What Management Said
Read the full Q1 2026 transcript ↗Our guidance incorporates the order trends that we've seen to date and what we believe today to be appropriate assumptions. Our guidance assumes, among other things, that we don't conclude any additional business acquisitions, restructurings, or legal settlements. It's not possible to accurately predict demand for our goods and services, and therefore, our actual results could differ materially from our guidance. Starting with AWS, growth continued to accelerate, up 28% year-over-year, the fastest growth rate in 15 quarters, up $2 billion quarter-over-quarter, the largest Q4 to Q1 AWS revenue increase ever.
The last time we saw growth at this clip, AWS was roughly half the size. To put our growth in perspective, three years after AWS launched, it had a $58 million revenue run rate. In the first three years of this AI wave, AWS's AI revenue run rate is over $15 billion, nearly 260x larger. That includes model building with SageMaker, which reduces training time by up to 40%, high-performance inference with the leading selection of frontier models in Bedrock, which saw 170% growth in customer spend quarter-over-quarter.
OpenAI has said they're already seeing unprecedented demand for this new product, and we're seeing heavy customer interest as well. Customers can deploy agents with enterprise scale, security, and reliability with AgentCore, which is being used to deploy an agent as frequently as every 10 seconds. We also offer turnkey agents for coding, software migrations, business operations, and knowledge workers in Amazon Q, AWS Transform, Amazon Connect, and Amazon Quick. They're choosing AWS because we've built the broadest and most capable core offerings by a wide margin.
What went well
- AWS growth accelerated to 28% YoY (fastest in 15 quarters), reaching a $150B annualized run rate, up $2B QoQ (largest-ever Q4-to-Q1 increase).
- AWS operating income was $14.2B, reflecting strong growth plus efficiency gains.
- Worldwide operating income was $23.9B at a 13.1% operating margin, the highest ever.
- Worldwide revenue grew 17% (+15% ex-FX) to $181.5B.
- The chips business grew ~40% QoQ to a $20B+ annual run rate (equivalent to ~$50B if sold standalone), now one of the top-three data-center chip businesses in the world.
- Backlog reached $364B, excluding the recently announced $100B+ Anthropic deal; Trainium revenue commitments now exceed $225B.
- Bedrock customer spend grew 170% QoQ and processed more tokens in Q1 than all prior years combined; AI revenue run rate is $15B+ (~260x AWS's revenue at the same stage of the cloud wave).
- North America operating income was $8.3B (7.9% margin); unit growth of 15% outpaced cost growth (outbound shipping +12%, fulfillment +9% FX-neutral).
What went wrong
- Guidance flagged an ~$1B YoY North America cost increase in Q2 2026 for Amazon Leo (satellite manufacturing/launch), with costs mostly expensed until capitalization begins in Q4.
- Memory and storage component costs have skyrocketed amid a supply/demand imbalance, pressuring the supply chain (managed via strategic suppliers, but a watch item for CapEx).
- Q2 guidance anticipates higher transportation costs from fuel inflation (partly offset by a new FBA fuel/logistics surcharge) plus the seasonal Q2 stock-based-comp step-up.
- International operating margin remained modest at 3.6% (op income $1.4B) amid continued stores investment.
- Cash CapEx climbed to $43.2B in the quarter as AI/AWS investment scaled, with a six-to-24-month lag before monetization.
Guidance Changes
| Metric | Period | Current guidance |
|---|---|---|
| Net sales | Q2 2026 | $194B-$199B (FX ~10bps headwind; assumes Prime Day in Q2 for most large geos) |
| Operating income | Q2 2026 | $20B-$24B (includes seasonal SBC step-up) |
| North America cost - Amazon Leo | Q2 2026 | ~$1B YoY cost increase; commercial service on track for Q3, capitalization begins Q4 |
| Trainium3 supply | 2026 | Nearly fully subscribed; much of Trainium4 (~18 months out) already reserved |
Performance Breakdown
| Metric | YoY | Note |
|---|---|---|
| Worldwide revenue | +17% (+15% ex-FX) ($181.5B) | Strong seasonal shopping events and 3P seller growth; $2.9B favorable FX. |
| Worldwide operating income | n/a ($23.9B) | Record 13.1% operating margin on efficiency across all segments. |
| AWS revenue | +28% ($37.6B) | Acceleration of 480bps on both core and AI services; strong AI/core correlation; $150B run rate, fastest growth in 15 quarters. |
| AWS operating income | n/a ($14.2B) | Strong growth plus efficiency gains, including 4x Trainium2 token-throughput improvement lifting effective capacity. |
| North America operating income | n/a ($8.3B) | 7.9% margin; 15% unit growth outpaced cost to serve as network efficiency improved. |
| International operating income | n/a ($1.4B) | 3.6% margin amid continued stores investment. |
| Chips business (Trainium + Graviton) | triple-digit YoY (+~40% QoQ) | $20B+ run rate (~$50B standalone-equivalent); one of the top-three data-center chip businesses. |
| AI revenue | triple digits ($15B+ run rate) | Bedrock spend +170% QoQ; more tokens processed in Q1 than all prior years combined. |
| Cash CapEx | n/a ($43.2B) | Primarily AWS and generative AI to support strong customer demand. |
Earnings Call Themes & Trends
| Topic | Previous mention | Current period | Trend |
|---|---|---|---|
| AWS growth trajectory | 24% YoY in Q4 2025 | 28% YoY in Q1 2026 (fastest in 15 quarters), $150B run rate | — |
| Backlog | $244B (Q4 2025) | $364B, excluding the new $100B+ Anthropic deal | — |
| Custom silicon / commitments | Trainium3 launching, $10B+ chip run rate (Q4) | $225B+ Trainium commitments; chips $20B+ run rate; Trainium4 largely reserved; expected to save tens of billions of CapEx/yr and several hundred bps of margin | — |
| Model choice / OpenAI | Bedrock had Anthropic, Nova, Mistral, Llama (Q4) | OpenAI GPT-5.4 added to Bedrock (5.5 soon); Bedrock Managed Agents (stateful) with OpenAI in preview | — |
| Supply chain | Not a focus previously | Memory/storage costs skyrocketing; pushing on-prem customers to the cloud since suppliers prioritize large cloud buyers | — |
| Operating margin | Strong segment margins (Q4) | Record 13.1% worldwide operating margin in Q1 2026 | — |
Q&A Summary
What investment levels are needed to scale compute/capacity to meet backlog, and how does custom silicon position you competitively? (Eric Sheridan, Goldman Sachs)
Jassy: pleased with 28% growth on a $150B base; customers choose AWS for broad functionality, proximity of inference to data, and security; core is also growing as AI drives post-training, RL and agentic/tool usage. The unusual Graviton (CPU) + Trainium (AI) chip set positions AWS well; capital plan unchanged, viewing this as a once-in-a-lifetime opportunity worth aggressive investment.
What is the AWS backlog and its breadth beyond the big labs, and what are your 2026 agentic-commerce milestones? (Brian Nowak, Morgan Stanley)
Jassy: backlog is $364B, excluding the recent $100B+ Anthropic deal, with reasonable breadth beyond one or two customers. On agentic commerce, focus is on making Rufus the best shopping assistant (monthly actives +115%, engagement +400% YoY) while working with third-party horizontal agents whose experience still lags (wrong pricing/product info, no personalization).
How big an unlock are OpenAI models on Bedrock versus your own Nova, and what is the opportunity to sell Trainium racks? (Justin Post, Bank of America)
Jassy: full OpenAI model suite on Bedrock is a big deal because customers want choice (GPT-5.4 live, 5.5 in weeks); the future is stateful models/APIs, and the OpenAI-built Bedrock Managed Agents (stateful) is unique. Given Trainium demand, there is a good chance AWS sells Trainium racks over the next couple of years, balanced against existing customer allocation.
Can you dimensionalize the Amazon Leo revenue opportunity, ramp governors, and long-term vision? (Rob Sanderson, Loop Capital)
Jassy: bullish; billions lack broadband and many enterprises/governments lack visibility. Leo (250+ satellites in space) will be one of two current-edge offerings, ~2x better downlink and ~6x better uplink with a cost advantage; enterprises/governments value pairing Leo data with AWS analytics/AI. The only governor is getting the constellation up (20+ launches in 2026, 30+ in 2027); Delta committed half its fleet from 2028; Globalstar brings scarce spectrum and an Apple direct-to-device relationship.
How are rising memory/storage prices and supply-chain inflation affecting CapEx, and how does agentic commerce affect advertising? (Shweta Khajuria, Wolfe Research)
Jassy: memory costs have skyrocketed on a demand/supply gap; AWS secured supply early via strategic partners and is not capacity-constrained, and the shortage is actually pushing on-prem customers to the cloud. On ads, agentic tools make campaign creation far easier (more advertisers), and multi-turn agentic conversations create multiple organic and sponsored placement opportunities (including sponsored prompts) - a net positive for advertising.
How does AI demand differ between early adopters and the broader enterprise base, and where is the biggest internal opportunity? (Colin Sebastian, Baird)
Jassy: AI labs are spending enormously; the largest absolute enterprise success today is cost-avoidance/productivity (customer service, business-process automation, fraud), with brand-new reinvented experiences increasingly reaching production. Internally, AI will have giant impact on every business and function; example - a service engine was rebuilt by five AI-forward engineers in 65 days versus the ~40-50 people over a year it would normally take.
More on Amazon Com Inc
Reported 2026-04-29 · figures from the Amazon Com Inc Q1 2026 earnings call.
See how Top Bucket AI works for your firm
Request DemoTop Bucket AI Research
Stay ahead of private markets
Research and market intelligence for private-markets professionals.
✓
You're subscribed.
Thanks for signing up.