DOCN earnings call summary

Prepared Highlights

• Q1 2026 营收 2.58 亿美元，同比增长 22%，超出指引上限（CEO）。
• AI 客户 ARR 达 1.7 亿美元，同比增长 221%；其中推理与核心云服务占比超 80%（CEO）。
• 百万美元以上客户 ARR 达 1.83 亿美元，同比增长 179%（CEO）。
• 调整后 EBITDA 为 1.05 亿美元，利润率 41%（CFO）。
• 调整后运营收入 6400 万美元，利润率 25%（CFO）。
• 过去 12 个月调整后自由现金流 1.71 亿美元，利润率 18%（CFO）。
• 剩余履约义务（RPO）达 2.43 亿美元，同比增长 1700%（CEO）。
• 一季度新增有机 ARR 6200 万美元，创历史新高（CEO）。
• 总承诺数据中心容量达 135 兆瓦，其中新锁定约 60 兆瓦，计划于 2027 年逐步贡献收入（CEO）。
• 一季度完成 8.88 亿美元股权融资，偿还全部 5 亿美元定期贷款 A，并计划赎回 3.12 亿美元 2026 年可转换债券（CFO）。
• Richmond 数据中心于 3 月开始贡献收入，当季贡献低于 50 万美元（CFO）。

Official Guidance

• Q2 2026: 营收 2.72–2.74 亿美元（同比增长 24%–25%）；调整后 EBITDA 利润率 37%–38%（约 1.02 亿美元）；非 GAAP 摊薄每股净收益 0.20–0.23 美元（CFO）。
• FY2026: 营收 11.3–11.45 亿美元（同比增长 25%–27%）；调整后 EBITDA 利润率 37%–39%（中位数约 4.32 亿美元）；调整后自由现金流利润率 9%–12%；非 GAAP 摊薄每股净收益 1.10–1.20 美元（CFO）。
• FY2027: 营收超过 17 亿美元（同比增长 50% 或以上）；调整后 EBITDA 利润率约 40%；调整后自由现金流利润率高十位数（CFO）。
• 2026 年营收增长完全由此前承诺的 31 兆瓦容量驱动，新锁定 60 兆瓦不贡献 2026 年收入（CEO）。

Mgmt Quotes

• “Q1 revenue was $258 million up 22% year-over-year, with million dollar plus customers growing 179% year-over-year to $183 million in ARR. AI customer ARR grew 221% to $170 million, and we beat every financial target we shared in our last call.” (CEO)
• “We delivered a record $62 million in incremental organic ARR, the highest in the company's history. RPO reached $243 million, up an extraordinary 1,700% year-over-year.” (CEO)
• “We recently onboarded Cursor, one of the fastest-growing AI applications ever built, for production inference, model fine-tuning and core cloud services.” (CEO)
• “First quarter adjusted EBITDA was $105 million, up 21% year-over-year with an adjusted EBITDA margin of 41%.” (CFO)
• “We are again raising our near- and medium-term outlook. For the second quarter of 2026, we expect revenue of $272 million to $274 million, representing 24% to 25% year-over-year growth.” (CFO)

Prepared Metrics

Q&A Batch (1-5 of 10)

Q1 — William Kingsley Crane

• Topic: CPU/GPU mix shift with Agentic workloads and revenue per megawatt upside
• Key points:
  • Management does not know if CPU-to-GPU ratio will shift from 1:12 to 1:1 but expects significantly more compute overall as software becomes more Agentic.
  • Current AI revenue per megawatt is ~$13M ARR; non-Bare Metal already accounts for >80% of AI customer ARR, which should lift that metric.
  • New capabilities (serverless inferencing, intelligent routing) can detach pricing from dollars per GPU hour, enabling higher revenue and margins.
• Mgmt stance: Bullish – sees clear demand from AI-native customers moving to Agentic era; expects ARR per megawatt to increase over time.

Q2 — Gabriela Borges

• Topic: Guidance beat/raise levers, product cycle peak timing, and ARR per megawatt as key metric
• Key points:
  • Management assumes cautious expectations on facility timing, sales velocity, and pricing when guiding; Richmond data center came online in March (vs. Q2 guidance), ahead of schedule.
  • No pricing compression; H100/H200 prices are increasing.
  • Agentic workloads scale more like cloud (ramp-up gradient is steeper than traditional cloud); industry is still early (OpenAI’s agent is barely 100 days old).
• Mgmt stance: Bullish – product cycle not peaking soon; front-row seat with marquee AI-native customers gives confidence; ARR per megawatt is the best metric.

Q3 — Mark Zhang

• Topic: Non-Bare Metal ARR mix shift, customer conversion dynamics, and go-to-market strategy
• Key points:
  • Mix includes both new AI-native customers (using core AI, storage, database, compute) and existing digital-native enterprises ramping AI workloads.
  • Bare Metal ARR decreased both as a percentage and in absolute dollars; management rotates customers off bare metal at contract renewal to repurpose capacity for higher-margin serverless inferencing.
  • Go-to-market relies on a small AI-native seller team, a startup ecosystem team, and a strong product-led growth flywheel.
• Mgmt stance: Bullish – new customers come in at higher altitude (not asking for bare metal); multiple front doors for customer entry; scaling existing motion.

Q4 — Jason Ader

• Topic: Differentiation vs. neoclouds and 2027 free cash flow margin including leases
• Key points:
  • Neoclouds are training-first with small number of concentrated customers and take-or-pay contracts; DOCN focuses on inferencing and an integrated open-source stack (announced last week).
  • 2027 adjusted free cash flow margin guided mid-to-high teens; lease-adjusted FCF depends on lease terms and mix (4/5-year lease vs. upfront), so no specific guidance given.
• Mgmt stance: Bullish on differentiation (open-source, full-stack integration, customer obsession); neutral on lease-adjusted FCF – cannot provide specifics yet.

Q5 — Wamsi Mohan

• Topic: AI workload penetration in customer cohorts and graduation rates
• Key points:
  • Good mix of AI and cloud-native customers in $500K and $1M+ cohorts.
  • Management actively drives customers from $100K → $500K → $1M by identifying blockers.
  • Expects AI adoption to push graduation rates higher over the next 2 years.
• Mgmt stance: Bullish – yes to both penetration and graduation; numbers expected to keep moving up.