Amazon raises prices of NVIDIA H200 GPU-based instances, ideal for AI applications, by approximately 15%

It has been reported that the prices of some instances of Amazon Web Services' (AWS) machine learning service,
Amazon EC2 Capacity Blocks for ML Pricing – AWS
https://aws.amazon.com/jp/ec2/capacityblocks/pricing/
AWS raises GPU prices 15% on a Saturday • The Register
https://www.theregister.com/2026/01/05/aws_price_increase/
According to technology media outlet The Register, the price increases apply to the 'p5e.48xlarge' and 'p5en.48xlarge' instances. The former has jumped from approximately $34.61 (approximately ¥5,415) per hour to approximately $39.80 (approximately ¥6,227) per hour in most regions, while the latter has jumped from approximately $36.18 (approximately ¥5,660) to approximately $41.61 (approximately ¥6,510). This represents a price increase of approximately 15%. Prices in the Asia Pacific (Tokyo) region were also subject to the price increase at the time of writing.
According to The Register, AWS's pricing page states, 'Current prices are scheduled to update in January 2026,' but does not specify whether this is a price increase or decrease. The current page states, 'Current prices are scheduled to update in April 2026.'

The P5e and P5en instances, which are subject to the price increase, are equipped with NVIDIA H200 Tensor Core GPUs and are marketed as being ideal for training and deploying complex, large-scale language models and diffusion models that power generative AI applications.
AWS's EC2 Capacity Blocks, which allow you to reserve usage for a specific period and pay a fixed fee, are popular with companies that want to focus on serious machine learning. The P5e and P5en models, with their powerful H200 GPUs and high pricing, are often used by teams with budgets of several hundred million yen.

An Amazon spokesperson explained, 'Prices for EC2 Capacity Blocks for ML fluctuate based on supply and demand trends as described on the product detail page. This price adjustment reflects expected supply and demand patterns for this quarter.'
Related Posts:
in Web Service, Posted by log1p_kr







