AWS just quietly increased EC2 Capacity Block prices – here's what you need to know

Amazon Web Services (AWS) has quietly raised the prices of EC2 Capacity Blocks for machine learning, upping them by around 15%.

First spotted by The Register, the changes mean the cost of a p5e.48xlarge instance has risen from $34.61 to $39.80 per hour across most regions, while the pricing for p5en.48xlarge has gone up from $36.18 to $41.61.

There's an even bigger increase in the US West (N. California) region, where p5e rates have risen from $43.26 to $49.75.

Amazon EC2 Capacity Blocks allow GPU capacity to be booked in advance at a locked-in price, and are mostly used by large organizations with substantial cloud budgets, generally for business-critical ML training.

This gives customers guaranteed GPU capacity at predetermined times, allowing them to reserve capacity for a fixed time window, ranging from one day to several weeks, paying an agreed rate up-front.

The blocks can be reserved up to eight weeks in advance.

"With Amazon EC2 Capacity Blocks for ML, you can reserve just the amount of accelerator capacity you need to run your machine learning workloads. EC2 Capacity Blocks pricing consists of a reservation fee and an operating system fee," AWS explained in promotional materials.

"The reservation fee is charged up front at the time you schedule the reservation. Your Capacity Block is charged at the prevailing rate at the time of purchase, even if the Capacity Block is scheduled to start after the price is updated."

Why has AWS increased prices?

Capacity Blocks have always been subject to variable pricing, based on supply and demand and adjusted every quarter based on expected market conditions.

AWS has previously reduced pricing three times: once in 2024, and twice in 2025.

While demand for compute power for AI and machine learning continues to grow, supply is limited, with a global shortage of advanced GPUs.

The change comes just a few months after AWS cut prices for some GPU instances, particularly for On-Demand and Savings Plans, by up to 45%.

The hyperscaler specifically highlighted price cuts at its annual re:Invent conference in Las Vegas in December.

Capacity Blocks were not included in those reductions, though.

"Regular price reductions on AWS services have been a standard way for AWS to pass on the economic efficiencies gained from our scale back to our customers," the company said at the time.

"These pricing updates reflect the AWS commitment to making advanced GPU computing more accessible while passing cost savings directly to customers."

The next review of pricing is set for April 2026.

A spokesperson for AWS told ITPro: “EC2 Capacity Blocks for ML pricing are dynamic and vary based on supply and demand patterns, as described on the product detail page.

“This price adjustment reflects the supply/demand patterns we expect this quarter. AWS’s commitment to not raise pricing on fixed pricing models like On Demand and Savings Plans remains unchanged.”

FOLLOW US ON SOCIAL MEDIA

Make sure to follow ITPro on Google News to keep tabs on all our latest news, analysis, and reviews.

You can also follow ITPro on LinkedIn, X, Facebook, and BlueSky.