AWS took back the #1 spot in commitments

AWS and Google Cloud report their backlog. Microsoft only discloses company total figures, so Azure portion is our estimate.

Q3 backlog, change vs. Q2, and years of revenue under contract (my math):

• Amazon Web Services (AWS): $578B, +$82B, 3.1 years

• Google Cloud: $561B, +$47B, 4.6 years

• Microsoft Azure: ~$468B, flat, ~3.5 years

The total grew $129B in one quarter. AWS supplied almost 2/3 of it and moved back ahead of Google, which led in Q1 and Q2.

Microsoft’s company-wide backlog stayed roughly flat at $677B. It usually falls in July–September. CFO Amy Hood said a large Azure commitment kept the decline smaller this time.

Revenue: the first $100B+ quarter for the big three

Q3 revenue and YoY growth:

• AWS: $46.0B, +39%

• Azure: $33.2B, +48%

• Google Cloud: $30.7B, +103%

Combined: $109.9B, up from $96.4B in Q2 and 56% more than a year ago.

Over 12 months, backlog grew far faster than revenue

• AWS: backlog 2.9x vs. revenue +39%

• Google Cloud: backlog 3.6x vs. revenue +103%

• Microsoft: commercial backlog +73% vs. Microsoft Cloud revenue +30% and Azure +48%

The backlogs are not identical. Google’s includes TPU hardware sales.

AWS plans to sell its Trainium chips outside AWS from late 2027, but its current backlog only includes future-service commitments. Meanwhile, AWS highlighted that revenue commitments for Graviton and Trainium each grew >20x in Q3.

How much of the committed spend comes from AI labs?

Andy Jassy calls the AWS backlog “well diversified.” Only Microsoft has publicly put a number on it: in January, ~45% of its commercial backlog came from OpenAI.

Our Azure estimate builds on that disclosure and includes ~$281B from OpenAI.

Stay tuned for a breakdown of what is publicly known about AI-lab commitments across all three.

What this means for alliance leaders

Not all of the $1.6T is enterprise spend. Part is AI-lab contracts, and at Google, part is hardware.

Still, customers can spend part of their commits on partner software through cloud marketplaces:

• AWS and Google Cloud: up to 25% of commits, on eligible purchases

• Azure: eligible marketplace purchases count 100% toward the commit

Even after excluding AI labs, that is hundreds of billions of committed budget that partners can reach through marketplaces. It is one reason marketplace sales keep accelerating.

• Map which cloud holds each top account’s commit, and make sure your offer can draw it down before the budget is allocated.

Is your 2027 plan built around your own pipeline, or around the $1.6T that customers have already committed to clouds?