DigitalOcean added a record $160 million of annual run-rate revenue (ARR) in the third quarter, it reported on 3 November. Its AI Customer ARR rose from $234 million at the end of June to $376 million at the end of September.
That means customers using DigitalOcean’s AI products accounted for $142 million of the $160 million, about 89%, by our arithmetic. ARR from all other customers rose by about $18 million.
The same day, the company filed to sell up to $1.0 billion of new shares through an at-the-market program, which the card circulating on X did not mention.
Business Pill · Neoclouds
An evergreen Business Engineer explainer on neoclouds: cloud providers built for AI computing that rent out chips by the hour. It explains the general idea only and does not discuss DigitalOcean.
The key insight: As we read the two releases, DigitalOcean’s growth is now almost entirely its AI customer base: the rest of the company’s ARR rose about 2% in the quarter, by our arithmetic, while AI customers added $142 million.
For the quarter to 30 September, revenue was $321 million, up 40%, and ARR was $1,285 million, also up 40%. AI Customer ARR, at $376 million, was up 288%. The company called it the fifth consecutive quarter of accelerating revenuegrowth.
Adjusted EBITDA was $144 million, a 45% margin, and operating income $33 million. The net loss attributable to common stockholders was $43 million, which included a $57 million induced conversion expense from repurchasing part of its 2030 convertible notes. Non-GAAP diluted earnings per share were $0.45.
Wall St Engine’s highlights card put the consensus estimates at $307 million for revenue and $0.32 for adjusted earnings per share.
DigitalOcean’s annual run-rate revenue (ARR) split between AI customers and all other customers. Left: ARR of $1,125 million at 30 June 2026, of which AI Customer ARR was $234 million. Right: the record $160 million of ARR added in the third quarter, of which $142 million came from the rise in AI Customer ARR to $376 million. Shares, the $891 million and $18 million for other customers and the $142 million are our arithmetic from the company’s second- and third-quarter releases. AI Customer ARR counts all revenue from customers using at least one AI offering.
At 12:12 UTC on 3 November, Wall St Engine (@wallstengine) posted its highlights card for DigitalOcean’s third quarter: revenue, adjusted earnings per share and adjusted EBITDA with consensus estimates, ARR, the full-year and fourth-quarter outlook, Managed Agents, data centre capacity, the $785 million equipment lease facility, AI Customer ARR of $376 million and remaining performance obligations of $800 million.
Where the New ARR Came From
DigitalOcean calculates AI Customer ARR by multiplying a quarter’s revenue from AI customers by four. It counts every customer that uses at least one of its AI or machine-learning offerings, including what those customers spend on its ordinary cloud products.
On that measure, AI customers made up about 21% of ARR at the end of June and about 29% at the end of September, by our arithmetic. ARR from other customers went from about $891 million to about $909 million, a rise of about 2% in the quarter.
“Inference Engine token volume grew roughly 10x quarter-over-quarter,” said Paddy Srinivasan, the chief executive. He said customers spending over $100,000 a year now represent 41% of revenue, and “they are not renting capacity; they are adopting the platform.”
The Capacity and the Money Behind It
DigitalOcean said it secured an incremental 20 megawatts of data centre capacity for 2028, bringing committed capacity to about 180 megawatts, and a $785 million equipment lease facility to fund capacity expansion. Remaining performance obligations were $800 million, $345 million of it due within 12 months, against $61 million a year earlier.
In the quarter it also repurchased about $472 million of its 2030 convertible notes, funded by a concurrent registered direct offering of shares.
Then, on 2 November, it signed an equity distribution agreement with eight banks, led by Goldman Sachs, J.P. Morgan and Morgan Stanley, to sell common stock with an aggregate offering price of up to $1.0 billion. The prospectus supplement says the proceeds are for general corporate purposes, including financing operations, growth initiatives and possible acquisitions.
At the $89.79 closing price of 2 November, the prospectus assumes sales of 11,137,097 shares. That is about 9.4% of the 117,920,731 shares outstanding at 30 September, by our arithmetic. The company is not obliged to sell any shares.
The Outlook
For the fourth quarter, DigitalOcean guides revenue of $348 million to $351 million, up 44% to 45%, against a consensus estimate of $333 million cited by Wall St Engine. For 2026 it expects revenue of $1.208 billion to $1.211 billion. The release says the company now expects revenuegrowth of more than 60% for 2027.
What the Risk Factors Say
The 10-Q filed the same day states that there have been no material changes to the risk factors in the annual report, by the reading of our risk-factor parser.
As we read the definition, the $376 million is a measure of customers, not of products: it counts every dollar an AI customer spends, so it overstates AI productrevenue and can grow when an existing customer tries one AI service.
As we read the filings, new financing is arriving alongside the demand: a $785 million equipment lease facility for capacity in the quarter and, on results day, a program to sell up to $1.0 billion of shares for general corporate purposes.
As we read the outlook, the company is guiding fourth-quarter revenue of $348 million to $351 million, about 9% above the third quarter at the midpoint, by our arithmetic.
Paddy Srinivasan, chief executive, DigitalOcean, in the results release, 3 November 2026
“they are not renting capacity; they are adopting the platform.”
Three Implications
FOR DIGITALOCEAN’S MIX On the company’s measure, AI customers rose from about 21% to about 29% of ARR in one quarter.
FOR SHAREHOLDERS The at-the-market program could add about 9.4% to the share count at the 2 November price, by our arithmetic, if fully used.
FOR AI-NATIVE CLOUDS DigitalOcean now discloses an AI customer ARR line, a revenue backlog of $800 million and about 180 megawatts of committed capacity.
The framework puts it this way: “Inference is a revenue engine: continuous, distributed across millions of applications and agents, incurring marginal cost with every query.”
As we read it, DigitalOcean’s quarter shows both halves: token volume up roughly tenfold in a quarter, by the chief executive’s account, and a new $785 million lease facility for the capacity that serves it, with a $1.0 billion share program for general corporate purposes alongside. The framework does not discuss DigitalOcean.
What Is Not Established
AI Customer ARR is not AI revenue: it includes everything an AI customer spends, so part of the $142 million may be customers who began using an AI product and brought existing spending into the category. The release does not split the change into new spending and reclassified customers.
The filings give no number of shares sold under the program so far and no timetable. We did not review the earnings call.
This Week In The AI Supercycle
Anthropic’s Reality Check, The Capital Squeeze, & The New World Order
This week’s issue: for two years the AI build-out was priced as a growthstory. This week it started being priced as a credit story.
Almost nine-tenths of DigitalOcean’s record quarterly ARR gain came from customers using its AI products, by our arithmetic, while the rest of the base grew about 2%. It added a $785 million equipment lease facility to fund capacity and, on the prospectus’s terms, a program to sell up to $1.0 billion of stock for general corporate purposes.
A note on sourcing. This piece is based on DigitalOcean’s third-quarter results release, its Form 8-K and prospectus supplement filed on 3 November 2026, and its second-quarter release of 4 August 2026. The $142 million ($376 million minus $234 million), the 89% share ($142 million divided by $160 million), the other-customer ARR ($1,125 million minus $234 million, and $1,285 million minus $376 million), the AI shares of ARR and the 9.4% (11,137,097 divided by 117,920,731 shares) are our arithmetic. Consensus estimates are as cited by Wall St Engine. We did not contact DigitalOcean.
Gennaro Cuofano is a CRO and tech executive who has worked in AI since late 2015, starting with bringing NLP, natural-language generation, voice and chatbot products to the marketing industry. His background is in law and finance: he holds a Master's degree in Law and an International MBA with an emphasis on corporate finance (LUISS Business School and the University of San Diego, 2012), and worked as a financial analyst at a real-estate investment firm in San Diego and as an assistant controller. His work focuses on business model strategy, business engineering and, more broadly, structural analysis: how companies make money, read from their own filings. He created FourWeekMBA and leads research there and at The Business Engineer, his newsletter on AI and business strategy, with over 95,000 subscribers and more than 1,000 published analyses. His writing has also appeared on Entrepreneur (for example, How to Digitize a Family Business in Europe) and HackerNoon. Find him on LinkedIn and Substack. See how we source.
