Chris ChoPublished: March 28th, 2026
Amazon Web Services is pouring an unprecedented $200 billion into artificial intelligence infrastructure this year—more than $50 billion above what Wall Street expected—as the cloud computing giant races to maintain its dominance in a market increasingly defined by AI capabilities.
The massive spending spree, focused primarily on data centers and servers, comes as AWS marks its 20th anniversary and confronts mounting pressure from Microsoft and Google. The company's market share slipped from 39% in 2023 to 37.7% in 2024, according to market research firm Gartner, even as AWS generated $128.7 billion in revenue last year and continues to produce 60% of Amazon's total profits.
Matt Garman, who became AWS CEO in June 2024 after spending two decades at the company, defended the extraordinary capital expenditure during an interview at Amazon's Seattle headquarters in mid-March. He said demand for AI computing power is so intense that AWS could stay busy for the next five to 10 years even if the technology stopped advancing.
“This is not a secret, it's data centers and servers,” Garman said, addressing concerns about whether the investments will pay off. At a recent meeting with around 150 senior technology leaders, 90% raised their hands when asked if they were seeing a solidly positive return on AI investments or expected to within six months.
The stakes for AWS
AWS launched in March 2006 with a simple pitch: Amazon would handle backend technical operations so companies could focus on their products and customers. It was a risky bet for a retailer, but one that transformed Amazon into the backbone of the internet. When AWS experiences outages, major services like Netflix, Starbucks ordering systems, and even NFL operations can grind to a halt.
Now AWS faces a similar inflection point with artificial intelligence. The company has signed massive deals with AI leaders—up to $50 billion in funding for OpenAI in exchange for a $100 billion commitment over eight years to use AWS services, announced in late February. AWS also invested up to $8 billion in Anthropic, maker of the Claude AI models, with the total growing from an initial $4 billion commitment announced in September 2023.
The company is deploying more than one million Nvidia GPUs over the next 12 months, offering what it claims is the broadest range of Nvidia GPU instances among cloud providers. AWS has also developed custom AI chips like Trainium to reduce dependence on Nvidia and lower costs for customers.
More than 100,000 companies now use AWS Bedrock, the platform that allows businesses to build their own AI applications and agents without massive infrastructure investments. In February, Garman announced Bedrock AgentCore Evaluations, a tool that uses large language models to evaluate production AI agents—part of what he predicts could become AWS's next multibillion-dollar business.
“Agentic AI has the potential to be the next multibillion-dollar business for AWS,” Garman told industry analysts, referring to AI systems that can act autonomously to complete complex tasks.
Organizational upheaval
The AI push has triggered significant restructuring at AWS. In December 2025, the company merged its AI model, chip, and quantum computing teams into one unit led by Peter DeSantis, following the departure of AI head Rohit Prasad. AWS also formed a dedicated AI agents unit under Swami Sivasubramanian.
Amazon cut roughly 30,000 jobs across two rounds of layoffs—one in October 2025 and another in January 2026—moves the company says were necessary to accelerate operations as AI reshapes work. While Amazon previously said AI advancements didn't drive most cuts, CEO Andy Jassy said in June that the company will need “fewer people” as AI changes workflows.
Garman said AI coding tools are dramatically speeding up software development at AWS. Projects that would have taken two to three years are now being completed in months with smaller teams. AWS teams are “building at a rate that we haven't seen for many years,” he said.
AI is also playing a bigger role in internal operations like supply chain planning and data center resource management, according to Garman.
But James Landay, cofounder of the Stanford Institute for Human-Centered AI, has said the role AI plays in software development has in some cases been overstated, suggesting the productivity gains may not be as dramatic as tech companies claim.
The competition heats up
AWS's market share decline, even if modest, signals intensifying competition. Google Cloud has become particularly attractive to AI startups because it's slightly easier to use and offers more generous credits for young companies, according to Jacob Colker, managing director of the Seattle-based A12 Incubator.
AWS counters that it has provided more than $8 billion in credits to 350,000 new companies through its AWS Activate program. The company also says more than 65% of billion-dollar startups tracked by Pitchbook in October were building on AWS infrastructure.
“The pace of innovation, obviously, is breakneck in the world of tech, and I think that's equally true for a lot of the cloud providers,” Colker said. “The tide could shift at any moment.”
Microsoft and Google are both racing to close the gap with AWS in AI capabilities. The competition has turned cloud providers into what Gartner analyst Nicole Greene calls “digital nation states”—entities that “control enough land, power, water, and talent to actually rival countries.”
AWS's spending has become so massive that it's fueling broader concerns about an AI bubble. With every earnings call, analysts press tech companies on when the billions being poured into AI infrastructure will translate into actual products and revenue.
Garman remains confident the bets will pay off. “I'm sure they exist,” he said about signs of an AI bubble. “But I have not seen them yet.”
The company continues to expand its physical footprint. In February, AWS announced a $12 billion investment in data centers in Louisiana, expected to create 540 direct jobs and 1,700 indirect positions.
What it means for businesses
For companies considering AI adoption, AWS's infrastructure investments lower the barrier to entry. Businesses can now access powerful AI models and computing resources without building their own data centers—the same value proposition that made AWS successful with cloud computing 20 years ago.
Startups can leverage AWS Activate credits to prototype AI applications, though they should also evaluate Google's offerings for potentially simpler onboarding. Larger enterprises face pressure to integrate AI tools for operations like supply chain management and data center optimization to remain competitive as development cycles accelerate.
The layoffs at Amazon signal a broader shift in the software industry. As AI coding agents become more capable, companies will likely need fewer traditional programmers and more workers skilled in orchestrating AI systems and managing automated workflows.
For investors, AWS's $200 billion capital expenditure represents either a visionary bet on AI's future or a dangerous gamble on unproven demand. The answer will likely emerge over the next few quarters as companies report whether AI investments are generating measurable returns.
Walking through AWS headquarters on an overcast Seattle afternoon, past a barista serving special lattes celebrating the division's 20th birthday, the scale of Amazon's ambitions becomes clear. A shelf in Garman's conference area displays autographed NFL helmets—a reminder that AWS has become as critical to live sports as it is to online shopping.
When Garman joined AWS full time in 2006 as its first product manager, he struggled to explain cloud computing to his parents. His father asked if it was “like the guy who comes to my office and fixes the printer.”
Twenty years later, those same conversations are happening again—this time about artificial intelligence. And Amazon is betting $200 billion that it can shape the answer.
