An interview with Google Cloud CEO Thomas Kurian about the unit finally turning a profit in Q1 2023, in part by making its data center operations more efficient
Jordan Novet / CNBC :
Context & Ripple Effects
When Thomas Kurian left Oracle to run Google Cloud, the unit was losing billions a year against AWS and Azure; his playbook, as profiled at the time, was classic enterprise selling — headcount, sales teams, customer trust — layered over Google's infrastructure advantage (a Bloomberg profile traced that build-out to 37,000 employees). Four years on, the bet paid its first dividend: Google Cloud posted its first-ever profitable quarter in Q1 2023, with $191M operating income on $7.45B revenue versus a $706M loss a year earlier.
This interview matters because Kurian attributes part of the swing to something unglamorous — running data centers more efficiently rather than winning dramatically more deals — which reframes cloud profitability as an operations problem, not just a scale problem.
First-order effects
- Kurian now has proof points for the two criticisms that dogged Google Cloud since his 2019 GeekWire interview — that it lacked enterprise credibility and bled cash — giving him leverage on pricing, hiring, and internal resource allocation.
- Data-center efficiency gains directly lower Google's cost per workload, letting the unit absorb discounts or invest in sales without re-entering losses.
Second-order effects
- AWS faces a rival whose margins are compounding from operations rather than price hikes; by late 2024 Google Cloud was growing 35% YoY at a 17% margin while AWS grew 19% at 38%, per CNBC — a gap that pressures AWS to defend share with its own cost story (Q3 2024 results made that comparison explicit).
- Profitability changes the internal politics: cloud stops being subsidized by ads revenue, freeing capital for the compute build-out Kurian defended in his 2024 Stratechery conversation about AI resetting competition (that Q&A framed AI as a chance to re-open the rankings).
Third-order effects
- If efficiency-driven margins hold — Google Cloud crossed $1B in quarterly operating profit by mid-2024 — hyperscale competition shifts from who spends most on capacity to who extracts most from each facility, favoring operators with the best utilization discipline.
- A consistently profitable Google Cloud makes the market structurally three-way viable, ending the assumption that only AWS could run cloud as a real business.
The trend: Hyperscale cloud is transitioning from a land-grab phase where losses were the price of share to a margin-discipline phase where data-center efficiency, not headline growth, is the durable competitive weapon.