# DigitalOcean account plan

> DigitalOcean is an NYSE-listed cloud for AI and digital native companies, with $901.4 million FY2025 revenue and $1,285 million ARR in Q3 2026.

## At a glance

DigitalOcean Holdings (NYSE: DOCN), headquartered in Broomfield, Colorado, now calls itself "an AI-Native Cloud, purpose-built for inference and agentic workloads". It sells consumption-based IaaS (Droplets, storage, networking), PaaS/SaaS (Cloudways managed hosting, Managed Databases, Managed Kubernetes, App Platform) and the Gradient AI Agentic Cloud (GPU Droplets, Bare Metal GPUs, inference and agent services) to growing technology companies in about 190 countries. FY2025 revenue was $901.4 million, up 15%, with net income of $259,262 thousand and adjusted EBITDA of $374,805 thousand (42% margin). It had 1,462 employees at 31 December 2025, 848 of them outside the United States, and reports a single operating segment.

## Business units

- **Digital Native Enterprise (DNE) customers** (60% of FY2025 revenue ($539,842 thousand); 69% in Q3 2026): Customers spending more than $500 a month, about 21,000 at the end of 2025 and 22,374 in Q3 2026. Their revenue grew 24% in FY2025 and 61% in Q3 2026, and the product roadmap, R&D, sales and support investments are focused on them, with a dedicated AI sales team for AI Native DNEs.
- **$1M+ customers** (11% of FY2025 revenue ($96,250 thousand); 30% in Q3 2026 ($96,547 thousand)): Customers spending above $83,333 a month, rising from 24 in 2024 to 41 at the end of 2025 and 51 in Q3 2026. Million+ Dollar Customer ARR reached $386 million in Q3 2026, up 273%.
- **Developers and other** (40% of FY2025 revenue ($361,585 thousand); 31% in Q3 2026): Users spending $500 a month or less, plus legacy Bare Metal CPU offerings and miscellaneous revenue. Its share has been shrinking as DNE customers grow, and the legacy dedicated bare metal CPU offering is being wound down.
- **Gradient AI Agentic Cloud (AI/ML products)**: GPU Droplets, Bare Metal GPUs, the Gradient AI Platform and, since 2026, an inference engine and Managed Agents, Agent Droplets, Agent Learning Services and Agent Data Services. AI Customer ARR, which counts all spend of customers using at least one AI/ML offering, reached $376 million in Q3 2026 from $97 million a year earlier.

## Strategic priorities

- **Inference and agentic workloads**: The company describes two "slipstreams", token flow (Inference Engine, with token volume up about 10x quarter on quarter in Q3 2026) and agentic flow (Managed Agents, Agent Learning Services, Agent Data Services) launched at the Open Intelligence Summit on 13 October 2026.
- **Grow AI Native and Cloud Native DNE customers**: Growth rests on expanding usage of existing customers, migrations from other clouds and winning AI-native accounts such as Cursor, Ideogram and Higgsfield AI through a dedicated AI sales team.
- **Full-platform adoption**: Customers on four or more products drove roughly 60% of revenue in Q3 2026, and 90% of AI customer revenue comes from customers also using core cloud primitives; the company says it is "not a GPU landlord".
- **Raise ARR per megawatt**: Management reports $18 million ARR per average active megawatt and aims to lift it through short-term contract pricing, newer GPU generations, higher utilization and pull-through of core cloud and data services.
- **60%+ revenue growth in 2027**: In November 2026 the 2027 growth outlook was raised from 50%-plus to 60% or more, after a May 2026 target of 2027 revenue above $1.7 billion with approximately 40% adjusted EBITDA margins.

## Latest results

In Q3 2026 (quarter to 30 September 2026) revenue was $321,192 thousand, up 40% year on year and the "fifth consecutive quarter of acceleration", with a record $160 million of ARR added and ARR at $1,285 million. Gross margin fell to 50% from 60% as data center costs came ahead of revenue; operating income was $33,025 thousand (10%) and adjusted EBITDA $144,212 thousand (45%). A GAAP net loss of $42,609 thousand reflected a $57.1 million induced conversion expense on repurchasing 2030 convertible notes. Remaining performance obligations reached $800 million, from $61 million a year earlier. Guidance: Q4 2026 revenue $348-$351 million (+44-45%) and FY2026 revenue $1.208-$1.211 billion (about +34%).

## Investments and deals

- **Data center capacity**: Richmond, Kansas City and Memphis went live in 2026; just over 75 MW is active, about 75 MW more comes in 2027 and 30 MW in 2028, for about 180 MW committed. Data center leases not yet commenced total $2,955.2 million, starting November 2026 to June 2028.
- **Equity raises**: A March 2026 follow-on offering raised $887.9 million, used to repay the $500.0 million Term Loan; in July 2026 a $1.47 billion share offering funded the repurchase of $471.8 million of 2030 convertible notes. A $1 billion at-the-market equity programme was also set up.
- **Equipment financing**: An equipment lease facility with MUFG of up to $725.0 million was signed on 10 September 2026 and raised to $785.0 million on 6 October 2026; equipment financing is the primary funding strategy for 2027.
- **AI native cloud product launches**: More than 15 new products in five integrated layers were launched at Deploy in San Francisco in 2026, followed by over 70 releases in 65 business days, including 21 new models in the catalogue.
- **Partnerships**: 2025 partnerships with Laravel, fal, AMD and Persistent Systems, the latter a "multi-year, eight figure average per annum strategic partnership".

## Challenges and risks

- **Margin pressure from capacity build-out**: Gross margin fell to 50% in Q3 2026 from 60% as depreciation (+$38.2 million) and co-location (+$19.1 million) costs came ahead of revenue from new data centers.
- **Rising customer concentration**: The top 25 customers made up 27% of revenue in Q3 2026, against 11% a year earlier and 10% in FY2025.
- **Competition for customers and GPUs**: The 10-K names AWS, Azure, GCP, Coreweave, Lambda Labs and niche clouds as competitors, and reliance on a limited number of GPU suppliers, with larger cloud companies competing for supply.
- **Data center capacity and local opposition**: Under- or over-estimating data center capacity, rising energy costs and, added in the Q3 2026 10-Q, political and local opposition to data center infrastructure and export controls are stated risks.
- **Integrating a new leadership team**: A CEO, Chief Ecosystem and Growth Officer and Chief Revenue Officer were hired in 2024 and a new Chief Product and Technology Officer in 2026; the 10-K lists integrating them as a risk.

## What leadership says

> "The center of gravity in AI has shifted from people chatting with models to agents doing work, and agents need a different kind of cloud" — Paddy Srinivasan, Chief Executive Officer, DigitalOcean Q3 2026 results press release, 3 November 2026

> "We are an AI-native inference cloud, not a GPU landlord" — Paddy Srinivasan, Chief Executive Officer, DigitalOcean Q1 2026 earnings call, 5 May 2026

> "Models and GPUs are not sticky, data is." — Paddy Srinivasan, Chief Executive Officer, DigitalOcean Q1 2026 earnings call, 5 May 2026

> "We believe that we've got 2027 covered" — Matt Steinfort, Chief Financial Officer, DigitalOcean Q3 2026 earnings call, 3 November 2026, on funding


## Where a supplier can help

- **Data center build-out suppliers**: The data center capacity programme (about 180 MW committed through 2028) creates demand for power, cooling, networking, construction and facilities services that help bring capacity online faster and lift ARR per megawatt.
- **Cost control during margin compression**: With gross margin down to 50% as data center costs run ahead of revenue, tools and services that cut energy, co-location or operating costs speak directly to a stated challenge.
- **Ecosystem and technology partnerships**: The full-platform adoption priority and partnerships with Laravel, fal, AMD and Persistent Systems show room for software, model and integration partners that bring workloads onto the AI native cloud.
- **Sales and go-to-market for AI-native accounts**: The push to win AI Native DNE customers through a dedicated AI sales team and migration services fits providers of sales intelligence, prospect data and migration tooling.
- **Hiring and onboarding at scale**: R&D and G&A rose 54% and 38% in Q3 2026 on headcount, while integrating a new leadership team is a stated risk; recruiting, onboarding and HR platforms fit that growth.

_Tamtam's reading of the priorities and challenges above, not the company's own words._

## Key documents

### [DigitalOcean Holdings Annual Report on Form 10-K for fiscal year 2025](https://www.sec.gov/Archives/edgar/data/1582961/000158296126000019/docn-20251231.htm)

Annual report · 2026-02-24

- DigitalOcean Holdings (NYSE: DOCN) describes itself as an "agentic inference cloud platform" for AI and Digital Native Enterprises: IaaS (Droplets, storage, networking), PaaS/SaaS (Managed Hosting via Cloudways, Managed Databases, Managed Kubernetes, App Platform) and the DigitalOcean Gradient® AI Agentic Cloud (GPU Droplets, Bare Metal GPUs, Gradient AI Platform, Gradient AI Agents).
- FY2025 revenue $901.4 million, up 15%; gross profit 60%; income from operations $156,989 thousand (17%); net income $259,262 thousand, boosted by a $69.9 million valuation-allowance release and a $48.1 million gain on partial extinguishment of the 2026 convertible notes; adjusted EBITDA $374,805 thousand (42% margin).
- ARR reached $970 million at 31 December 2025 (from $820 million); net dollar retention rose from 98% to 100%.
- About 21,000 Digital Native Enterprise (DNE) customers (spend above $500 a month) generated 60% of revenue (55% in 2024); revenue from DNE customers grew 24%; $1M+ customers rose from 24 to 41 and generated 11% of revenue.
- One operating and reporting segment; revenue by billing address: North America 38%, Europe 28%, Asia 23%, rest of world 11%; US customers 33%. Top 25 customers made up 10% of revenue.
- Growth strategy: expand usage of existing customers, win AI Native and Cloud Native DNE customers through a dedicated AI sales team and migration services, keep a self-service funnel (sales and marketing about 9% of revenue), partnerships (Laravel, fal, AMD, a multi-year eight-figure-per-year deal with Persistent Systems) and acquisitions.
- Infrastructure: data centers in nine geographic regions; new US facilities planned in 2026 in Memphis, Richmond and Kansas City; $599.4 million of data center leases not yet commenced.
- Capital: issued $625.0 million 2030 convertible notes and drew $380.0 million Term Loan A in August 2025 to repurchase $1,187.7 million of 2026 notes; $312.3 million of 2026 notes still due within 12 months; cash $254.5 million.
- Key risks named: sustaining profitability, expanding DNE usage, AI/ML products not succeeding, competition from AWS, Azure, GCP, Coreweave, Lambda Labs and niche clouds, data center capacity planning and GPU supply from a limited number of suppliers, integrating a new leadership team.

### [DigitalOcean Holdings Quarterly Report on Form 10-Q for Q3 2026](https://www.sec.gov/Archives/edgar/data/1582961/000162828026069141/docn-20260930.htm)

Interim report · 2026-11-03

- Q3 2026 (quarter to 30 September 2026) revenue $321,192 thousand, up 40% year on year; nine-month revenue $860,281 thousand, up 31%.
- ARR $1,285 million (from $919 million); AI Customer ARR $376 million (from $97 million). AI Customer revenue counts all spend of customers that use at least one AI/ML offering.
- DNE customers generated 69% of revenue (60% a year earlier) and their revenue grew 61%; the $1M+ tier rose to 30% of revenue ($96,547 thousand) and 51 customers (from 26).
- Top 25 customers made up 27% of revenue (11% a year earlier); North America rose to 49% of revenue and the US to 45%.
- Gross margin fell to 50% from 60% as data center costs (depreciation +$38.2 million, co-location +$19.1 million) came ahead of revenue; operating income $33,025 thousand (10%); adjusted EBITDA $144,212 thousand (45%).
- GAAP net loss $42,609 thousand, driven by a $57.1 million induced conversion expense on repurchasing $471.8 million of 2030 convertible notes with a $1.47 billion share offering.
- Funding: $887.9 million March 2026 follow-on offering; $500.0 million Term Loan repaid; $725.0 million MUFG equipment lease facility (raised to $785.0 million); cash $666.1 million plus $178.6 million marketable securities.
- Commitments: $2,955.2 million of data center leases not yet commenced (November 2026 to June 2028); remaining performance obligations $799,707 thousand.

### [DigitalOcean Announces Third Quarter 2026 Financial Results (press release)](https://www.sec.gov/Archives/edgar/data/1582961/000162828026069125/a2026-q3dopressrelease.htm)

Interim report · 2026-11-03

- Q3 2026 revenue $321 million, up 40% year on year, the "fifth consecutive quarter of acceleration"; record $160 million of ARR added in the quarter.
- ARR $1,285 million (+40%); AI Customer ARR $376 million (+288%); Million+ Dollar Customer ARR $386 million (+273%).
- Adjusted EBITDA $144 million (45% margin); net loss $43 million including a $57 million induced conversion expense; adjusted free cash flow $130 million (41% margin).
- RPO $800 million, of which $345 million expected within 12 months (RPO was $61 million a year earlier).
- $100K+ customers generate 41% of revenue; their revenue grew 133%.
- Launched Managed Agents, Agent Droplets, Agent Learning Services and Agent Data Services at the first Open Intelligence Summit (13 October); Memphis data center opened ahead of schedule; 20 MW more secured for 2028, total committed capacity about 180 MW.
- Q4 2026 guidance: revenue $348-$351 million (+44-45%), adjusted EBITDA margin 42-43%. FY2026: revenue $1.208-$1.211 billion (about +34%), adjusted free cash flow margin 20-22%; 2027 revenue growth now expected at 60%+.

### [DigitalOcean Q3 2026 earnings call (transcript)](https://www.webull.com/news/15860220643023872)

Executive talk · Paddy Srinivasan, Chief Executive Officer · 2026-11-03

- CEO Paddy Srinivasan: Q3 2026 revenue $321 million, up 40%, "our highest in a decade and our fifth consecutive quarter of acceleration"; record $160 million ARR added; AI customer ARR $376 million (+288%).
- Strategy: two "slipstreams", token flow (Inference Engine; token volume up about 10x quarter on quarter) and agentic flow (Managed Agents, Agent Learning Services, Agent Data Services); 90% of AI customer revenue comes from customers also using core cloud primitives.
- Platform adoption: customers on four or more products drove roughly 60% of revenue; roughly half of unique visitors to digitalocean.com are agents and more than 37,000 customers deployed through an agent this year; Amplitude builds Amplitude Wave on DigitalOcean.
- Unit economics: $18 million ARR per average active megawatt; multiple deals "well north of 20 million in ARR per megawatt".
- Capacity: Richmond, Kansas City and Memphis live in 2026; just over 75 MW active, about 75 MW more in 2027 and 30 MW in 2028, about 180 MW committed.
- Differentiation: top 25 customers only 27% of revenue, no bare-metal long-term contracts, a "$900 million ARR core cloud business"; CFO Matt Steinfort: net leverage 1.2x, $1 billion ATM equity programme set up as "option value".
- Guidance: Q4 2026 growth 44-45%, FY2026 revenue $1.208-$1.211 billion, 2027 growth raised to 60% or more; capex per MW "mid-30s" for GB300-class gear.

### [DigitalOcean Q1 2026 earnings call (transcript)](https://www.fool.com/earnings/call-transcripts/2026/05/05/digitalocean-docn-q1-2026-earnings-transcript/)

Executive talk · Paddy Srinivasan, Chief Executive Officer · 2026-05-05

- CEO Paddy Srinivasan, Q1 2026 call (5 May 2026): revenue $258 million, up 22%, over 400 bp faster than Q4 2025's 18%; record $62 million organic incremental ARR; $1 million-plus customer ARR $183 million (+179%); AI customer ARR $170 million (+221%).
- Launched the "DigitalOcean AI native cloud" at Deploy in San Francisco: more than 15 new products in five integrated layers (infrastructure, core cloud, inference engine, data and learning, managed agents), open source options at every layer.
- New AI-native customers: Cursor (production inference, fine-tuning, core cloud), Ideogram (migrated inference from a hyperscaler), Higgsfield AI.
- Inference and core cloud are over 80% of AI customer ARR (70% in Q4); "We are an AI-native inference cloud, not a GPU landlord"; Bare Metal declining in absolute dollars.
- Raised $888 million of equity, repaid the $500 million Term Loan A, and committed 60 MW more across four locations (135 MW total) ramping in 2027.
- Guidance raised: FY2026 revenue $1.13-$1.145 billion (25-27%), exit approaching 30%; 2027 revenue above $1.7 billion (50%+), about 40% adjusted EBITDA margin.
- Pipeline "3 to 4x" available capacity; no GPU price compression, with short contracts allowing repricing; inferencing in "the top of the second inning" and agentic "just in the national anthem".

## Questions about DigitalOcean

### What does DigitalOcean do?

DigitalOcean runs a cloud platform for AI and digital native companies, combining GPU and CPU infrastructure, core cloud services, an inference engine, data services and managed agents. Customers are billed mostly on consumption.

### What is DigitalOcean's revenue?

FY2025 revenue was $901.4 million, up 15%. In Q3 2026 revenue was $321,192 thousand, up 40%, and FY2026 guidance is $1.208-$1.211 billion.

### How many employees does DigitalOcean have?

DigitalOcean had 1,462 employees at 31 December 2025, including 848 outside the United States.

### Who leads DigitalOcean?

Padmanabhan (Paddy) Srinivasan has been CEO since 12 February 2024, and W. Matthew (Matt) Steinfort is CFO.

### What is DigitalOcean's strategy?

It aims to grow AI Native and Cloud Native DNE customers by serving both inference and agentic workloads on one integrated platform, while expanding data center capacity to about 180 MW committed and targeting 60% or more revenue growth in 2027.

_Written by Tamtam from the company's own published documents listed above, read in full. Figures are as those documents state them, for the periods they name. (2026-11-17)_
