DigitalOcean (DOCN) Q2 RPO tops $800m, 10X YoY on nine-figure AI customer wins

DigitalOcean Holdings, Inc. (NYSE: DOCN) shares climbed 9.3 percent on July 8, 2026 to close at 143.64 dollars, touching an intraday high of 157.99 dollars on volume of 3.28 million shares, after the Broomfield, Colorado-based cloud infrastructure company disclosed preliminary second-quarter 2026 metrics on July 7, 2026 that materially reset expectations for both current-quarter delivery and forward-year growth trajectory. Remaining performance obligations are expected to exceed 800 million dollars for the second quarter, up more than tenfold from approximately 53 million dollars in the year-ago period, with the weighted average life of contracted revenue extending from 1.6 years to more than three years, reflecting durable commercial commitment from an increasingly sophisticated AI customer base. Revenue growth is expected to accelerate to 29 percent versus 14 percent in the year-ago quarter, and the implied Q2 revenue of approximately 283 million dollars sits about 9 million dollars above the previous high end of the 272 to 274 million dollar range that DigitalOcean Holdings had provided investors in May.

The company also confirmed that adjusted earnings before interest, taxes, depreciation, and amortisation margin and non-GAAP diluted earnings per share are expected at or above the high end of prior guidance, and management expects positive customer momentum to lift the previously communicated exit 2026 revenue growth rate materially higher. Chief executive officer Paddy Srinivasan attributed the acceleration to multiple nine-figure annual customer commitments added during the quarter for inference and cloud products, and highlighted the Inference Router capability that balances price and performance across both closed and open source models as the software advantage that distinguishes DigitalOcean Holdings from what he characterised as bare-metal GPU rental companies.

For a mid-cap cloud infrastructure operator whose market capitalisation had languished at approximately 16.1 billion dollars while the mega-cap hyperscaler AI narrative captured the majority of AI infrastructure investor attention, the preliminary Q2 disclosure represents one of the clearest datapoints available on how AI-Native Cloud pricing and platform economics are materialising below the hyperscaler tier.

What does the $800 million RPO forecast actually change for DigitalOcean’s revenue predictability

Remaining performance obligations are contracted future revenue that has not yet been recognised on the income statement, and they represent one of the cleanest available indicators of forward revenue visibility for any recurring-revenue business. DigitalOcean Holdings has moved from approximately 53 million dollars of RPO at the end of the second quarter of 2025 to more than 800 million dollars projected at the end of the second quarter of 2026, an approximately 15-times expansion that fundamentally changes how the equity should be valued for revenue predictability. The company is no longer a consumption-billed cloud infrastructure operator with limited forward visibility. It is a company where AI-Native workloads are being committed on multi-year contracts by sophisticated enterprise customers who now underwrite the majority of near-term growth trajectory.

The weighted average life of the RPO base is at least as informative as the total dollar amount. Moving from 1.6 years to more than three years means the contracted revenue base is not just larger but also more durable, and the incremental contract terms being signed today are longer than the average contract terms that had been carrying the business through the developer-oriented phase of DigitalOcean Holdings’ commercial evolution. Enterprise AI infrastructure customers are willing to commit to three-plus year relationships because they are architecting production workloads on the platform, not experimenting with it, and that structural shift changes the appropriate valuation multiple that should be applied to the equity’s forward revenue.

The read-across to sell-side estimate revisions should be substantial and rapid. Consensus estimates for the second quarter had clustered near 274 million dollars in revenue, approximately 38 percent adjusted EBITDA margin, and around 0.25 dollars adjusted earnings per share. The preliminary disclosure exceeds all three metrics simultaneously, and the RPO expansion signals that the beat is not a single-quarter event but the surface manifestation of a durable commercial trajectory. Analysts covering the stock will need to raise both current-year and next-year revenue estimates, and the implied fair value framework built on those higher estimates supports a materially higher share price than the current level.

Why is the 10X year-over-year RPO expansion the analytically important number in this preliminary release

A 10-times year-over-year expansion in remaining performance obligations is an unusually large move by any measure, and it is particularly notable in the context of a company that had been perceived as a developer-first infrastructure operator competing at the SMB and startup tier of the cloud market. The magnitude of the RPO expansion reflects two structurally distinct forces working in the same direction. First, the AI-Native Cloud repositioning has attracted a fundamentally different customer profile than DigitalOcean Holdings had been serving through most of its history. Second, those new customers are signing contracts at scales and durations that the company had rarely seen previously.

The composition of the RPO growth appears to be driven by a small number of very large deals rather than a broad expansion across the existing customer base. Management specifically referenced multiple nine-figure annual customer commitments added during the quarter for inference and cloud products, and those individual contract sizes are meaningful contributors to the total RPO number. A nine-figure annual commitment implies a minimum 100 million dollar annual value per contract, and multiple such contracts signed in a single quarter effectively account for the majority of the RPO expansion when compounded by the multi-year contract terms.

The strategic implication is that DigitalOcean Holdings has moved into a genuinely different competitive position in the AI infrastructure market. The company is no longer just serving the developer and SMB tier where AWS, Google Cloud, and Microsoft Azure have not focused their AI infrastructure attention. It is now competing for and winning enterprise workloads at scales that overlap with what the hyperscalers, CoreWeave Inc., Vast.ai, and specialised AI infrastructure providers are targeting. Whether this positioning is durable will be the defining commercial question through the balance of 2026 and into 2027, and the RPO trajectory is the leading indicator that will answer it.

How does the multiple nine-figure customer commitments confirm the AI-Native Cloud market thesis

The customer commitments disclosed by DigitalOcean Holdings are structurally the most important data point in the preliminary release, because they confirm that the AI-Native Cloud thesis is actually being validated by paying customers rather than by capacity commitments waiting for demand. Multiple nine-figure annual customer commitments for inference and cloud products signed in a single quarter is a concentration of enterprise commercial commitment that had not been previously visible in the DigitalOcean Holdings customer profile. Those customers are effectively underwriting the platform economics against multi-year workload trajectories.

The specific pattern of customers seeking DigitalOcean Holdings’ total cost of ownership advantage is meaningful. Chief executive officer Paddy Srinivasan emphasised that customers appreciate the ease of scaling their businesses and seek the total cost of ownership advantage that the platform provides. That is a specific commercial argument that positions DigitalOcean Holdings against alternatives where the compute cost is either higher, the platform is harder to integrate with production workflows, or the pricing model is less predictable at scale. Bare-metal GPU rental providers command lower base rates for compute time but typically require substantial additional integration, software, and operational investment on the customer side.

The AI-Native Cloud thesis rests on the argument that inference and agentic workloads require a fundamentally different infrastructure profile than the training workloads that dominated the previous phase of AI infrastructure investment. Inference workloads run continuously, serve production applications, and require both cost and performance optimisation at scale. The Inference Router capability that DigitalOcean Holdings has been building addresses this specific requirement by balancing price and performance across closed and open source models, and it is the type of software layer that enterprise AI customers increasingly need but do not want to build in-house. The multiple nine-figure customer commitments confirm that this software differentiation is generating measurable commercial value.

What role does the Inference Router play in DigitalOcean’s differentiation from bare-metal GPU providers

The Inference Router is DigitalOcean Holdings’ answer to one of the specific operational challenges that enterprise AI customers face at scale. Running production inference workloads across multiple foundation models, whether closed proprietary models from OpenAI, Anthropic, and other frontier developers or open source models from Meta Platforms Inc., Mistral AI, and the broader open source ecosystem, requires intelligent routing of individual queries to the model that offers the best combination of accuracy, latency, and cost for the specific use case. Building that routing logic in-house is possible but resource-intensive, and DigitalOcean Holdings’ platform delivers it as an integrated capability.

The competitive positioning against bare-metal GPU rental providers is deliberate and strategically important. CoreWeave Inc., Vast.ai, and other providers offer raw GPU compute at attractive pricing but leave the software integration, orchestration, and workload management responsibility with the customer. That trade-off works for sophisticated AI development teams building custom infrastructure, but it does not work for enterprise customers deploying production AI applications who need predictable cost and performance without the operational burden of building their own AI infrastructure stack. DigitalOcean Holdings is explicitly targeting the enterprise segment that wants the AI-Native Cloud outcome rather than the bare-metal GPU input.

The strategic implication for the AI infrastructure market is that different customer segments will pay different premiums for different levels of platform integration. The lowest-cost GPU rental options serve highly sophisticated buyers who can build their own inference infrastructure. Enterprise customers who need production AI applications delivered at scale are willing to pay platform premiums for the operational simplicity, integration, and reliability that AI-Native Cloud providers deliver. DigitalOcean Holdings is positioning itself as the primary alternative to hyperscaler AI-Native cloud offerings at a materially better total cost of ownership, and the nine-figure customer commitments suggest the market is prepared to pay for that positioning.

Why does the 20 MW additional committed data center capacity extend the growth runway into 2027 and 2028

DigitalOcean Holdings signed an additional 20 megawatts of committed data center capacity for 2027 and 2028 during the second quarter, bringing total committed data center capacity to approximately 155 megawatts across the platform. That capacity commitment is analytically important because it extends the operational runway for revenue growth well beyond the immediate 2026 fiscal year and provides the physical infrastructure basis for continuing to serve enterprise AI workloads at scale. The 155 megawatt total committed footprint places DigitalOcean Holdings in a substantial mid-tier of the AI infrastructure market and gives management the physical capacity to convert the RPO backlog into recognised revenue over the coming years.

The composition of the capacity commitment reflects a specific investment strategy. Rather than building 100 percent of capacity through direct ownership and construction, DigitalOcean Holdings has been contracting for capacity through partnership arrangements and colocation providers, which allows the company to scale capacity in step with committed customer demand without the balance sheet impact of large greenfield data center construction. That capital-efficient scaling model is one of the operational differentiators that allows a mid-cap infrastructure operator to compete against hyperscalers on capacity without needing hyperscaler-scale capital investment.

The near-term implication of the 2027 and 2028 capacity commitments is that DigitalOcean Holdings has extended the growth runway well beyond the current fiscal year. The exit 2026 growth rate that management now expects to lift will be supported by capacity coming online through 2027 and 2028, and the RPO base that will convert into recognised revenue over the next three-plus years is now backed by contracted infrastructure. That combination of committed customer demand and committed infrastructure capacity is the analytical basis for the sell-side and buy-side expectations that DigitalOcean Holdings can sustain the 26 percent projected fiscal 2026 revenue growth and accelerate to more than 50 percent growth in fiscal 2027.

How does DigitalOcean’s positioning compete against Cloudflare, Snowflake, CoreWeave, and hyperscalers

The competitive read-through from the preliminary Q2 disclosure was visible in relative share price performance on the announcement day. DigitalOcean Holdings shares climbed 9.3 percent to 143.64 dollars while Cloudflare Inc. and Snowflake Inc. shares moved substantially less on the day, and CoreWeave Inc. shares declined. That relative move pattern reflects how the market is now beginning to differentiate between different positioning within the AI infrastructure category, and DigitalOcean Holdings’ AI-Native Cloud approach is being rewarded relative to alternatives that are either more mature and slower-growing or that face specific competitive headwinds.

Cloudflare Inc. serves a partially overlapping developer and enterprise infrastructure market with a different product architecture focused on edge computing, security, and application delivery. Snowflake Inc. is anchored on the data platform layer and increasingly extends into AI and analytics workloads. Both companies compete for enterprise cloud spend but neither is precisely positioned in the inference-first AI-Native Cloud category that DigitalOcean Holdings is now defining. The relative share price stability of Cloudflare Inc. and Snowflake Inc. on the DigitalOcean Holdings news reflects that market participants view them as different competitive stories rather than direct comparables.

CoreWeave Inc.’s decline on the DigitalOcean Holdings news is more analytically significant. CoreWeave Inc. has been the leading pure-play AI infrastructure provider serving hyperscale AI workloads including OpenAI, and its business model is anchored on GPU-heavy compute delivered at scale to large frontier AI customers. DigitalOcean Holdings’ explicit positioning against bare-metal GPU rental providers, combined with its multiple nine-figure customer commitments, suggests that some enterprise AI customers are choosing the AI-Native Cloud platform over pure GPU rental alternatives. That commercial shift, if it continues, would compress the addressable market for CoreWeave Inc. among enterprise customers and force the company to either compete on platform capabilities or accept a narrower hyperscale customer focus.

What are the execution and capacity risks that could complicate the raised exit 2026 growth rate

The primary execution risk is capacity delivery. DigitalOcean Holdings has committed to bringing 20 megawatts of additional capacity online in late 2027 and early 2028 and to serving the existing 800 million dollar RPO base through the platform’s current capacity envelope. Any friction in the data center construction, commissioning, or partner contracting timelines could delay the conversion of RPO to recognised revenue, and specialty data center construction has historically been subject to permitting, power interconnection, and cooling infrastructure delays that can compress delivery windows. The sequential capacity build-out from approximately 135 megawatts today to 155 megawatts by 2028 is aggressive but not unprecedented in the AI infrastructure category.

The commercial concentration risk is that a small number of very large customers now account for a disproportionate share of the RPO base and the incremental revenue growth. Multiple nine-figure customer commitments are strategically valuable but they also increase counterparty concentration and correlation risk. If any single large customer chooses to reduce its commitment, migrate workloads to a hyperscaler alternative, or pursue in-house infrastructure development, the impact on DigitalOcean Holdings’ revenue trajectory would be materially larger than any single customer loss would have been in the previous commercial phase of the business. Managing that concentration risk while continuing to build the large enterprise pipeline will be a strategic focus through 2027.

The competitive response risk is that the hyperscalers, CoreWeave Inc., and specialised AI infrastructure providers will respond to the DigitalOcean Holdings commercial momentum with more aggressive pricing, platform investment, and enterprise sales focus. Amazon Web Services, Microsoft Azure, and Google Cloud each carry commercial and technical resources that dwarf DigitalOcean Holdings, and their willingness to compete for enterprise AI-Native workloads with more integrated offerings could compress the price and margin premium that DigitalOcean Holdings is currently earning. Sustaining the software differentiation advantage represented by the Inference Router and adjacent capabilities will require continued product investment against those larger competitors.

Key takeaways on what the DOCN preliminary Q2 signals for AI-Native Cloud infrastructure investors

• DigitalOcean Holdings, Inc. disclosed on July 7, 2026 that second-quarter 2026 remaining performance obligations are expected to exceed 800 million dollars, up more than tenfold from approximately 53 million dollars in the year-ago quarter, with weighted average contract life extending from 1.6 years to more than three years.
• Revenue growth is expected to accelerate to 29 percent from 14 percent in the year-ago quarter, and implied Q2 revenue of approximately 283 million dollars sits about 9 million dollars above the previous high end of the 272 to 274 million dollar range provided in May.
• Adjusted EBITDA margin and non-GAAP diluted earnings per share are expected at or above the high end of prior guidance, and management expects positive customer momentum to lift the previously communicated exit 2026 revenue growth rate materially higher.
• Chief executive officer Paddy Srinivasan attributed the acceleration to multiple nine-figure annual customer commitments added during the quarter for inference and cloud products, and highlighted the Inference Router capability as the software differentiation against bare-metal GPU rental providers.
• DigitalOcean Holdings signed an additional 20 megawatts of committed data center capacity for 2027 and 2028, bringing total committed data center capacity to approximately 155 megawatts, providing the physical infrastructure basis to convert RPO backlog into recognised revenue.
• Shares climbed 9.3 percent to 143.64 dollars on July 8, 2026 with an intraday high of 157.99 dollars, valuing DigitalOcean Holdings near 16.1 billion dollars, and Cloudflare Inc. and Snowflake Inc. moved substantially less on the day while CoreWeave Inc. declined.
• The AI-Native Cloud positioning against bare-metal GPU rental providers is generating measurable commercial value, and the multiple nine-figure customer commitments confirm that enterprise AI customers are willing to pay premium pricing for platform integration, inference routing, and operational simplicity.
• Consensus analyst expectations of 26 percent fiscal 2026 revenue growth and more than 50 percent fiscal 2027 revenue growth are underwritten by the RPO base and the committed capacity trajectory, supporting a longer-term market capitalisation target that analysts have projected at approximately 35 billion dollars by 2030.
• Execution risks include capacity delivery timelines through 2027 and 2028, customer concentration risk associated with multiple nine-figure commitments, and competitive response from Amazon Web Services, Microsoft Azure, Google Cloud, CoreWeave Inc., and specialised AI infrastructure providers.
• The DigitalOcean Holdings preliminary Q2 disclosure provides one of the cleanest available datapoints on how AI-Native Cloud pricing and platform economics are materialising below the hyperscaler tier, and it validates the thesis that inference and agentic workloads require a fundamentally different infrastructure profile from training workloads that dominated the previous AI infrastructure investment phase.