An increase in Amazon Web Services spending cut into the cloud giant’s profits in the second quarter as AWS raced to meet demand for AI against Microsoft, Google, and others.
While Amazon beat Wall Street’s overall expectations, its shares fell in after-hours trading as investors focused on rising costs within the cloud unit, the company’s biggest profit engine.
AWS revenue grew 17.5% year-over-year to $30.9 billion, in line with estimates. But the costs of competing in the AI arms race were evident in the cloud division’s profits. AWS operating income grew less than 9% to $10.2 billion as its operating expenses surged to $20.7 billion, up from $16.9 billion a year earlier.
This resulted in an operating margin of 32.9% — the cloud giant’s lowest profit-to-revenue ratio since late 2023, and well below the operating margin of nearly 40% in the first quarter.
On the company’s earnings call with analysts, Amazon CFO Brian Olsavsky attributed the pressure on AWS’s profit margin to a seasonal increase in stock-based compensation, unfavorable foreign exchange rates, and higher depreciation costs from ongoing investments in AI infrastructure.
on the call, analysts pressed Amazon CEO Andy Jassy to respond to a “Wall Street narrative” that the company is falling behind rivals Microsoft and Google in pursuing the AI opportunity.
Amazon’s results came a day after Microsoft disclosed annual revenue of more than $75 billion for its Azure cloud platform, with quarterly growth of 39%, far exceeding expectations.
Google Cloud’s annual revenue run rate is more than $50 billion, with quarterly growth of 32% year-over-year, according to numbers disclosed by Google parent Alphabet last week.
AWS is a significantly larger business, with an annual revenue run rate of more than $123 billion, making it much more difficult to generate eye-popping growth rates.
But even at that size, Morgan Stanley analyst Brian Nowak asked Jassy if it would be possible for AWS growth to accelerate for the remainder of 2025, “given the size of the opportunity” and the volume of the generative AI workloads expected to come online in the next 12 months.
While Jassy declined to give specific guidance for the AWS segment, he said he is “optimistic about the AWS business” and its potential to accelerate.
He cited a combination of factors, including more enterprises resuming their migrations from on-premises data centers to the cloud, the expected increase in companies deploying AI applications into production, and more AWS capacity coming online in the coming months.
Jassy described long-term strategic advantages for AWS, including the fact that AI inference will eventually be treated as “another building block” like compute, storage and database.
Because of this, he said, many customers will want to run their AI applications close to where their other applications and data already reside — creating a powerful advantage for Amazon, because “there’s just so many more applications and data running in AWS than anywhere else.”
Jassy said AWS has a key advantage with its custom-built chips, which he said offer 30% to 40% better price-performance for inference than other GPU providers.
He also pointed to what he called a “very big difference” in security — alluding to problems that have plagued Microsoft without naming the company directly — and noted that AWS has a more significant, end-to-end set of AI services “from the bottom of stack all the way to top.”
AI milestones for the company in the recent quarter included the launch of Kiro, a new agentic integrated development environment (IDE) that lets developers code in natural language, while the system automatically creates documentation and specs, and scans for security issues.
The company reported $31.4 billion in capital expenditures for the second quarter, up from nearly $25 billion in the first quarter, most of it related to technology infrastructure. Amazon’s capex number is not directly comparable to Microsoft and Google due to its additional investments in fulfillment centers and its Project Kuiper satellite network.
Olsavsky said AWS will continue to invest more capital in data centers and other resources “to pursue this unusually large opportunity that we have in generative AI.” Operating margins, he said, will “fluctuate over time, driven in part by the level of investments we are making.”
