Massive CPU Shortages in Cloud Segment Driven by Agentic AI Surge

The landscape of technology supply chains is experiencing significant turmoil, with GPUs, memory, and now CPUs facing acute shortages, all attributed to the burgeoning demand for Agentic AI.

Amazon and Cloud Providers Run Out of CPUs Amid Surging Agentic AI Demand

The meteoric rise of Agentic AI is transforming the tech industry, causing serious challenges for cloud service providers and semiconductor manufacturers who are struggling to fulfill the soaring demand. Even as the industry grapples with previous supply shortages, it now faces a critical deficit of CPUs, marking a significant shift in resource scarcity.

Dylan Patel says GPUs are no longer the biggest bottleneck. According to @dylan522p, now CPUs are the constraint. In the early AI era, CPUs were the laggers. You used them for storage, checkpointing, pre-processing, etc.(pretty light workloads) The models weren’t agentic and… pic.twitter.com/yR6g6hh74F

— Ivan Burazin (@ivanburazin) April 13, 2026

According to a report by Dylan Patel of Semianalysis, the bottleneck for cloud providers has shifted from GPUs to CPUs. Initially, GPUs were used primarily for simpler inferencing tasks. However, with the increasing complexity of AI models and their requirements, CPUs are now crucial for critical operations, including database interactions and complex simulations, which has resulted in a dramatic rise in demand for CPU resources within cloud data centers.

Traditionally, cloud servers operated with a significant imbalance—often relying on numerous GPUs for every single CPU. For instance, a common configuration might involve eight GPUs per one CPU. However, this ratio is narrowing, as both CPUs and GPUs are now being utilized more equally in AI applications, such as reinforcement learning (RL) training and agentic inference.

This surge in demand has triggered instability in platforms like GitHub, as users have begun to report frequent downtimes and failures to commit changes.

Yeah, so we’ve been like checking GitHub’s like stats on like how often is it down, how often does it fail to commit, like, you know, whatever, right? It’s terrible. And that’s because Microsoft sold all their CPUs that they had spare to other people, right? Either internal use for their lab, but, you know, not really, more so like external labs that sign deals with Entropic and OpenAI.

And so they just have like no CPUs left, right? And we’ve seen, you know, the same at many other firms, right? So whereas before you’d have like, you know, many GPU servers per CPU server. And so you’d be like, you know, 100 megawatts of GPUs would be served by even like one megawatt or less of CPUs. Nowadays, the ratio is like getting much, much closer, both for RL training and for inference, agentic inference. So then you’ve just seen everyone run out of CPUs. Amazon’s volumes on CPUs.

Dylan Patel (Semianalysis)

The driving force behind this CPU crisis appears to be the overwhelming demand that has left cloud providers like Amazon and Microsoft completely devoid of available CPU stock. The high consumption rates reflect a shift towards AI firms, such as OpenAI and Entropic, which have leveraged their relationships to secure CPU resources. Even with Amazon’s aggressive plans to triple CPU server offerings annually, they are unable to keep pace with the ever-increasing demand.

OpenAI ported their entire code base to ARM so that they can use $AMZN graviton CPUs as CPUs capacity is very short source @dylan522p pic.twitter.com/pOyole5BFQ

— MacroValue (@pradeeepk) April 14, 2026

Additionally, OpenAI’s transition from using x86 to ARM architecture, driven by the availability of Amazon Graviton CPUs, has exacerbated the situation. This migration to ARM was intended to alleviate the CPU shortfall. Unfortunately, it has led to intensified demand for ARM chips, particularly from AI firms that have shifted their codebases to capitalize on these resources.

The implications for the tech sector are stark; an acute shortage of CPUs is on the horizon. Responses from various vendors to amplify production will be necessary to meet this demand. The situation will affect not only ARM chips but also x86 processors, with AMD and Intel pressuring their supply chains to support cloud providers. NVIDIA is also stepping up its production of Vera CPU racks, which comprise multiple chips and substantial DRAM capacity. This focus on AI will likely divert resources from consumer and business markets, resulting in higher prices and limited availability for the general market.

Source & Images