DigitalOcean Holdings, Inc. Q4 FY2025 earnings call

• AI is reshaping industries, and DigitalOcean is built for this shift. They are serving AI and cloud - native companies. - Top customers are now the growth engine, with DNEs growing faster. - On the right side of software disruption driven by AI, with traction from leading AI - native companies like Character AI and Hippocratic AI. - Putting the cloud in Neo Cloud by having an integrated agentic inference cloud. - Strengthened executive team with Vinay Kumar as Chief Product and Technology Officer, and R&D team shipping products and features to help customers scale.

In Q4, revenue was $51,000,000 with 18% growth. Full - year revenue was $901,000,000 with 18% growth. Incremental organic ARR was $51,000,000 in Q4, highest in company history. Million - dollar customers had $133,000,000 in ARR, growing at 123% year - over - year. ARR from DNEs reached $640,000,000 in Q4, 62% of total ARR, growing 30% year - over - year. DNE NDR was 102%. $100,000 customers grew at 58%, $500,000 customers at 97%, million - dollar customers at 123% year - over - year. AI customer ARR in Q4 2025 was $120,000,000, growing 150% year - over - year, making up 12% of total ARR, with 70% from inference services or general - purpose cloud products rather than bare metal GPU rentals.

• First quarter 2026 revenue expected to be $249,000,000 to $250,000,000, up 18% to 19% year - over - year. Adjusted EBITDA margins expected 36% to 37%, non - GAAP diluted net income per share $0.22 to $0.27. - Full - year 2026 revenue growth 19% to 23% (21% midpoint), adjusted EBITDA margins 36% to 38%, unlevered adjusted free cash flow margins 18% to 20%, non - GAAP diluted net income per share $0.75 to $1.00. - 2027 revenue growth expected to be 30%, with projected 20% plus unlevered adjusted free cash flow margins.

Q: Talk about inference market evolution and open - source models.

A: AI - native companies use mix of open - source and closed - source models. Open - source models are 90% cheaper with comparable accuracy. 30% of traffic already served by open source.

Q: Weighted rule of 50 numbers.

A: Weighted rule of 50 is revenue growth * 1.5 + free cash flow margin * 0.5. Projected 30% revenue growth in 2027 with 20% unlevered free cash flow margins.

Q: Support for open - source models.

A: Day - zero support with automated test harness. Operational overhead manageable.

Q: Payback period on GPUs.

A: Leasing gear to align investment with revenue, earning more ARR per megawatt, paying back within months.

Q: 30% growth drivers.

A: Customer demand, capacity ramp, RPO growth, real - world inference demand.

Q: Margin progression.

A: Near - term margin compression due to capacity ramp, but adjusted EBITDA and free cash flow margins expected to improve as utilization increases.

Q: Share of wallet in AI inference compute.

A: Integrated cloud, mix of AI models, 70% of AI customer revenue from non - bare metal.

Q: Capacity and free cash flow.

A: 31 megawatts of new capacity coming online, free cash flow generation strong with proper alignment of revenue and cost.

Q: Pricing dynamic.

A: Pricing varies by stack layer. Higher up stack, pricing per token with more degrees of freedom.

Q: AI - native overlap in large customers.

A: About half of million - dollar customers are AI customers, half core cloud. NDR not including AI yet, but AI customer revenue and RPO as leading indicators.

Q: Growth drivers since joining.

A: Work to retain top customers, participation in AI - native ecosystem, inventing alongside them.

Q: Capacity commits and utilization.

A: Committed 31 megawatts, utilization depends on GPU generation and ramp, expected healthy utilization in 2027.