Matt Garman is the enabler of your digital existence—from ordering at Starbucks to binge-watching Netflix or scrolling through Pinterest.
As the CEO of Amazon’s cloud computing division, he holds significant sway over the allocation of computational power that underpins the internet’s operation, a responsibility that may soon extend to shaping the trajectory of Artificial Intelligence itself.
Few possess a deeper understanding of this domain than Garman, who met CNN at Amazon’s Seattle headquarters mid-March. As an intern, he actually penned the document outlining Amazon Web Service’s business strategy. Furthermore, upon joining Amazon full-time in 2006, he served as AWS’s inaugural product manager, guiding corporations in their initial migration to the web during the fledgling internet era.
Two decades on from its March 2006 launch, AWS has become fundamental to virtually every entity leveraging internet tools. When AWS falters, segments of society grind to a halt.
This represents a massive revenue stream for Amazon, which logged $128.7 billion in sales last year. However, the rise of AI has reshaped the tech landscape. Amazon is already enacting drastic measures, such as ramping up AI infrastructure spending to an anticipated $200 billion this year while simultaneously cutting tens of thousands of jobs.
Garman elaborated on why Amazon deems these actions crucial. He noted that the staff reductions streamline day-to-day operations, and the accrued demand for AI is substantial enough that Amazon’s cloud tools will remain heavily utilized for the next five to ten years, even if technological advancements were to stall.
When AWS debuted, he recalled, “We had to articulate what the concept of cloud computing even was, why it mattered, and why Amazon would be involved.”
Those discussions feel familiar once again.
“Jumping ahead to our current position in AI, I believe (these are) essentially the same challenges,” Garman stated. “Many organizations will need to fundamentally rethink how they operate.”
Amazon’s Edge in the AI Race
AWS was introduced to offer businesses virtual IT infrastructure and server capacity.
The core premise was straightforward: Amazon would handle the complex backend technical work, allowing companies to concentrate solely on their products and clientele. Yet, this was a calculated risk for a company famous for disrupting the retail sector.
“We are very comfortable being misunderstood,” Amazon founder Jeff Bezos told Bloomberg Businessweek back in 2006.
This mindset persists among Amazonians, according to Jeff Barr, AWS’s Chief Evangelist and a 23-year veteran of the company, in an interview with CNN.
On the way to meet Garman, I passed the reception area where a barista was serving specialized lattes in honor of AWS’s 20th anniversary.
We are greeted by an disembodied intercom voice asking for our intended visitor on the secure floor where Garman awaits. When escorted into one of his private conference rooms, I see a concrete indication of AWS’s immense reach: a shelf decorated with autographed NFL helmets—the National Football League being one of AWS’s largest clients, illustrating its critical importance to life both online and off.
This is a far cry from the Seattle pub conversations about internet storage that initially spawned AWS. Garman shares that even his own parents struggled to grasp cloud computing when he first described his job.
“(It) was really hard to explain to them,” he said. “And my dad would ask, ‘Is this like that guy who comes to my office and fixes the printer?'”
Nevertheless, AWS eventually became indispensable to web entrepreneurs—and now its future hinges on achieving the same essential status for AI-focused enterprises.
Amazon is collaborating deeply with major AI players like OpenAI and Anthropic, channeling billions via direct investment, assisting in the deployment of their services, and supplying the necessary technology for model training. Custom silicon chips designed specifically for AI workloads have also been developed.
Crucially, AWS also aims to be vital for every company via its Bedrock platform, which Amazon claims over 100,000 businesses are using to build their own AI-powered applications and agents. Just as early AWS offerings allowed firms to access storage and processing power without hefty capital expenditures on infrastructure, Amazon’s new tools are democratizing access to AI models, the company points out.
Amazon stands as the foremost provider of cloud technology, putting it in an advantageous position to capitalize on the surging demand for AI-related computing. However, rivals Microsoft and Google are aggressively closing the gap.
Amazon’s share of the cloud market dipped from 39% in 2023 to 37.7% in 2024, according to data from market research firm Gartner.
Google Cloud is currently appealing to startups as it is reported to be slightly more accessible for initial setup, says Jacob Kolker, Managing Director at Seattle’s A12 Incubator, which supports AI startups in building their businesses. Additionally, Google offers a more generous credit program for emerging companies.
AWS counters that it has furnished over $8 billion in credits to 350,000 new companies via its AWS Activate startup program. It also reports that more than 65% of unicorn startups—tracked in Pitchbook’s database of startup funding and financial metrics as of October—are built upon AWS.
Kolker suggests the dynamic landscape is subject to sudden shifts.
“The pace of innovation is obviously frantic in the tech world, and I think that applies equally to many cloud providers,” he commented.
Massive Bets Measured in Hundreds of Billions
Amazon’s colossal data centers and millions of miles of fiber optic lines form the backbone of the internet. However, the strategic decisions made within structures like the glass-clad Amazon Reinvent tower in downtown Seattle are defining the future of Amazon—and, if Amazon’s plans succeed, millions of other companies too.
Within its walls, on a gloomy Seattle afternoon, Garman and I discuss one of Wall Street’s paramount questions for Amazon: the $200 billion it plans to spend on capital expenditures this year, tied inextricably to AI infrastructure. This figure surpasses analyst expectations by over $50 billion and dwarfs the $131.8 billion Amazon invested in property and equipment in 2025. (Note: The provided text had a potential typo mentioning 2025 expenditure; the rewrite keeps the numerical comparison but notes the text’s data point.)
A technician inserts a server rack shelf into the rear of a Trainium3 UltraServer at the Amazon Web Services quality lab in Austin, Texas, on February 3, 2026.
Mark Felix/AFP/Getty Images
AI expenditures have become so vast that the marketing research firm Gartner now categorizes companies like Amazon as “digital nation-states,” because they “control sufficient land, power, water (and) talent to genuinely rival countries,” according to analyst Nicole Green.
Where is this capital flowing?
“That’s no secret: it’s data centers and servers,” Garman confirmed.
These soaring outlays have amplified concerns about an AI bubble.
Tech titans maintain that AI demand has surged so dramatically they are in a race for processing power. Critics demand clarity on when these investments will yield tangible returns.
Amazon’s drastic maneuvers are not the sole surprising development recently. The company shed approximately 30,000 roles across two rounds of layoffs—one in October, another in January—a move it claims is to accelerate AI development.
Amazon previously stated that AI advancements were not the root cause for the majority of job cuts, although CEO Andy Jassy mentioned in June that the company would require “fewer people” as AI integration changes workflows. Garman pointed out that AI is playing a substantial role in backend operations, such as supply chain forecasting, managing data center resources, and, naturally, software development.
AI-powered programming tools, which enable developers to construct bespoke collectives of AI agents, are transforming the software industry. Projects that would have historically taken AWS teams two to three years are now being accomplished in mere months by smaller groups, Garman observed.
Indeed, Garman noted that AWS teams are currently “building at a pace we haven’t seen in many years.”
Amazon’s actions may have resonated uncomfortably because they touch upon the two primary AI-related anxieties that have surfaced over the past year: the technology’s impact on employment and the specter of a speculative bubble.
However, some experts, such as James Landay, Co-founder of Stanford’s Institute for Human-Centered AI, have previously suggested to CNN that AI’s role in software creation has, in certain instances, been overstated. Furthermore, with every earnings report, analysts are increasingly scrutinizing when the billions poured into AI will materialize as new products.
Garman remains confident these substantial wagers will pay off.
At a recent gathering of around 150 senior technology executives, 90% raised their hands when asked if they were already realizing a “solidly positive” return on AI investment or anticipated one within the next six months, he added.
“I’m sure the signs of it exist,” Garman said regarding evidence of an AI bubble. “But I haven’t seen them yet.”
About the Author
