The cloud infrastructure race has always been about scale. Now, however, it is increasingly about margin. A new report from Semianalysis shows that Amazon.com Inc. (NASDAQ:AMZN) has pulled measurably ahead of its cloud rivals, and the engine driving that gap is not simply raw compute capacity. It is the structural economics of how Claude operates on Amazon Bedrock.

AWS Margin Growth Is Not an Accident

In Q1 2026, AWS posted revenue of $37.6 billion, up 28% year-over-year. That growth rate marked the fastest expansion in 15 quarters. More importantly, AWS operating margin reached 37.7%, rising for three consecutive quarters. Meanwhile, Semianalysis research shows AWS EBIT margins expanded by 213 basis points quarter-over-quarter. Competitors, by contrast, are not keeping pace. Microsoft Corporation's (NASDAQ:MSFT) Intelligent Cloud gross margin percentage declined slightly in Q1, due to the cost of scaling AI infrastructure, per Microsoft's own investor disclosures. That is a meaningful contrast: AWS is expanding margins as AI spending rises, while Azure is compressing them.

The Bedrock Deal Structure Is the Key

Most coverage of Claude's success focuses on user growth and revenue milestones. However, the more consequential story for investors lies in how the AWS-Anthropic commercial structure actually works. According to Semianalysis, Anthropic acts as the seller of record when Claude tokens move through Amazon Bedrock. AWS, therefore, earns both an infrastructure fee and a distribution revenue share for every token sold.

This arrangement functions like a high-margin royalty stream layered on top of AWS's existing compute revenue. In other words, AWS gets paid twice, once for the compute and once for the distribution, without absorbing the model development cost. That asymmetry creates an advantage few AI distribution arrangements in tech history can match.

Claude's Revenue Growth Is Feeding That Engine

The volume flowing through this structure is now substantial. Anthropic's run-rate revenue surpassed $30 billion as of April 2026, up from roughly $9 billion at the end of 2025, according to Anthropic's own announcement of its expanded AWS compute agreement. Additionally, over 100,000 customers now run Claude on Amazon Bedrock, per that same disclosure.

Semianalysis also notes that Anthropic's inference gross margins have climbed into the mid-60s, up sharply from 38% in 2025 and negative territory in 2024. Higher inference margins for Anthropic mean more profitable tokens flowing through Bedrock, which in turn feeds AWS's revenue share upside.

Custom Silicon Compounds the Advantage

Beyond the Anthropic deal structure, AWS is also benefiting from a hardware cost advantage. Amazon CEO Andy Jassy stated on the Q1 2026 earnings call that Trainium chips now power more than 50% of Amazon Bedrock token usage. According to Futurum Research's analysis of Amazon's Q1 results, Trainium2 delivers roughly 30% better price performance than comparable GPUs. Jassy further noted that custom silicon could save Amazon “tens of billions of capex dollars per year” and provide “several hundred basis points of operating margin advantage.” Graviton CPUs, which 98% of the top 1,000 EC2 customers already use, are also increasingly relevant for the CPU-intensive demands of agentic AI workloads. Together, these chips reduce AWS' unit cost of inference precisely as demand for inference is accelerating.

What This Means for Investors

The practical investment implication is straightforward. AWS is compounding two structural advantages at the same time: a high-margin distribution arrangement with the fastest-growing AI company in enterprise history, and a custom silicon portfolio that lowers the cost of delivering that AI at scale. Neither advantage is easily replicated in the near term.

Microsoft's Azure benefits from OpenAI demand growth, but also absorbs more infrastructure cost directly into its cloud business as AI usage scales. Google Cloud, despite 63% year-over-year growth in Q1, is still earlier in its margin maturity curve. AWS, by contrast, is demonstrating that scale and margin can expand together when the business model is structured correctly. For retail investors evaluating cloud exposure, that distinction matters significantly more than headline revenue growth rates alone.

The Bigger Structural Shift

Cloud computing began as a commodity infrastructure business. Today, it is rapidly becoming an AI distribution business. The players that control the pipes through which frontier AI flows are accruing margin advantages that go beyond typical infrastructure economics. AWS's EBIT expansion of 213 basis points in a single quarter reflects that shift in real time.

Furthermore, with Anthropic's AWS compute commitment now exceeding $100 billion over ten years, per Anthropic's April 2026 announcement, the economics of this partnership are designed to compound well into the next decade. Investors who focus only on revenue growth risk missing the more durable signal. Over the next decade, the biggest winners may be the platforms that monetize AI demand most efficiently. In the cloud AI era, margin architecture is the moat.

Benzinga Disclaimer: This article is from an unpaid external contributor. It does not represent Benzinga’s reporting and has not been edited for content or accuracy.