Amazon raised prices for a key AI cloud service, the latest sign of strong demand for this new technology.
Amazon Web Services recently announced price increases for EC2 Capacity Blocks for ML. This lets companies reserve GPUs in advance, so they can avoid workloads getting interrupted — a bit like booking a hotel room for a future trip and paying to ensure it’s there when you arrive.
This AWS service is popular among serious AI developers with big-budget projects, such as training huge new models, or fine-tuning foundation models.
The changes mean hourly rates for renting this type of AI compute capacity will jump by roughly 20% starting in July. AWS had already raised prices by about 15% in January.
“Amazon EC2 Capacity Blocks for ML reservation prices are updated periodically based on supply and demand,” the company said in its announcement.
Amazon stressed that this change applies to one purchasing option. Cloud customers have alternatives they can pick for AI workloads, and Amazon said it’s committed to maintaining fixed prices on those.
Amazon said the AWS pricing change shows there’s high demand for GPUs to run AI workloads. As the world’s largest cloud provider, AWS underpins many AI-powered services.
Price increases are happening in other parts of the tech industry, as tech giants pass memory price pressure on to customers. Apple raised prices this week, blaming soaring memory chip costs. Xbox did the same, and Elon Musk complained about unprecedented memory price increases.
This highlights a broader shift happening in tech. AI is increasingly constrained by physical limitations, rather than software availability. Tight memory chip supply and strong GPU demand are raising costs for cloud providers.
One of the biggest physical constraints right now is high-bandwidth memory, a critical component packaged alongside advanced AI chips. AI cloud services run on these chips and servers, so shortages and price increases like this have a big impact on data center expansion plans and, ultimately, the supply of AI.
“As there is a limit to how much memory can be produced, then there is a limit to how many GPUs can be produced, which means that there’s a limit to how many data centers can be built,” Peter Berezin, chief economist at BCA Research, wrote on X on Friday.
Berezin added that cloud providers can pass on higher infrastructure costs because customers have few alternatives when GPU capacity is scarce, giving hyperscalers AWS, Microsoft, Google, and Oracle greater pricing power.
“While the memory shortage raises their costs, it also keeps the demand for compute above the available supply, which gives them greater pricing power over access to cloud computing,” Berezin wrote on X.
The same shortages pushing up AI cloud prices have propelled memory-chip makers such as Micron and SK Hynix to records, reflecting investor expectations that AI-driven demand will keep the market tight, and prices high, for years.
Update, June 26, 2026 — This story was updated to add comments from Amazon throughout.
Have a tip? Contact this reporter via email at ekim@businessinsider.com or Signal, Telegram, or WhatsApp at 650-942-3061. Use a personal email address, a nonwork WiFi network, and a nonwork device; here’s our guide to sharing information securely.
Read the full article here
