Amazon CEO Andy Jassy has said that AWS's large committed cloud capacity is largely sold out through 2027, with a good chunk of 2028 already allocated, while Linthicum distinguished pre-reserved capacity from what enterprises can allocate day to day and developers reported constraints on ordinary instances.
Jassy made the remarks in a video podcast with Andreessen Horowitz co-founder Ben Horowitz and general partner Erik Torenberg, reported by Reuters on October 23. He said capacity for 2026 is already gone and that the company could invest well beyond the roughly $220 billion in capital expenditure planned for this year without it being a poor use of capital.
"2027 is largely sold out. We always leave a certain amount for on demand and our inference service bedrock, or EC2, but largely sold out all the big chunks. A good chunk of 2028 is sold out," Jassy said.
AWS sells capacity assurance through On-Demand Capacity Reservations and Capacity Blocks for ML, where a customer reserves instances in a named Availability Zone. In January, AWS raised Capacity Blocks for ML rates by about 15% uniformly across every region.
Cloud analyst David Linthicum argued that the reporting overstated the position, drawing a distinction between the pre-reserved blocks Jassy described and what an enterprise can allocate day to day. "Of course, I suspect I can still go into the AWS console today and allocate GPU capacity for AI workloads. Maybe not the exact massive, pre-reserved, hyperscale capacity block some large model provider wants for the next three years, but that's not exactly the same thing as 'AWS is sold out,'" Linthicum said. He added that the market already carries enough confusion around reserved instances, on-demand services and committed capacity.
Discussion on Reddit turned to current availability rather than the 2027 commitments. A developer posting as cbigsby wrote: "We've been seen a big rise in compute capacity issues the past month or so at work, just for regular ec2 instances. Forget spot instances we have trouble getting enough on demand. Luckily mostly tied to CI but there have been a few days where the production k8s clusters have been concerningly constrained."
Similar reports appeared elsewhere. Richard T., a technical operations manager, commented that Microsoft has released only limited information about Azure East US capacity constraints, and that his team adopted reserved instances to secure capacity. A commenter in a parallel thread reported trouble obtaining sufficient compute from any of the three major providers, and said the constraint was not limited to GPUs.
Ian Rae, replying to Linthicum, pointed to an architectural consequence: the next thing you know, folks reserve cloud resources they are worried they might need, which is an anti-cloud pattern. AWS already ships tooling for the problem Rae describes: EC2 Capacity Manager aggregates capacity data into a cross-account, cross-region dashboard and reports the ratio of used to unused reserved capacity.
The $220 billion figure covers total capital expenditure rather than AI spending alone, a distinction raised by commenters. Amazon reports third-quarter results on October 29, where its remaining performance obligations disclosure will show how much of the committed capacity is contracted revenue.
Analysis
Why This Matters
- Cloud capacity constraints increasingly affect enterprise customers, who may face delays or higher costs for compute resources.
- The AI boom is driving unprecedented demand for GPU instances, pushing providers to invest in data centers and manage capacity tightly.
- If major cloud providers are sold out on large committed blocks, it could slow AI startups and large-scale model training.
Background
Andy Jassy’s comments come as Amazon plans roughly $220 billion in capital expenditure this year, much of it for AWS data centers to meet AI demand. AWS sells reserved capacity through Capacity Blocks for ML and On-Demand Capacity Reservations, giving customers guaranteed compute. Analysts and developers have reported growing difficulty securing both GPU and regular instances from multiple cloud providers.
Key Perspectives
Andy Jassy (Amazon CEO): Says AWS's large committed capacity is sold out through 2027, with a chunk of 2028 already allocated, while the company leaves headroom for on-demand and inference services.
David Linthicum (cloud analyst): Argues the reporting overstates the position; ordinary on-demand GPU capacity is still available in the console, and the scarcity applies to massive pre-reserved hyperscale blocks, not day-to-day use.
Developers (Reddit and comment threads): Report rising compute capacity issues for regular EC2 instances, including difficulty getting on-demand and spot instances, with constraints on production clusters.
What to Watch
- Amazon’s Q3 earnings on October 29: the remaining performance obligations disclosure will reveal how much committed capacity is under contract.
- Whether AWS raises prices further on Capacity Blocks or introduces new reservation options to manage demand.
- Signs of capacity crunches in other cloud providers (Azure, GCP) and whether enterprise customers shift deployment strategies.
Checked
12 claims in this story checked against 1 source: 12 supported.
Show each claim
- SupportedJassy said AWS’s large committed cloud capacity is largely sold out through 2027.InfoQ: “Amazon CEO Andy Jassy recently said that AWS’s large committed cloud capacity is largely sold out through 2027”
- SupportedA good chunk of AWS capacity for 2028 is already allocated.InfoQ: “with a good chunk of 2028 already allocated”
- SupportedLinthicum distinguished Jassy’s pre-reserved capacity from what an enterprise can allocate day to day.InfoQ: “drawing a distinction between the pre-reserved blocks Jassy described and what an enterprise can allocate day to day”
- SupportedA developer reported constraints on regular EC2 instances, including difficulty getting enough on-demand capacity.InfoQ: “We’ve been seen a big rise in compute capacity issues the past month or so at work, just for regular ec2 instances. Forget spot instances we have trouble getting enough on demand.”
- SupportedJassy made the remarks in a video podcast with Ben Horowitz and Erik Torenberg, reported by Reuters on October 23.InfoQ: “Jassy made the remarks in a video podcast with Andreessen Horowitz co-founder Ben Horowitz and general partner Erik Torenberg, reported by Reuters on October 23.”
- SupportedJassy said capacity for 2026 is already gone.InfoQ: “Capacity for 2026 is already gone, Jassy said”
- SupportedJassy said Amazon could invest well beyond roughly $220 billion in planned capital expenditure for this year without it being a poor use of capital.InfoQ: “the company could invest well beyond the roughly $220 billion in capital expenditure planned for this year without it being a poor use of capital.”
- SupportedJassy said AWS leaves some capacity for on-demand services, Bedrock and EC2 while the big chunks are largely sold out.InfoQ: “We always leave a certain amount for on demand and our inference service bedrock, or EC2, but largely sold out all the big chunks.”
- SupportedAWS offers capacity assurance through On-Demand Capacity Reservations and Capacity Blocks for ML, with instances reserved in a named Availability Zone.InfoQ: “AWS sells capacity assurance through On-Demand Capacity Reservations and Capacity Blocks for ML, where a customer reserves instances in a named Availability Zone”
- SupportedAWS raised Capacity Blocks for ML rates by about 15% uniformly across every region in January.InfoQ: “AWS raised Capacity Blocks for ML rates by about 15% uniformly across every region in January”
- SupportedRichard T. said his team adopted reserved instances to secure capacity, and said Microsoft had released only limited information about Azure East US capacity constraints.InfoQ: “Richard T., a technical operations manager, commented on Linthicum’s post that Microsoft has released only limited information about Azure East US capacity constraints, and that his team adopted reserved instances to secure capacity against projected growth.”
- SupportedAmazon’s October 29 third-quarter results are expected to include an RPO disclosure showing how much of the committed capacity is contracted revenue.InfoQ: “Amazon reports third-quarter results on October 29, where its remaining performance obligations disclosure will show how much of the committed capacity is contracted revenue.”