $242M (+18% YoY); net income $259M (+207% YoY). AI customer ARR $120M (+150% YoY). Stock ~$84, market cap ~$9B. Reached $1B annualized monthly revenue in December 2025.
Our Verdict
A profitable SMB cloud platform adding an AI inference layer through Paperspace GPU integration — AI customer ARR of $120M (+150% YoY) is promising but represents only ~13% of total revenue, making DOCN a tangential AI play trading at ~10x revenue where the thesis depends on whether inference demand from smaller customers can meaningfully accelerate the 18-21% total growth rate.
Structural trends
Structural
43
/ 100
Moat
3/10
Developer brand + SMB niche
AI Exp.AI Exposure
Embedded~13% AI
Play Type
EmergingAI Growth
~150% YoY
Rel. Value
44
FAIRPriceLIVE
$73.45
-8.04%
Live via Yahoo Finance · refreshes every 5 min
Market Cap
$7.6B
P/E Ratio
31.7
P/S Ratio
8.5x
52W High
$94.66
52W Low
$25.56
52W Chg
187.4%
Beta
1.43
DigitalOcean was founded in 2011 to simplify cloud infrastructure for individual developers and small teams. While AWS, Azure, and GCP built increasingly complex platforms for enterprise customers, DigitalOcean carved out a niche with transparent pricing, simple APIs, and a developer community orientation. The company went public in March 2021 and has grown steadily, reaching $901M in FY2025 revenue with approximately 600,000+ customers globally.
The core business is traditional cloud infrastructure — virtual machines (Droplets), managed Kubernetes, databases (MySQL, PostgreSQL, MongoDB, Redis), object storage, and CDN — sold primarily to SMB technology companies and independent developers. Average revenue per customer (ARPU) has climbed to approximately $112, reflecting a deliberate push upmarket from hobbyists to revenue-generating businesses. Net dollar retention improved to 100% in early 2025, up from the high-90s in prior years, indicating customers are spending more over time.
The AI pivot began with the $111M acquisition of Paperspace in July 2023, which brought GPU cloud capabilities (NVIDIA H100, L40S, AMD MI300X) and a machine learning developer community. DigitalOcean integrated Paperspace as its GPU and AI platform, launching GPU Droplets and a GenAI platform for deploying AI agents, serverless inference, and fine-tuning models. By Q4 2025, AI customer ARR reached $120M, growing 150% YoY — though more than 70% of this comes from inference services and core cloud products used by AI-building customers, not bare-metal GPU rentals.
The company is adding 31MW of new data center capacity in 2026 to support GPU workloads, with the first facility ramping in Q2 and remaining capacity online in H2 2026. Management guided FY2026 revenue growth of 19-23% (21% midpoint) with plans to exit 2026 at 25%+ growth and reach 30% growth in 2027 — an acceleration that depends heavily on AI customer expansion.
The Catch
DigitalOcean's core challenge is scale mismatch. The company positions itself as an "agentic inference cloud" but is adding 31MW of GPU capacity in 2026 while CoreWeave adds 850MW+ and hyperscalers add thousands of MW. AI customer ARR of $120M is impressive growth from a small base, but it represents less than 0.1% of the estimated $200B+ AI cloud market. The real risk is that DigitalOcean's AI positioning is genuine but irrelevant — SMB inference demand exists, but it may be too small and too price-sensitive to generate the margin expansion needed to justify a growth re-rate. Meanwhile, hyperscalers are aggressively simplifying AI services for smaller customers (AWS Bedrock, Google Vertex AI Agent Builder, Azure AI Studio), threatening to pull DOCN's best AI customers upmarket before the inference economy reaches scale.
If They Win
If DigitalOcean's "agentic inference cloud" thesis plays out — SMB and startup developers choose DOCN over hyperscalers for AI inference workloads because the platform is simpler, cheaper, and purpose-built for smaller teams — the company becomes the Shopify of AI infrastructure. In this scenario, AI customer ARR reaches $500M-1B by 2028-2029, total revenue exceeds $2B with 30%+ growth, and the GPU capacity expansion (from 31MW to hundreds of MW) drives a margin mix shift as inference workloads carry higher ARPU. DOCN trades at $200-250 as the market re-values it from a mid-market cloud platform ($10B) to the dominant SMB AI infrastructure provider ($25-30B) — the company that made AI accessible to the 600,000+ businesses that could never afford CoreWeave or Azure.
Others in Run the Workload
Not financial advice. All scores generated via AI algorithms using public data.