Short write-up on DigitalOcean (DOCN), the cloud infrastructure company. They’re essentially an alternative for the hyperscalers, especially for smaller companies that want simplicity, transparent pricing etc.
After the Covid boom, growth was tepid and the majority of C-suite was replaced in early 2024. Revenue growth bottomed in Q1 2024 at 12% YoY and accelerated to 18% in the last quarter.
What is interesting is that new management has positioned DigitalOcean well for the AI era:
- Build out product suite for AI inference (they skipped “training” phase as the durable growth will be in inference)
- Addressed the churn of its highest customers by expanding product suite. None of its >1M $ + customers churned in 2025. (historically those customers would upgrade to the hyperscalers at some point)
- Onboarded many “AI native” companies that started generating revenue
Management has now guided to exit 2026 with 25% revenue growth and is already forecasting 2027 revenue growth of 30%. This is driven by its AI native customers that contributed 120M $ ARR, growing 150% YoY.
They’re confident in the growth outlook as it’s based on the incremental data center and GPU capacity investments already committed and coming online in 2026. (31 megawatts of new data center capacity and 3 new facilities in 2026).
The 30% growth rate is based on existing committed capacity only. It appears that they’re supply constrained, so it’s possible that they will commit to additional investments that will further accelerate its revenue growth.
EBITDA margins 40%
FCF margins around 20% (delta with EBITDA is mostly all those CAPEX investments)
Valuation = 6.6x 2026 expected revenue. If market focus shifts more towards the “AI inference era”, DOCN may become a proxy to invest along those fast growing (private) companies.
Any thoughts?
18 Likes
Yeh interesting one. I held this and advocated for Digital Ocean back in the day when it was a 30%+ grower but as growth dropped I exited.
I find the re-accel interesting but I’m not entirely convinced on the pivot for a couple of reasons… 1) Their chip asset base is or was 2 generations behind now so not sure how they can compete for frontier inference demands. 2) They never broke out of their SMB business segment even though they attracted cloud native start ups.
Interesting one to watch tho!
Ant
7 Likes
DOCN used to see their large customers graduating out of the platform..
New CEO (~2 years I believe) has been fixing this for a while and been taking very disciplined approach to investing and building both HW infrastructure and SW feature set needed to retain their largest customers.
They have been upgrading HW but using mostly AMD (some NVDA) in a cost conscious manner as they avoid training workloads arm race.
I see this somewhat similar to Cloudflare (which continued to be recognized by higher multiple awarded by the market, for their 30%+ sustained growth with security as well as now workers platform for AI natives).. which is inference focused, reinvest small % of their revenue into HW and keep building SW feature..
Old chips are not so much of an issue for inference workloads.. most overlay AI startup companies build their business on DOCN for ease of use and reasonable prices.. if they have to use public LLM in the background, they will simply attach that with API (MCP now!) calls to LLMs (ChatGPT.. Gemini.. etc.) .. and so those LLMs will still be executed on AWS, Azure, etc.
DOCN is somewhat better than NET as it is under-appreciated and does not have binary risk of security problem.. and revenue growth is accelerating.. at-least for next 6 to 8 quarters, growth should be visibly inching up with high conversion to cash
11 Likes