DigitalOcean: Inference Growth Meets the Lease Test

Executive assessment

DigitalOcean Holdings, Inc. presents a striking split between operating momentum and its share-price trend. Third-quarter revenue increased 39.9% to $321.2 million, and management raised its 2027 revenue-growth expectation to more than 60%. Yet DOCN closed November 4 at $84.02, down 7.16% for the session and below both its 14-period Kaufman adaptive moving average and 50-session simple moving average. Growth is accelerating; the stock is not confirming it. Neither observation establishes why investors sold. [1] [2]

The strongest evidence is the $160 million quarterly increase in annual run-rate revenue and a $800 million remaining performance obligation balance. The biggest qualification is that the company's $130 million adjusted quarterly free cash flow excludes equipment acquired under financing arrangements, finance leases and future contractual commitments. That distinction matters more as the asset base expands. These are company-reported non-GAAP measures, not interchangeable measures of cash available to shareholders. [1]

Company and industry: from developer cloud to inference platform

DigitalOcean identifies itself as an AI-native cloud for inference and agentic workloads. Its integrated offering spans GPU and CPU infrastructure, core cloud, inference, data and managed agent orchestration. The company reports more than 680,000 customers; the opportunity is to retain its developer-oriented entry point while selling a broader production platform to larger customers. Its November 3 release independently identifies the listed issuer as DigitalOcean Holdings, Inc., NYSE: DOCN. [1]

The industry contest is not merely access to GPUs. Customers need reliable deployment, data access, orchestration and attractive economics for repeated inference. Integration can make a platform useful beyond rented capacity, but it does not establish a durable moat by itself. Larger cloud platforms, specialist providers and evolving model economics can all pressure returns. DigitalOcean's stated open architecture is a positioning claim; retention and realized margins are the financial tests.

The mix is shifting materially: customers spending more than $100,000 annually generated 41% of third-quarter revenue, with revenue from that cohort up 133%. Customers above $1 million represented 30% of revenue. Those categories overlap and must not be added together. Bigger accounts can support platform adoption while also making results more sensitive to deployment schedules and individual purchasing decisions. [1]

Fundamentals: growth outruns operating profit

The financial statements below are unaudited company disclosures for the quarter ended September 30, 2026, released November 3—not November 4 financial observations. Dollar values are millions, except per-share figures. [1]

Revenue growth has not yet produced operating leverage on a GAAP basis. Gross profit rose in dollars, but the gross-margin contraction absorbed much of the incremental revenue before operating expenses. Adjusted operating income also grew far more slowly than revenue. The net loss should not be read as a clean measure of recurring deterioration: the conversion charge matters, and the prior-year quarter included a $48.373 million debt-extinguishment gain and $68.057 million income-tax benefit. Equally, excluding the conversion charge does not erase the operating-margin decline. [1]

Run-rate demand is not contracted recurring revenue

Reported ARR reached $1.285 billion and AI Customer ARR reached $376 million, up 288%. DigitalOcean defines ARR as quarterly revenue multiplied by four. AI Customer ARR includes all revenue from customers using at least one AI/ML offering, including their infrastructure and platform spending; it is not a pure measure of AI product revenue. The $800 million RPO balance includes $345 million expected to be recognized over the next 12 months. The company cautions that consumption timing and shifts to commitment-based agreements can change RPO without equivalent incremental growth. The earlier comparison periods were recast for the updated RPO definition. [1]

Balance sheet and the capital-cost trap

At September 30, cash and equivalents were $666.096 million and marketable securities were $178.645 million, or $844.741 million combined. Current and long-term debt totaled $461.630 million. Looking only at those figures would miss $923.599 million of finance-lease liabilities and equipment-financing obligations, plus $666.611 million of operating-lease liabilities. These are carrying amounts, not a schedule of all future undiscounted payments; they should not be mistaken for a single immediately payable debt bill. [1]

Net property and equipment increased to $1.408 billion from $589.094 million at December 31, 2025. The company secured a $785 million equipment lease facility and reported approximately 180 MW of committed capacity, including an incremental 20 MW expected in 2028. A facility is financing capacity, not revenue or cash already earned; committed megawatts are not the same as energized, utilized or profitable capacity. [1]

For the first nine months, operating cash flow was $318.116 million. Subtracting $105.124 million of property-and-equipment expenditure, $19.279 million of internal-use software and $0.754 million of intangible purchases leaves $192.959 million before other investing and financing flows. Subtracting the separately reported $75.143 million equipment acquisition under financing arrangements would reduce that illustrative amount to $117.816 million. This is an analytical bridge, not the company's adjusted free cash flow or a complete measure of owner earnings: it still does not capture all leased equipment additions or future commitments. Principal repayments of finance leases and financing arrangements were another $39.895 million in the financing section. [1]

Valuation: the growth hurdle remains high

Management's November 3 outlook calls for full-year 2026 revenue of $1.208–$1.211 billion, adjusted EBITDA margin of 41–42%, adjusted free-cash-flow margin of 20–22% and non-GAAP diluted EPS of $1.60–$1.62. At the November 4 close, the midpoint EPS implies a price/non-GAAP-guided-earnings ratio of approximately 52.2 times ($84.02 ÷ $1.61). This is not a GAAP trailing P/E, an independent forecast or a target price. [1] [2]

The midpoint revenue and free-cash-flow-margin assumptions imply about $254.0 million of annual adjusted free cash flow ($1.2095 billion × 21%). That is a mechanical midpoint illustration, not separate guidance, and it retains the company's exclusions for financed equipment and leases. A premium earnings multiple requires more than fast sales: shareholders need durable unit economics and growth in earnings per diluted share. Full-year 2026 guidance assumes roughly 122–123 million diluted weighted-average shares, while Q4 guidance assumes 128–129 million. Those are different averaging periods, not a calculation of incremental shares issued. [1]

Technicals: a falling price beneath both trend measures

The 90-calendar-day window is August 7–November 4, 2026, containing 63 observed trading sessions. No weekend or holiday bars are inserted. DOCN closed at $84.02, versus $90.50 on November 3: a decline of $6.48, or 7.16%. November 4 volume was 6,218,943 shares. The 14-period KAMA was $99.31967 and the 50-session SMA was $118.2912; the close stood 15.4% and 29.0% below them, respectively. Indicators use daily history beginning June 1 for initialization; only the stated 90-day window is displayed. [2]

USD; daily observed closes and trend indicators. Source: StockApp daily series, through November 4, 2026. KAMA is path-dependent; its initialization window is part of the measurement. SMA observations are shown from August 12; earlier dates are left blank.

The decline from the September 21 window-high close of $146.16 to $84.02 is 42.5%. Recovering a prior price requires a different percentage gain. More importantly, the final close is the lowest in this displayed window. A bounce toward KAMA would still leave the stock below its slower trend measure. Those averages describe trend, not intrinsic value or guaranteed resistance. A more constructive setup would require sustained price improvement alongside capital-efficiency evidence; a single green session would not resolve either issue. [2]

Catalysts and evidence to watch

• Q4 delivery: management guides to $348–$351 million revenue, 44–45% year-over-year growth, and 42–43% adjusted EBITDA margin for the quarter ending December 31, 2026. Revenue conversion alongside gross margin is more informative than revenue alone. No unannounced earnings-release date is assumed. [1]
• 2027 acceleration: the stated expectation of more than 60% revenue growth is a management forecast, not an achieved result. Watch whether incremental demand is diversified and whether its financing supports per-share economics. [1]
• Platform adoption: Managed Agents, Agent Learning Services and Agent Data Services were announced October 13, according to the November 3 release. Customer usage and retention will test whether these services deepen the business beyond infrastructure rental. [1]
• Capacity utilization: management says all three 2026 data centers were delivered on time or early. Future utilization, lease servicing and incremental gross profit—not delivery alone—will determine the investment return. The additional 2028 capacity is a longer-dated commitment, not a near-term earnings catalyst. [1]

Scenarios: conditional outcomes, not price targets

No numerical probabilities, analyst consensus or target prices are assigned. These are analytical conditions anchored to the disclosed business model, not observed future outcomes.

Risks and decision framework

Capital intensity: adjusted free cash flow can look strong while equipment-financing and lease obligations rise. Execution: power, GPU availability and data-center timing can disrupt delivery. Demand: larger accounts and commitment agreements can create uneven recognition and concentration risk. Competition: the value of a differentiated platform must survive pricing pressure and changing inference technology. Shareholder economics: stock compensation and financing transactions can dilute participation in operating growth. These concerns are consistent with the issuer's disclosed risk categories, but their future severity is uncertain. [1]

For portfolio construction, treat DOCN as a volatile growth exposure, not a defensive cash-flow substitute. The practical checklist is straightforward: reconcile revenue growth to gross profit, reconcile cash generation to financed equipment and lease service, and reconcile enterprise growth to diluted per-share outcomes. The November 4 technical picture argues against assuming that improved guidance alone has repaired the risk/reward balance.

Scope: financial disclosures available by November 4, 2026, centered on the November 3 Q3 release; prices end November 4. No subsequent trading, unsupported macro observations, social-sentiment inference or quantified peer comparison enters this assessment. This analysis is not personalized investment advice.