On September 10, DigitalOcean Holdings (NYSE:DOCN) secured a $725 million equipment finance facility to fund the GPU and CPU capacity its AI-Native Cloud business needs. The stock jumped 7.45% on the news. That reaction makes more sense once you look back a month, to August 4, when DigitalOcean reported second-quarter revenue growing 29% year over year and raised its full-year outlook on the back of accelerating AI demand. These are really two chapters of the same story.
Capacity Chasing Real Orders
DigitalOcean structured this financing deliberately. The $725 million facility, arranged with MUFG Bank as lead administrative and collateral agent alongside Axos Bank, BMO Bank and Wells Fargo as joint lead arrangers and PNC Bank as document agent, matures on September 10, 2030, and carries an accordion option for another $300 million. CFO Matt Steinfort described it as a way to align cash outflows with revenue at an attractive cost of capital, raised from a position of strength. That framing matters because DigitalOcean is not borrowing to survive; it is borrowing to keep pace with demand it already has on the books.
The numbers back that up. Remaining performance obligations reached $894 million in the second quarter, up twelvefold from a year earlier, and the weighted average contract length stretched from 1.6 years to more than three as DigitalOcean signed its first nine-figure annual commitments. Early customers of its newly launched Inference Engine pushed token consumption up roughly 30 times in just 60 days, and 85% of AI customer ARR now traces to inference and core cloud work rather than lower-margin bare metal. AI customer ARR itself grew 212% year over year, and customers spending $500,000 or more grew their revenue contribution 160% even as their headcount grew a smaller 35%.
Growth That Comes At A Cost
That growth is not free. Even as total revenue rose 29%, operating income fell 18%, and operating margin slipped to 10%, while net income dropped 4% to $35 million. Adjusted free cash flow margin also narrowed, from 26% in last year’s second quarter to 22% this year. Capacity does not build itself, and the hardware behind this expansion, GPUs and CPUs with real depreciation clocks, is now tied to a financing obligation running through September 2030 that could grow by another $300 million.
If AI infrastructure demand cools or newer chips make today’s equipment obsolete faster than expected, DigitalOcean still owes the money. The company is also leaning harder on a small number of very large customers. Accounts spending $1 million or more now generate 23% of total revenue, and revenue from that group grew 214% in the quarter. Nine-figure annual commitments from a handful of AI-Native customers are driving that shift, and any pullback from even one or two of those accounts would show up quickly in the results.
What The Market Is Pricing In
Hedge fund ownership rose from 47 funds to 53 in the most recent quarter, a sign of building institutional interest. Short interest sits at 8.76% of float, enough to reflect real skepticism without looking like a crowded short. The stock trades at 67.11 times forward earnings, as of September 17, a multiple that assumes DigitalOcean’s growth acceleration keeps going rather than plateaus. Rising fund ownership paired with a rich multiple suggests the market has already given DigitalOcean substantial credit for the AI story playing out as management describes it.
The Tension The Numbers Leave Open
DigitalOcean’s financing move and its accelerating AI metrics describe a company scaling into demand it already has, not speculative capacity for demand it hopes to attract. Margin compression and a fixed obligation running through 2030 are the price of that scaling, and they are visible in the numbers today even as growth points forward. The nine-figure commitments and inference workload gains need to keep outrunning the equipment bill for the current story to keep paying off.
READ NEXT: 10 Best Future Stocks to Buy Under $10 and 12 Best Performing Semiconductor Stocks to Invest In.
