DigitalOcean (DOCN) Q2 2026 earnings review
Hypergrowth Returns, But Capital Intensity Follows
DigitalOcean has successfully transformed from a developer-focused host into an AI-Native Cloud, and the financial results are staggering. Q2 revenue growth accelerated to 29% YoY, blowing past expectations, driven by an insatiable demand for AI inference. The company signed its first nine-figure annual commitments, causing Remaining Performance Obligations (RPO) to explode by 12x to $894M. Guidance points to sustained acceleration, with a projected Q4 exit growth rate of over 35%. However, this growth is not cheap. GAAP net income and operating income both declined YoY, as heavy depreciation and stock-based compensation from massive infrastructure investments overshadowed top-line gains.
🐂 Bull Case
The Inference Engine saw token volume increase 30x in just 60 days, adding over 6,000 customers. Management asserts demand is far outstripping current capacity, providing immense visibility into 2027.
The company is shedding its 'small developer' limits. ARR from $1M+ customers grew an astonishing 214% YoY and now represents nearly a quarter of total revenue.
🐻 Bear Case
Despite adding $62M in YoY revenue, GAAP operating income dropped 18%. The costs to build and staff the AI cloud are weighing heavily on the bottom line.
The company secured another 20 MW of capacity, bringing the total to 155 MW. Delivering on a 50%+ growth target in 2027 requires flawless data center deployment and zero softening in AI demand.
⚖️ Verdict: 🟢
Bullish. The speed of revenue acceleration and the massive RPO spike validate DigitalOcean's strategic pivot. While near-term cash flow and GAAP margins are under pressure from infrastructure investments, the high attach rate of core cloud products to AI workloads suggests strong long-term unit economics.
Key Themes
The Macro 'Valuemaxxing' Shift to Open-Weight Models
A massive structural shift is underway in the AI market: companies are moving away from expensive frontier models toward open-weight models to control unit economics. Management noted that open-weight models have surged from 15% to 75% of token volume on their platform. This 'valuemaxxing' trend perfectly suits DigitalOcean, which is optimizing infrastructure specifically to serve these open models efficiently.
Inference Engine and Model Synthesis
Technological innovation is directly driving top-line results. The new Inference Engine acts as a full production runtime, solving complex batching and routing problems. Notably, the new 'Model Synthesis' feature allows customers to run requests across multiple models in parallel to achieve frontier-grade quality at a fraction of the cost. The platform also secured a 'Day 0' launch for the massive Kimi K3 open-weight model, immediately attracting 400 net new customers.
AI Flywheel Pulls Core Cloud
DigitalOcean is proving it is not just a 'GPU landlord.' A critical metric: 85% of AI customer ARR comes from inference and core cloud products, not bare metal. Furthermore, ~70% of AI customers spending over $100K are attaching core cloud primitives (storage, databases, networking) to their workloads. This twin-stack approach creates sticky, high-margin recurring revenue.
Hyper-Scaling Large Customers
The thesis that DigitalOcean's best customers outgrow the platform has been definitively broken. The higher the spend cohort, the faster the growth. $100K+ customers grew 98%, $500K+ grew 160%, and $1M+ customers surged 214%. The $1M+ cohort now represents 23% of total revenue.
Nine-Figure Commitments Surge RPO
The company landed its first nine-figure annual customer commitments during the quarter. This caused Remaining Performance Obligations (RPO) to explode to $894M, up 12x from a year ago. Crucially, the weighted average contract life extended from 1.6 years to over 3 years, providing unprecedented revenue visibility.
GAAP Profitability Reversing
A glaring contradiction in the hypergrowth narrative: despite a 29% surge in revenue, GAAP Operating Income actually fell 18% YoY (from $35.6M to $29.4M), and GAAP Net Income fell 4%. This margin compression was driven by heavy depreciation ($51.1M vs $32.7M a year ago) and surging stock-based compensation ($32.7M vs $21.0M). The company is buying its growth through heavy infrastructure and talent costs.
Massive CapEx and FCF Compression
The aggressive build-out of 155 MW of data center capacity is consuming cash. While Adjusted Free Cash Flow was healthy at 22% in Q2, full-year FY26 guidance calls for a compressed 11-13% margin. This indicates heavy upcoming capital expenditures and equipment lease payments in the second half of the year, breaking the historical ~20% FCF margin trend.
Concentration in Volatile AI Startup Ecosystem
While management noted that the top 25 customers only represent 20% of ARR, the explosive growth is heavily reliant on 'AI-Native' companies and agentic app developers. This ecosystem is inherently volatile; if AI startup funding dries up or consolidation occurs rapidly, token consumption could violently contract.
Other KPIs
Accelerating dramatically. Up 212% YoY, this segment now represents over 20% of the company's total $1.125 billion ARR. The growth proves the successful transition from a general-purpose host to a specialized AI destination.
Accelerating. A record quarter for the company, representing a 191% YoY increase in the absolute dollar amount of new recurring revenue added in a single 90-day period.
Stable to slightly decelerating in margin. While the absolute dollar amount grew 19% YoY, the operating cash flow margin slipped from 42% in Q2 2025 to 39% this quarter, reflecting the rising upfront costs of operating the expanded data center footprint.
Guidance
Accelerating. The midpoint implies ~33% YoY growth, a clear step up from Q2's 29%, showing that new capacity is being monetized as fast as it comes online.
Accelerating. Raised from prior expectations. Management is guiding for full-year growth of ~30.5%, but crucially noted an expected exit growth rate in Q4 of 35% or more, which underpins their informal confidence in >50% growth for 2027.
Stable. Demonstrates that while GAAP profitability is taking a hit from depreciation, the underlying cash-proxy operating economics of the business remain resilient.
Decelerating. A significant step down from the 22% achieved in Q2 and historical norms. Management is absorbing massive upfront cash outflows to secure and build the 2027/2028 data center capacity.
Key Questions
Margin Profile of Nine-Figure Deals
With the average contract life extending beyond 3 years on these massive new commitments, did the company have to sacrifice gross margin or offer significant volume discounts to secure this RPO spike?
Free Cash Flow Trough Timing
With FY26 Adjusted FCF margin guided down to 11-13%, when exactly does management expect this metric to bottom out? Will the continuous addition of 20 MW increments keep FCF permanently suppressed below historical levels?
Open-Weight Revenue Economics
As token traffic shifts aggressively (from 15% to 75%) toward open-weight models, how does the revenue realization per million tokens compare to serving closed-source frontier models? Is this a volume-over-price play?