Amazon AI Revenue Has Surpassed $25 Billion

CEO Andy Jassy cited shifting market demands as the company pivots toward custom silicon to lower computing costs.

Updated on Sept. 21, 2026 in Artificial Intelligence

Amazon AI Revenue Has Surpassed $25 Billion

Live Poll

Do you worry that AI will permanently reduce the availability of professional jobs in your field?

Amazon has achieved an AI revenue run rate exceeding $25 billion during the first three years of the current generative AI wave. This milestone marks a significant acceleration from the company's early infrastructure growth.

Why it matters

By developing its own budget AI-accelerator chips, Amazon is attempting to mitigate the high costs of computer processing while offering customers a wider variety of models. This strategy challenges the assumption that a single industry-standard model will dominate the market.

Amazon reported an AI revenue run rate of over $25 billion within the first three years of the current adoption cycle. This figure contrasts with the $58 million revenue run rate recorded during the first three years of Amazon Web Services.

The players

Andy Jassy

The CEO of Amazon who has overseen the company's transition from an e-commerce giant into a dominant cloud infrastructure and AI provider.

Amazon

A global technology conglomerate whose stack includes the AWS cloud platform and a rapidly growing portfolio of custom AI hardware.

The details

Amazon is producing its own budget-conscious AI-accelerator chips to lower the cost of processing for generative AI, which typically requires multiple models to function. By building custom hardware, the company aims to move away from expensive, generalized computing architectures. Future shifts in the labor market are expected as AI agents begin performing tasks currently handled by humans.

1997: Andy Jassy joined Amazon.
2002: Amazon Web Services launched.
August 2026: Amazon reached a $3 trillion market valuation.
Thursday, September 17, 2026: Andy Jassy spoke at Harvard Business School.

The Tech Race

Amazon is positioning its custom silicon as a necessary answer to the compute-intensive nature of modern generative AI applications. This strategy aligns with the firm's historical evolution, mirroring the early growth trajectory of the AWS platform.

The development of custom AI silicon is aimed at reducing processing costs for businesses that utilize multiple generative AI models simultaneously. These optimizations are expected to reach enterprise customers as the hardware is integrated into the broader Amazon Web Services stack.

The takeaway

Amazon believes the future of AI will favor diversity, with six to eight competing models becoming the standard. Investors should watch for the performance impact of new custom chips on overall AI operating margins in upcoming quarterly filings.