AI inference infrastructure demand surge: DigitalOcean RPO 10x, AI ARR 221%, inference shift from training
Key Questions
What signals surging AI inference demand at DigitalOcean?
DigitalOcean's Q2 preview shows RPO up 10x to $800M, contract durations doubling to 3+ years, and AI ARR growing 221% with 155MW capacity expansion.
How does inference demand validate infrastructure buildout?
The shift from training to inference at enterprise level supports continued hyperscaler and cloud spending beyond initial training phases.
What profitability concerns accompany the growth?
Despite 171% YTD stock gains, EPS dropping 56% raises questions on long-term margins amid heavy capacity investments.
What is the inference cost paradox in enterprise AI?
Token prices continue to plummet while enterprise inference bills soar, with Gartner forecasting $20.6B in spend and examples like Uber budget blowouts.
Why is inference seen as the next phase after training?
Enterprise spend is moving from model training to ongoing inference workloads, confirming sustained demand for AI infrastructure capacity.
DigitalOcean Q2 preview shows AI inference demand surging: RPO up 10x to $800M, contract duration doubling to 3+ years, AI ARR up 221%. 155MW capacity expansion underway. 171% YTD stock gain. EPS dropping 56% raises profitability concerns. This signals enterprise AI spend shifting from training to inference, validating infrastructure buildout narrative beyond hyperscalers. New data: inference cost paradox—token prices plummeting but enterprise bills soaring; Gartner $20.6B inference spend forecast; Uber budget blowout.