Earnings call transcript: DigitalOcean tops Q3 2026 estimates, lifts outlook

DigitalOcean’s Q3 revenue rose 40% year over year to $321 million, beating the $306.66 million consensus, while adjusted EPS of $0.45 exceeded the $0.32 estimate and adjusted EBITDA margin reached 45%, above 38%–39% guidance. Management raised 2026 revenue guidance to $1.208 billion–$1.211 billion and lifted its 2027 growth outlook to 60% or more; AI customer ARR grew 288% to $376 million. GAAP net loss was $43 million, attributed mainly to a one-time induced conversion expense, and shares were up 0.65% in premarket trading.

Analysis

DOCN’s key underwriting question is whether software attach can sustain returns as GPU capacity expands—not whether AI usage is growing. Management’s ARR-per-megawatt framing is a revenue run-rate, not a measure of cash return: utilization, depreciation, power, financing costs, and the timing of customer ramp determine actual economics. New capacity can therefore lift revenue while temporarily compressing gross margin and free cash flow. If platform attach is real, DOCN should be less exposed than GPU-only providers to falling inference prices; if workloads remain capacity-led, supply normalization could quickly weaken pricing and utilization.

Near term, the raised outlook supports estimates, but the limited initial reaction and the stock’s prior run-up raise the bar for incremental upside. Over 1–3 months, watch Q4 delivery, gross-margin direction, and evidence that AI customers expand into higher-value services. Over 6–18 months, 2027 capacity commissioning and utilization will test the claimed software-driven yield. Larger cloud platforms and specialized infrastructure providers remain competitive threats; GPU vendors and colocation operators may benefit from deployment, but supplier economics are not established by DOCN’s demand claims.

Treat management’s platform and demand assertions as hypotheses until cohort expansion and cash conversion corroborate them. The $1 billion ATM provides funding flexibility but creates a potential dilution overhang. Amplitude’s use case is a product validation signal, not enough evidence to infer a material financial impact for AMPL.

Key Decisions for Investors

• Consider a staged long in DOCN after the opening volatility settles rather than chasing the first print. The thesis is estimate revision and platform attach; size against the risk that elevated expectations leave little room for execution slippage.
• Reassess after Q4 results: a miss against the company’s 44%–45% growth outlook, weakening gross margin without improving utilization, or deterioration in free-cash-flow conversion would falsify the near-term long thesis.
• Track 2027 capacity starts and utilization alongside ARR per megawatt. Delays, falling inference prices, or weaker customer expansion would challenge the assumption that software can offset GPU supply normalization.
• Do not trade AMPL on the customer example alone. Verify whether DigitalOcean adoption produces measurable revenue or retention impact before treating it as an AMPL catalyst; monitor DOCN ATM usage for dilution risk.