Crypto Briefing reported Monday that Amazon Web Services could generate better returns on AI infrastructure spending than Microsoft Azure over the next several years, arguing that AWS’s scale, custom silicon efforts, and rapid Bedrock adoption may improve the economics of its data-center buildout.
That is an investment thesis, not a forecast from either company. Amazon and Microsoft do not publish a comparable, AI-only return-on-investment figure for their cloud capital expenditures, and neither has said AWS is set to overtake Azure on that measure.
The underlying growth data are real. Amazon said in its first-quarter 2026 results that AWS revenue grew 28% year over year, its fastest rate in 15 quarters. It also said Bedrock customer spending rose 170% quarter over quarter and that Bedrock processed more tokens during the quarter than in all previous years combined. Microsoft reported Azure and other cloud-services revenue growth of 39% in its fiscal second quarter, while warning that AI infrastructure investment and a shift toward Azure reduced cloud gross-margin percentage.
The comparison hinges on what each company gets for its spending rather than who spends more. Microsoft is pushing Azure capacity alongside its own Copilot services, Azure AI Foundry, and major model-provider workloads. That can increase utilization quickly, but it also puts expensive accelerators, networking and power capacity ahead of revenue recognition.
Amazon has a different mix: AWS sells its own AI platform services, but it is also emphasizing Trainium and Inferentia chips as alternatives to the most constrained GPU supply. Amazon said its chips business reached a $20 billion annual revenue run rate in the first quarter. If AWS can meet more demand with internally designed hardware, it may have more control over costs and margins than a cloud provider dependent on premium third-party accelerators.
There is no sign that AWS has displaced Azure in AI-led cloud growth, however. Synergy Research Group’s first-quarter figures put AWS first with 28% of worldwide cloud-infrastructure spending, followed by Microsoft at 21% and Google at 14%. Synergy also said Microsoft and Google were growing faster than Amazon.
Crypto Briefing’s use of a $75 billion Amazon capital-expenditure plan is dated: it refers to a 2024 spending plan, not a current 2026 projection. That matters because the return profile of AI infrastructure is highly sensitive to build timing, equipment costs, depreciation, power availability and actual utilization.
The practical question is whether an AI workload belongs in a managed Azure service, on virtual machines, in a multi-cloud design, or on-premises. Buyers should compare committed-use discounts, data-egress exposure, regional GPU availability, model support, identity controls and support commitments—not assume that hyperscaler capex automatically produces lower prices.
Amazon and Microsoft will provide their next meaningful evidence in quarterly earnings and capacity updates, rather than in third-party predictions about a future ROI crossover.
That is an investment thesis, not a forecast from either company. Amazon and Microsoft do not publish a comparable, AI-only return-on-investment figure for their cloud capital expenditures, and neither has said AWS is set to overtake Azure on that measure.
The underlying growth data are real. Amazon said in its first-quarter 2026 results that AWS revenue grew 28% year over year, its fastest rate in 15 quarters. It also said Bedrock customer spending rose 170% quarter over quarter and that Bedrock processed more tokens during the quarter than in all previous years combined. Microsoft reported Azure and other cloud-services revenue growth of 39% in its fiscal second quarter, while warning that AI infrastructure investment and a shift toward Azure reduced cloud gross-margin percentage.
Large buildouts, different economics
The comparison hinges on what each company gets for its spending rather than who spends more. Microsoft is pushing Azure capacity alongside its own Copilot services, Azure AI Foundry, and major model-provider workloads. That can increase utilization quickly, but it also puts expensive accelerators, networking and power capacity ahead of revenue recognition.Amazon has a different mix: AWS sells its own AI platform services, but it is also emphasizing Trainium and Inferentia chips as alternatives to the most constrained GPU supply. Amazon said its chips business reached a $20 billion annual revenue run rate in the first quarter. If AWS can meet more demand with internally designed hardware, it may have more control over costs and margins than a cloud provider dependent on premium third-party accelerators.
There is no sign that AWS has displaced Azure in AI-led cloud growth, however. Synergy Research Group’s first-quarter figures put AWS first with 28% of worldwide cloud-infrastructure spending, followed by Microsoft at 21% and Google at 14%. Synergy also said Microsoft and Google were growing faster than Amazon.
Crypto Briefing’s use of a $75 billion Amazon capital-expenditure plan is dated: it refers to a 2024 spending plan, not a current 2026 projection. That matters because the return profile of AI infrastructure is highly sensitive to build timing, equipment costs, depreciation, power availability and actual utilization.
What it means for Windows and Azure shops
For Windows administrators and IT buyers, the AWS-versus-Azure ROI argument does not change any near-term deployment decision. Azure remains tightly integrated with Microsoft Entra, Microsoft 365, Windows Server, GitHub and the company’s enterprise management stack. Those operational links can outweigh a modest difference in infrastructure economics for organizations already standardized on Microsoft tooling.The practical question is whether an AI workload belongs in a managed Azure service, on virtual machines, in a multi-cloud design, or on-premises. Buyers should compare committed-use discounts, data-egress exposure, regional GPU availability, model support, identity controls and support commitments—not assume that hyperscaler capex automatically produces lower prices.
Amazon and Microsoft will provide their next meaningful evidence in quarterly earnings and capacity updates, rather than in third-party predictions about a future ROI crossover.
References
- Primary source: Crypto Briefing
Published: 2026-07-20T10:30:50+00:00
Amazon's ROI expected to surpass Microsoft's in AI cloud spending
Analysts expect AWS to deliver better ROI than Azure on AI spending. Here's what the cloud war means for investors and decentralized compute tokens.cryptobriefing.com - Official source: microsoft.com
FY26 Q1 - Intelligent Cloud Performance - Investor Relations - Microsoft
FY26 Q1 - Intelligent Cloud Performance - Investor Relations - Microsoftwww.microsoft.com
- Related coverage: crn.com
Cloud Market Share Q1 2026: AWS, Microsoft, Google Battling In AI Era
Global cloud market share leaders AWS, Microsoft Azure, Google Cloud, Oracle and Alibaba compete in first quarter 2026 with AI driving enterprise cloud sales.www.crn.com
