Amazon’s stock has been under pressure — down eight consecutive trading sessions — as investors digest the staggering implications of the company’s plan to spend roughly $200 billion on capital expenditures, with the lion’s share directed toward artificial intelligence infrastructure. The announcement, made during Amazon’s most recent earnings report, has reignited a familiar debate on Wall Street: Are the hyperscalers spending too much, too fast, on a technology whose returns remain uncertain?
But Matt Garman, the chief executive of Amazon Web Services, isn’t flinching. In a CNBC exclusive interview on “Money Movers” with Jon Fortt, Garman laid out a forceful case for why the spending is not only justified but essential — and why he believes AWS is better positioned than any of its rivals to capture the enormous wave of enterprise demand that is building across the cloud and AI sectors.
A $142 Billion Run Rate and Accelerating Demand
Garman’s confidence is rooted in numbers that, by any measure, are formidable. AWS reported 24% year-over-year revenue growth in its most recent quarter, putting the division on a $142 billion annual run rate. For a business of that scale, sustaining that kind of growth is remarkable — and Garman insists the trajectory is only steepening. “I’m just as bullish about the AWS business as I’ve ever been, frankly,” he told CNBC. “Every single customer is seeing more and more demand for moving their workloads to the cloud, putting their data on AWS, and then layering AI on top of that to get value for their business.”
The key distinction Garman emphasized is the breadth of demand. Unlike some competitors whose cloud revenue growth is heavily concentrated in a handful of massive customers — think large AI model builders consuming enormous quantities of compute — AWS’s growth is distributed across tens of thousands, even hundreds of thousands, of customers. “This demand is not isolated in one or two customers,” Garman said. “We have tens of thousands, hundreds of thousands of customers that are driving this demand. And all of them are growing pretty rapidly as they’re thinking about how they deliver more value to their business.”
The $200 Billion Question: Why Wall Street Is Nervous
The $200 billion capital expenditure figure — disclosed during Amazon’s earnings call on Thursday, as reported by Reuters — has become the central point of contention between Amazon’s leadership and a skeptical investor base. Amazon shares have declined for eight straight days, reflecting anxiety that the company is pouring money into data centers and servers at a pace that could compress margins and delay the return of free cash flow to shareholders.
The concern is not unique to Amazon. Microsoft, Alphabet, and Meta have all announced massive AI infrastructure spending plans, and the collective capital expenditure commitments from the major hyperscalers now run well into the hundreds of billions of dollars. According to Bloomberg, the combined AI-related capital spending by the four largest U.S. tech companies is exceeded $300 billion in 2025 alone, a figure that has prompted some analysts to draw uncomfortable parallels to previous technology investment cycles that ended badly.
Garman’s Rebuttal: This Is an Investment, Not a Gamble
Garman, however, frames the spending not as speculative but as a direct response to customer demand that AWS is struggling to meet. “The 200 billion in capex spending, that’s to invest in that future for those customers to be able to land their workloads in the cloud,” he explained on CNBC. “For them to be able to ramp up, we need to build data centers, we buy servers. And so this is an investment for us. It’s obviously a large number, but it’s something that we think is a really great investment. And it bodes — it’s an indication of how bullish we are about the future and the growth of AWS.”
This framing is deliberate. Amazon has long operated with a willingness to sacrifice short-term profitability for long-term market dominance, a philosophy inherited from founder Jeff Bezos. The original AWS business itself was built on years of heavy investment before it became the profit engine that now generates the majority of Amazon’s operating income. Garman appears to be making a similar argument: the AI infrastructure buildout is the next phase of that same playbook, and the customers are already there, waiting for capacity.
The Hyperscaler Horse Race: How AWS Measures Its Competitive Position
One of the more revealing moments in the CNBC interview came when Fortt pressed Garman on how AWS internally gauges its competitive standing against Microsoft Azure and Google Cloud Platform — the two other members of the so-called “Big Three” of cloud computing. Investors have grown increasingly sophisticated in parsing the quarterly results of each hyperscaler, comparing percentage growth rates, absolute dollar additions, and the composition of revenue.
Garman acknowledged that AWS’s percentage growth rate is lower than some competitors, a mathematical inevitability given its much larger revenue base. But he pivoted to what he considers a more meaningful metric: absolute dollar growth. “If you look at the absolute dollars added last quarter, and the quarter before that and the before that, AWS was bigger than everybody else,” Garman said. “And so we continued to grow in absolute dollars faster than our competitors do.” This is a significant claim. While Microsoft’s Intelligent Cloud segment and Google Cloud have posted higher percentage growth rates in recent quarters, AWS’s sheer scale means that each percentage point of growth translates into billions more in revenue than the same percentage point for a smaller competitor.
Diversification as a Strategic Moat
Beyond raw financial metrics, Garman pointed to diversification of the customer base as a critical competitive advantage — and one that he believes will prove decisive over the long term. While AWS counts major AI companies like OpenAI and Anthropic among its customers, Garman stressed that the company’s strategic focus extends far beyond serving AI model builders. “We’re very focused on how do we get all of the largest enterprises in the world, how do we get all of the startups in the world building on top of AWS,” he said.
This emphasis on breadth is more than a talking point. It reflects a fundamental strategic calculation. AI model training — the workload that companies like OpenAI and Anthropic represent — is enormously capital-intensive and tends to be concentrated among a relatively small number of customers. If one or two of those customers were to shift their spending or build their own infrastructure, a cloud provider heavily dependent on that revenue could face a sudden and painful shortfall. By contrast, a customer base that spans banks, healthcare companies, retailers, manufacturers, and startups is inherently more resilient. “For the long-term health of the business,” Garman said, AWS is focused on penetrating “every bank, every healthcare company, every retail company, every manufacturing company.”
The AI Layer: From Infrastructure to Intelligence
What makes the current moment different from previous cloud infrastructure buildouts is the AI dimension. According to The Wall Street Journal, AWS has been aggressively expanding its AI service offerings, from its Bedrock platform for accessing foundation models to its custom Trainium and Inferentia chips designed to reduce the cost of AI training and inference. These custom silicon efforts are particularly important because they reduce AWS’s dependence on Nvidia, whose GPUs currently dominate the AI training market but whose supply constraints and pricing power have become a strategic vulnerability for all hyperscalers.
Garman’s vision, as articulated in the interview and in previous public appearances, is that the real value of AI for most enterprises won’t come from training their own large language models. Instead, it will come from deploying pre-trained models on top of their existing data — data that, increasingly, resides on AWS. This creates a powerful flywheel: the more data enterprises store on AWS, the more natural it becomes for them to use AWS’s AI services to extract value from that data, which in turn drives more compute consumption and more revenue for Amazon. As CNBC reported, AWS’s AI-related revenue is growing at a triple-digit percentage rate, albeit from a smaller base.
Fears in the Software Space: Who Gets Disrupted?
The CNBC interview also touched on a topic that has been generating significant anxiety across the technology sector: the potential for AI to disrupt traditional software companies. As AI agents become more capable of performing tasks that previously required specialized software applications, there is a growing concern that many of the subscription-based software businesses that have powered the enterprise technology sector for the past two decades could see their value propositions erode. According to reporting from The Financial Times, venture capital firms are already redirecting investment away from traditional SaaS startups and toward AI-native companies that promise to deliver the same functionality at a fraction of the cost.
For AWS, this dynamic cuts both ways. On one hand, if traditional software companies lose market share, some of the workloads they run on AWS could diminish. On the other hand, the AI-native companies that replace them are likely to be even more compute-intensive, potentially driving even greater demand for cloud infrastructure. Garman has previously suggested that the net effect will be positive for AWS, as the overall volume of computing required by the economy continues to grow. The transition from traditional software to AI-powered alternatives doesn’t reduce the need for infrastructure — it increases it, because AI workloads are fundamentally more resource-hungry than the applications they replace.
The Capacity Constraint: Building Fast Enough
One of the underappreciated dynamics in the current AI infrastructure race is the constraint imposed by physical reality. Building data centers takes time — typically 18 to 24 months from site selection to operational capacity, and sometimes longer when permitting, power supply, and supply chain issues are factored in. According to Data Center Dynamics, AWS has been on a global data center construction spree, with new facilities under development across the United States, Europe, Asia, and the Middle East.
The $200 billion capex figure reflects not just the purchase of servers and networking equipment but the enormous cost of land acquisition, construction, and — perhaps most critically — securing reliable power supply. AI data centers consume vastly more electricity than traditional cloud computing facilities, and the race to secure power has become one of the most significant bottlenecks in the industry. Amazon has been signing power purchase agreements with nuclear, solar, and natural gas providers at an unprecedented pace, a strategy that Garman has described as essential to ensuring that AWS can deliver the capacity its customers are demanding.
What Investors Are Missing, According to Garman
The eight-day stock decline suggests that investors remain unconvinced — or at least uncertain — about the return profile of Amazon’s massive spending commitment. But Garman’s argument, distilled to its essence, is that the market is underestimating both the durability and the breadth of the demand signal. This is not, in his telling, a speculative bet on a technology that might generate returns someday. It is a response to concrete customer demand that is already showing up in the revenue numbers.
The 24% growth rate on a $142 billion run rate is, by itself, a powerful data point. It means AWS added roughly $27 billion in annualized revenue over the past year — more than the entire annual revenue of many large technology companies. If that growth rate can be sustained or even modestly accelerated, the return on the $200 billion investment could be substantial. The question, as always, is whether the growth will persist. Garman is betting it will. “We’re incredibly bullish on AWS growth over the next few years,” he said flatly.
The Broader Stakes: AI Infrastructure as the New Foundation of the Economy
Zooming out, the spending commitments from Amazon and its peers represent something more than a corporate capital allocation decision. They are, collectively, a massive bet that artificial intelligence will become as fundamental to the global economy as the internet itself — and that the companies that build and control the infrastructure on which AI runs will occupy a position of extraordinary strategic importance. According to Goldman Sachs Research, global AI infrastructure investment could exceed $1 trillion cumulatively by 2028, a figure that would make it one of the largest capital formation events in economic history.
For Amazon, the stakes are particularly high. AWS is not just a business unit — it is the financial engine that subsidizes Amazon’s lower-margin retail and logistics operations. If the AI infrastructure buildout delivers the returns Garman is projecting, AWS could become even more dominant, potentially generating hundreds of billions in annual revenue and cementing Amazon’s position as one of the most valuable companies in the world. If the spending proves excessive and demand fails to materialize at the expected scale, the consequences could be severe — not just for Amazon’s stock price, but for the broader technology sector that has hitched its fortunes to the AI narrative.
The Road Ahead: Execution Will Determine Everything
Matt Garman’s message to investors, customers, and competitors is clear: AWS sees a generational opportunity in AI, and it intends to spend whatever is necessary to capture it. The $200 billion figure, while eye-popping, is presented not as a ceiling but as a floor — a minimum investment required to meet the demand that is already visible in the pipeline. Whether that confidence is vindicated will depend on execution: building data centers on time, securing power, delivering AI services that enterprises actually adopt, and maintaining the customer diversification that Garman considers AWS’s greatest competitive advantage.
The market, for now, is voting with its feet — or at least pausing to catch its breath. Eight consecutive down days for Amazon stock suggest that investors want to see more proof before they fully buy into the vision. But if Garman is right — if the demand is as broad, as deep, and as durable as he claims — then the current stock weakness may look, in retrospect, like a buying opportunity. That is the bet Amazon is making. And at $200 billion, it is one of the largest bets in the history of corporate America.