AWS raises AI cloud prices 20% as memory chip shortage bites

Amazon Web Services has raised prices for a key AI cloud service by roughly 20%, marking the second significant increase in six months as memory chip shortages drive up infrastructure costs across the technology industry.

The price hike affects EC2 Capacity Blocks for ML, a service that allows companies to reserve graphics processing units in advance for machine learning workloads. Hourly rates for several types of cloud servers will jump starting in July, following a 15% increase AWS implemented in January.

"Amazon EC2 Capacity Blocks for ML reservation prices are updated periodically based on supply and demand," the company stated in its announcement, according to Business Insider, which first reported the pricing changes.

Why it matters

As the world's largest cloud provider, AWS underpins countless software services used by millions of developers and businesses. A 20% price increase on AI infrastructure won't stay contained to Amazon's billing statements—it will ripple through the broader technology economy as companies pass costs downstream or scale back AI ambitions. The move also confirms that AI development is increasingly constrained by physical supply chains rather than software innovation, fundamentally changing the competitive dynamics of the industry.

The memory bottleneck

The price increases reflect a critical shortage of high-bandwidth memory, a specialized component packaged alongside advanced AI chips. This memory is essential for the GPUs that power AI cloud services, creating a supply chain bottleneck that limits how quickly cloud providers can expand their data center capacity.

"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.

The shortage has created unusual pricing power for cloud providers. With GPU capacity scarce and few alternatives available, hyperscalers including AWS, Microsoft, Google, and Oracle can pass infrastructure costs to customers who have limited options.

"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 noted.

Industry-wide pressure

AWS isn't alone in raising prices due to memory costs. Apple increased prices this week, directly citing soaring memory chip expenses. Xbox implemented similar increases, and Elon Musk publicly complained about unprecedented memory price hikes.

Memory chip manufacturers like Micron and SK Hynix have seen their stock prices reach record levels as investors anticipate that AI-driven demand will keep supply tight and prices elevated for years.

The pricing pressure marks a broader shift in how AI development unfolds. Rather than software breakthroughs driving progress, physical constraints on chip production and memory supply are increasingly determining which companies can scale AI capabilities and at what cost.

Business Insider reported these details on June 26, 2026.