“As we see more customers using more artificial intelligence, they actually also use more of our core AWS services. And the more core AWS services they use, they tend to generate more data, more applications, more context, and that drives more artificial intelligence use,” said AWS AI head Matt Wood. “More AI use [means] more core use, and round and round it goes.”
Amazon’s $220 billion CapEx investment is boosting AWS services sales, opening doors for new customer wins and elevating AWS to the top of the pyramid when it comes to AI power and leadership.
This is according to AWS’ AI leader Matt Wood, who explains why Amazon’s massive $220 billion spending on AI infrastructure like data centers wasn’t made “on a whim.”
“As we see more customers using more artificial intelligence, they actually also use more of our core AWS services. And the more core AWS services they use, they tend to generate more data, more applications, more context, and that drives more artificial intelligence use,” Wood’s said. “More AI use [means] more core use, and round and round it goes.”
“So we’re really not investing this amount on a whim,” Wood, AWS’ Chief AI and technology officer tells CRN.
[Related: AWS AI Chief: Why AWS Bests Microsoft And Google For Channel Partners In AI Era]
Demand for computing power around new AI applications, modern workloads and infrastructure to run AI models is significant this year, with increasing demand expected in 2027.
“This is a very, very evident demand that we see from our customers for today and for the future. So we want to ensure that there is enough capacity—whether that is training capacity for actually training these models in the first place, whether it is inference capacity, running the predictions and generations of these models, but also the agent capacity,” said Wood.
AWS generated a record $42.2 billion in revenue during the second quarter of 2026, an increase of 37 percent year over year, marking the company’s highest growth rate in 18 quarters.
AWS AI Flywheel Vision
Wood said AI agents tend to write and run a lot of their own code, with AWS seeking to supply those agents with the most powerful and efficient hardware.
“So chips like [AWS] Graviton are a really, really good place to run those agents. They are more efficient in terms of power and energy usage, and they’re a lot higher in terms of performance,” said Wood.
The AWS AI Chief said the $169 billion cloud market share leader is creating a flywheel effect by giving customers nearly everything they need to make their AI dreams come true—from leveraging data correctly on the cloud to providing agentic AI platforms like Bedrock.
“So we bring all of that together and you have: AI usage drives more agents; agents which drives more CPUs and more core AWS service usage; which tends to drive more data; more data drives more AI—and round and round it goes,” he said.
Wood said channel partners will benefit greatly from this AI and cloud spending flywheel effect.
Compute And Inference Capacity For AI Is ‘Scare’ Today
Solution provider Innovative Solutions sees Amazon’s $220 spending as a huge win for AWS partners.
“People don’t realize how scarce the compute capacity and the [AI] inference capacity is today,” said Justin Copie, CEO of Innovative Solutions, an AWS Premier Tier Services Partner.
“We see limitations in what the technology is able to do for us at a cost-effective rate because it’s so expensive. We just don’t have the compute resources that are available to do all the things that we want to do,” Copie said. “It’s still early days with the amount of capacity that’s available for what we want AI to do.”
Copie said Amazon’s big CapEx investment in creating brand new compute and inference resources will change that dynamic over the next few years.
Key to creating new capacity for hungry customers is Amazon building new data centers at a global scale.
For example, last month, Amazon unveiled plans to invest $6 billion to build a third data center campus in Louisiana. This is on top of a $12 billion investment AWS announced in February inside the state, making AWS’ total data center investment in Louisiana alone this year approximately $18 billion.
“You’re seeing Amazon sink so much capital expense into data centers today to see the ROI in two, three, maybe four years from now because they know that it’s absolutely needed,” Copie said.
Amazon’s $500 Billion Backlog; ‘We Have Direct Line Of Sight Into Demands Across AI And Our Core Business’
During Amazon’s Q2 2026 earnings report, CEO Andy Jassy said his company doesn’t have enough computing power to meet the demand in 2026.
“We will still not have enough capacity to meet all the demand we have in 2026,” Amazon CEO Jassy said. “I believe this dynamic will also be true in 2027, too. In fact, the [customer] demand we already have for 2028 is striking.”
Amazon’s backlog hit nearly $500 billion as of the end of the second quarter, growing at a triple-digit pace year over year.
Wood said it is critical that Amazon maximizes its spending now as it will lead to greater growth down the road as AI workloads ramp-up considerably.
“We have a direct line of sight into the demands across AI and our core business. And it looks remarkably healthy,” said Wood.
“So we’re able to make a very, very sizable investment in order to build out the infrastructure through 2026 in order to be able to deliver against that demand going forward,” he said.
