DigitalOcean secures $800M AI backlog, growth jumps to 29%

DigitalOcean, long known for affordable cloud services targeting developers and small teams, is undergoing a rapid shift toward enterprise-scale AI infrastructure. Preliminary second-quarter figures released this week reveal a dramatic expansion in its business model, marked by large, multi-year contracts and a surge in growth metrics. The company’s trajectory now hinges on its ability to retain high-value AI customers rather than its traditional self-serve user base.

What the numbers show

The most striking change is DigitalOcean’s contracted backlog, measured as remaining performance obligations (RPO). The company reported RPO exceeding $800 million, a more than tenfold increase from the same period a year earlier. These commitments are not only larger but also longer-term: the weighted-average contract length has extended from 1.6 years to over three years. Revenue growth for the quarter accelerated to approximately 29%, up from 14% year-over-year, while adjusted EBITDA margins and non-GAAP earnings are expected to meet or exceed prior guidance.

To support this growth, DigitalOcean has secured an additional 20 megawatts of data center capacity for late 2027 and early 2028, bringing its total committed capacity to roughly 155 MW. The company’s customer base is also evolving. While it generated around $900 million in revenue last year, a handful of new contracts—each worth $100 million or more annually—are reshaping its profile. These deals reflect a deliberate pivot away from the smallest, self-serve tiers toward higher-spending enterprise clients.

A platform, not just GPUs

DigitalOcean is positioning itself as more than a provider of raw compute resources. CEO Paddy Srinivasan emphasized the company’s "AI-Native Cloud platform" as a key differentiator, arguing that customers are drawn to its software layer—including an Inference Router that optimizes performance and cost across proprietary and open-source models. This approach aligns with a broader industry trend: as GPU capacity becomes more commoditized, providers are competing on the tools and services built around the hardware rather than the hardware itself.

The strategy carries risks. While the software layer may offer stickiness, DigitalOcean’s ability to retain nine-figure customers against hyperscale competitors remains unproven. The company’s historical strength—its simplicity and accessibility for smaller users—could also be diluted as it prioritizes enterprise clients. Meanwhile, the backlog’s concentration in a few large deals introduces new vulnerabilities. Demand from these customers can shift rapidly, and the company’s capacity commitments are locked in for years.

What’s next

The full second-quarter results will clarify whether this shift is sustainable. For now, DigitalOcean’s preliminary figures suggest a business in transition: one that has successfully attracted high-value AI contracts but must now prove it can balance enterprise demands with its developer-friendly roots. The company’s growth narrative is no longer driven by incremental self-serve signups but by a smaller number of large, lumpy deals—a model that offers visibility but also exposes it to greater volatility.