Amazon has raised prices on several key AI cloud products, the latest sign that a memory chip shortage is driving up the cost of some technologies.
Amazon Web Services recently announced price increases for EC2 capacity blocks for ML. A cloud service that allows companies to reserve GPUs in advance.
This change means hourly rates for renting several types of cloud servers will increase by approximately 20% starting in July. AWS had already raised prices for the service by about 15% in January.
“Reservation prices for Amazon EC2 Capacity Blocks for ML are updated periodically based on supply and demand,” the company said in an announcement. Amazon did not respond to a request for comment on Friday.
Similar price increases are occurring in other areas of the tech industry as tech giants pass on memory price pressures to customers. Apple raised prices this week, citing rising costs for memory chips. Same goes for Xbox, with Elon Musk complaining about unprecedented memory price increases.
Moving to AWS is more important than paying a few hundred dollars more for your next MacBook or gaming console. As the world’s largest cloud provider, AWS powers many software services and millions of developers rely on cloud services to deliver apps and other technology products. Price increases of 15% and now 20% could spill over into these areas in the coming quarters.
The price increase reflects broader changes in technology, with AI increasingly constrained by physical limitations rather than software availability. Tight supply of memory chips and strong demand for GPUs are driving up costs for cloud providers.
One of the biggest physical constraints right now is high-bandwidth memory, which is a critical component packaged with advanced AI chips. Because AI cloud services run on these chips and servers, these shortages and price increases will have a significant impact on data center expansion plans and ultimately the supply of AI.
“Just as there is a limit to the amount of memory that can be produced, there is also a limit to the number of GPUs that can be produced, which means there is also a limit to the number of data centers that can be built,” Peter Berezin, chief economist at BCA Research, wrote in Friday’s X.
Berezin added that hyperscalers AWS, Microsoft, Google and Oracle will be given greater pricing power as cloud providers will be able to pass on higher infrastructure costs as customers will have few alternatives when GPU capacity is in short supply.
“Memory scarcity drives up costs, while computing demand outstrips available supply, giving us greater pricing power over access to cloud computing,” Berezin wrote for X.
Similar shortages are driving up AI cloud prices, pushing memory chip makers like Micron and SK Hynix to record levels, reflecting investors’ expectations that AI-driven demand will tighten the market and keep prices low for many years.
Any tips? To contact this reporter via email, please specify the following address: ekim@businessinsider.com or on Signal, Telegram or WhatsApp 650-942-3061. Use a personal email address, non-work WiFi network, and non-work device. Here’s a guide to sharing your information securely.