Amazon (AMZN) Q4 FY25 earnings call: guidance and Q&A

How will investors see strong long-term ROI from the current CapEx cycle, considering its duration, profitability levels, and free cash flow generation guardrails?

Brian Olsavsky and Andy Jassy explained that AWS capacity is immediately useful and monetized, driven by strong demand for AI services, with AWS operating margin at 35% (up 40 bps YoY). Jassy emphasized the 'unusual opportunity' in AI, expressing confidence in strong ROI based on past AWS experience, custom chips, hardware, and power inventions, without specific financial guardrails.

“I passionately believe that every customer experience that we know of today is going to be reinvented with AI, there are going to be a whole bunch of customer experiences that none of us ever imagined that are going to become the norms of how we all operate every day and what we use.”

asked by Mark Stephen Mahaney · answered by Andrew Jassy

Update on Project Rainier with Anthropic, clarification on Trainium chip numbers, and whether there are financial guardrails around the significant AI spend.

Andy Jassy highlighted Trainium's strong performance, offering 30-40% better price performance than comparable GPUs, and its multibillion-dollar run rate. He confirmed Trainium3 supply commitment by mid-2026 and substantial interest in Trainium4 (2027). He reiterated the aggressive investment in AI as an 'extraordinarily unusual opportunity' for Amazon and its customers, without detailing specific financial guardrails.

“Trainium has 30% to 40% better price performance than comparable GPU. So it's very compelling to customers.”

asked by Douglas Anmuth · answered by Andrew Jassy

How is the AI market changing from being top-heavy, and how might Amazon's relationship with companies like OpenAI extend to support its AI efforts?

Andy Jassy described the AI market as 'barbelled,' with AI labs and productivity workloads at the ends, and enterprise production workloads in the middle representing the largest future opportunity. He noted the broad nature of the AI movement, acknowledging the agreement with OpenAI and respecting the company, but emphasized that the AI ecosystem will involve 'thousands of companies' over time.

“I would tell you that this movement in what's happening in AI is -- it's very broad. It's going to be a lot of companies. It is a lot of companies already.”

asked by Ross Sandler · answered by Andrew Jassy

How will Agentic protocols impact the retail business and on-site ads, potentially leading to a compression in the shopping funnel?

Andy Jassy expressed optimism about Agentic shopping, citing Rufus's success (300 million users in 2025, 60% higher purchase completion). He discussed potential relationships with third-party horizontal agents but believes customers will ultimately prefer retailer-specific agents due to better shopping history, data accuracy, selection, pricing, and trust.

“I think a lot of customers are ultimately going to choose to use a great shopping agent from that retailer. Because if you think about what consumers really want in retail in a retailer, they want really broad selection. They want low prices. They want really fast delivery.”

asked by Michael Morton · answered by Andrew Jassy

What are the sources of efficiency and areas of investment planned for the global retail business this year?

Andy Jassy outlined investment areas including expanding selection (luxury brands, everyday essentials), quick commerce (Amazon Now), and perishable grocery delivery. He detailed efficiency drivers such as honing regionalization in the fulfillment network, increasing units per box to save shipments, and expanding the use of robotics, with over 1 million robots currently in the network.

“If you look at what we've done there, we've extended the number of regions. It was 8. It's now 10. We've extended regionalization to what we do with our inbound delivery to be much more efficient in being able to get more items closer to customers more quickly.”

asked by Brian Nowak · answered by Andrew Jassy

Can you discuss the AWS revenue backlog, internal and external AI use cases, and any supply/demand imbalance for AI efforts, and how to close that gap?

Andy Jassy reported AWS backlog at $244 billion, up 40% YoY and 22% QoQ. He clarified that external customers consume most capacity, but Amazon uses AI extensively internally (over 1,000 applications). He acknowledged supply constraints for AI but highlighted aggressive capacity expansion, adding 3.9 gigawatts of power in the last 12 months and 1.2 gigawatts in Q4 alone, with plans to double again by end of 2027.

“We're growing 24% year-over-year on a $142 billion annualized run rate business. So we're growing at really an unprecedented rate yet, I think every provider would tell you, including us that we could actually grow faster if we had all the supply that we could take.”

asked by Eric Sheridan · answered by Andrew Jassy