AMZN · Q1 2026 Earnings

Bullish

Amazon

Reported April 29, 2026

30-second summary

AWS grew 28% YoY to $37.6B — a fourth consecutive reacceleration (17.5% → 20.2% → 24% → 28%) and the fastest pace in 15 quarters, with AWS operating margin at 37.7% (+270bps QoQ from Q4's 35.0%, though -180bps YoY vs. Q1 2025's 39.5%). Q1 revenue of $181.5B (+17% YoY) beat consensus by 2.5% and cleared the high end of the guided range by $3.0B; GAAP operating income of $23.9B printed $2.4B above the high end of guidance. The Q2 FY2026 guide of $194–199B implies 16–19% YoY growth against the $167.7B prior-year base, sustaining the run rate, and the $364B AWS backlog (excluding a separately-disclosed $100B+ Anthropic commitment) is the metric that now does the bull case's heavy lifting.

Headline numbers

EPS

Q1 FY2026

$2.78

Revenue

Q1 FY2026

$181.52B

+17.0% YoY

+2.5% vs est.

Operating margin

Q1 FY2026

13.1%

Key financials

Q1 FY2026

Metric	Q1 FY2026	YoY	Q4 FY2025	QoQ
Revenue	$181.52B	+17.0%	$213.40B	-14.9%
EPS	$2.78	—	$1.95	+42.6%
Operating margin	13.1%	—	11.7%	+140bps

Guidance

Amazon delivered a strong Q1 FY2026 beat across revenue ($181.5B, +17% YoY) and operating income ($23.8B), exceeding prior guidance ranges on both metrics; forward Q2 FY2026 guidance implies sustained high-teens revenue growth (16-19% YoY) and operating income of $20-24B.

Guidance is issued one quarter forward. The Prior-guide column references the guide issued last quarter for the period just reported; the New-guide column is for next quarter.

Actuals vs prior guidance

Metric	Period	Prior guide	Actual	Δ	Result
Revenue	Q1 FY2026	$173.5 billion to $178.5 billion	$181.519 billion	+$3.0 billion above high end of guide	Beat
Revenue YoY growth	Q1 FY2026	11% to 15%	17%	+2-6 pts above guide	Beat
Operating income	Q1 FY2026	$16.5 billion to $21.5 billion	$23.786 billion	+$2.3 billion above high end of guide	Beat

New guidance

Metric	Period	Guide	YoY
Revenue	Q2 FY2026	$194.0 billion to $199.0 billion	+15.7% to +18.6% YoY
Revenue YoY growth	Q2 FY2026	16% to 19%	16% to 19% YoY
Operating income	Q2 FY2026	$20.0 billion to $24.0 billion	—

Reaffirmed unchanged this quarter: Capital expenditures (~$200 billion)

Segment performance

Q1 FY2026

Segment	Q1 FY2026	YoY
North America	$104.143B	+12.0%
International	$39.789B	+19.0%
AWS	$37.587B	+28.0%
Online Stores	$64.254B	+12.0%
Third-party Seller Services	$41.578B	+14.0%
Advertising Services	$17.243B	+24.0%
Subscription Services	$13.427B	+15.0%

Platform metrics

Q1 FY2026

Segment	Q1 FY2026
WW Paid Units Growth	15%
Advertising Services TTM Revenue	~$70 billion
AWS Chips Business Annual Revenue Run Rate	$20+ billion
Same-day or Overnight Delivery Items in 2026	1+ billion items

Profitability

Q1 FY2026

Segment	Q1 FY2026
AWS Operating Margin	37.7%
North America Operating Margin	7.9%
International Operating Margin	3.6%
Operating Cash Flow TTM	$148.5 billion

Management tone

Narrative arc: Q2 "demand exceeds supply" → Q3 "AWS reaccelerated, AI is revenue" → Q4 "we are spending $200B to own the stack" → Q1 "the stack is monetizing faster than we built it."

The custom-silicon disclosure regime stepped up another level. Three quarters ago, Trainium economics were a qualitative claim. Two quarters ago, Trainium 2 was "fully subscribed" and Trainium + Graviton combined at "$10B+ annualized run rate." Last quarter the Trainium roadmap was sketched through Trainium 4/5. This quarter Andy Jassy put a single discrete number on the chip business — "$20+ billion annual revenue run rate" — and disclosed that Trainium 2 has 30% better price-performance than comparable GPUs, Trainium 3 offers another 30–40% improvement on top, and Trainium 4 already has substantial reservations 18 months out. The progression from "subscribed" → "$10B run rate" → "$20B+ run rate with named generations sold out" is the strongest evidence yet that AWS owns a differentiated economic position in AI infrastructure, not just a backlog story.

Backlog scale doubled in three quarters with explicit breadth language. Q2's $195B backlog and Q3's emphasis on "demand exceeds supply" framed the AWS story around constrained capacity. This quarter Jassy disclosed $364B in Q1 backlog with "reasonable breadth beyond just one or two customers" — and then noted the Anthropic $100B+ commitment is on top of that figure. The Q4 brief flagged single-anchor-tenant concentration risk; the Q1 answer is that the underlying $364B is broad-based and Anthropic is the addition, not the core. Combined with $225B+ in revenue commitments specifically for Trainium, the demand-concentration debate is materially weaker than it was 90 days ago.

Confidence language stepped up from "extraordinary opportunity" to "biggest inflections of our lifetime." Q4's Jassy framed AI as "an extraordinarily unusual opportunity" while declining to disclose minimum FCF guardrails. Q1's framing — "we're in the middle of some of the biggest inflections of our lifetime, we're well positioned to lead, and I'm very optimistic" — is more emphatic, paired with results that for the first time deliver a >$2B operating-income beat. The conviction-without-guardrails posture from Q4 is being validated by a print that materially exceeded the wide guidance band.

A new growth vector, Leo, was articulated as a multi-billion-dollar revenue business with AWS-like economics. In Q2 Kuiper was "optionality." In Q4 it was a disclosed cost line ($1B incremental). This quarter Jassy laid out the commercial substance: 250+ satellites in orbit, 20+ launches in 2026 and 30+ in 2027, named enterprise commitments (Delta committing half its fleet by 2028, JetBlue, AT&T, Vodafone, DirecTV LatAm, NBN Australia, NASA), the GlobalStar-enabled Apple iPhone/Watch satellite partnership, and a self-described "many billion dollar revenue business" with high-capex/long-term-FCF economics. Leo has crossed from line-item investment to disclosed strategic growth vector inside one quarter.

Q&A highlights

Eric Sheridan · Goldman Sachs

Asked about the level of investment needed over the next couple of years to scale compute and capacity to meet the current revenue backlog, and how Amazon's custom silicon approach positions it competitively.

Andy emphasized AWS's 28% YoY growth (fastest in 15 quarters) and explained that the combination of Tranium (AI chips) and Graviton (CPU chips) positions Amazon uniquely for the AI inflection. He stated that AWS has over $225 billion in revenue commitments for Tranium, with Tranium 2 nearly sold out and Tranium 3 nearly fully subscribed. He characterized this as a 'once in a lifetime opportunity' and indicated significant capital investment will continue, noting this will benefit customers, shareholders, and Amazon long-term.

AWS revenue: $37.6B, growth accelerated to 28% YoY$225B+ in revenue commitments for TraniumTranium 2: 30% better price performance than comparable GPUs, largely sold outTranium 3: 30-40% better price performance than Tranium 2, nearly fully subscribed

Brian Novak · Morgan Stanley

Asked about AWS backlog visibility and breadth of customers beyond the big labs; also asked about Rufus and agentic commerce milestones for 2026.

Andy reported Q1 backlog of $364 billion with reasonable breadth beyond just one or two customers, though he noted the recent Anthropic deal ($100B+) was not included in this figure. On Rufus, he highlighted 115% growth in monthly active users and 400% growth in engagement YoY. He emphasized that Rufus is becoming competitive with third-party horizontal agents and that retailer-specific agents (like Rufus) will likely outperform because they have better product data, personalization, and account control. He drew parallels to early search engine days in e-commerce adoption.

AWS backlog: $364B in Q1 (excluding recent Anthropic $100B+ deal)Rufus monthly active users: up 115% YoYRufus engagement: up 400% YoYBedrock used by 125,000+ customers

Justin Post · Bank of America

Asked about the significance of having the full OpenAI model suite on Bedrock and Amazon's focus on its NOVA model; also asked about timing and opportunity for selling Tranium racks given capacity constraints.

Andy described having all OpenAI models available in Bedrock as 'a big deal' because customers demand choice across models. He emphasized that the stateful Bedrock Managed Agents (built with OpenAI) represents the future of agentic applications and is differentiated from competitors. On Tranium racks, he acknowledged this as 'very much a possibility' but said Amazon must balance current Tranium demand from existing customers against selling racks, and will decide on the mix over the next couple of years.

OpenAI GPT-5.4 model added to Bedrock, GPT-5.5 coming in weeksStateful Bedrock Managed Agents (new product announced) are differentiated from competitorsBedrock processes more tokens in Q1 than all prior years combinedBedrock customer spend grew 170% QoQ

Rob Sanderson · Lube Capital Markets

Asked to dimensionalize the revenue opportunity for Amazon Leo in consumer and enterprise segments over coming years; asked about new services enabled by GlobalStar spectrum; asked about longer-term vision including orbital data centers.

Andy expressed strong bullishness on Leo, characterizing it as solving global broadband access for billions of people. He stated Leo will launch with 2x better downlink and 6x better uplink performance than existing alternatives, plus cost advantages. He noted over 250 satellites in space, 20+ launches planned for 2026, and 30+ for 2027. Multiple enterprise commitments already signed (Delta, JetBlue, AT&T, etc.), with Delta committing at least half its fleet starting 2028. He described Leo as potentially becoming a 'many billion dollar revenue business' with AWS characteristics (capital-intensive upfront, strong long-term FCF and ROIC). GlobalStar acquisition provides direct-to-device spectrum and satellite expertise, and enabled partnership with Apple for iPhone and Apple Watch satellite services.

250+ satellites in space20+ launches planned in 2026, 30+ in 2027Downlink performance: 2x better than alternativesUplink performance: 6x better than alternatives

Shweta Kajuria · Wolf Research

Asked about impact of memory and storage price increases and supply chain inflation on capex in 2026-2027; also asked about agentic commerce opportunity for advertising.

On memory/storage, Andy acknowledged skyrocketing costs and capacity constraints, but stated Amazon has worked closely with strategic suppliers to secure significant supply early. He noted this actually benefits AWS by pushing on-premises enterprises to the cloud because suppliers prioritize large cloud providers. On agentic commerce and advertising, he expressed belief this will be positive for advertising, noting: (1) AI tools are making advertising creation faster/easier for SMBs; (2) multi-turn agentic conversations create multiple opportunities to surface relevant products; (3) sponsored prompts and dynamic product suggestions will drive monetization similar to existing ads success.

Memory and storage costs have 'skyrocketed'Supply shortage is pushing enterprises to cloud migrationAmazon secured significant supply agreements with strategic suppliersAgentic commerce will generate multiple monetization touchpoints per conversation

Answers to last quarter's watch list

AWS growth sustaining 22%+ — AWS printed 28%, a further 400bps reacceleration on top of Q4's 24%, and management characterized this as the fastest growth in 15 quarters. The $364B backlog (excluding $100B+ Anthropic) and $225B+ in Trainium-specific commitments corroborate that this is structural backlog conversion, not a one-quarter catch-up.

Resolved positively

AWS operating margin trajectory — AWS operating margin came in at 37.7%, up 270bps sequentially from 35.0% in Q4 but still down 180bps vs. the 39.5% Q1 2025 prior-year mark. The chip-led cost advantage has reversed the QoQ direction, but YoY compression remains and bears continued monitoring. Status: Partially resolved

Q1 operating income inside the $16.5–21.5B band — Operating income landed at $23.9B, $2.4B above the high end of the band. This is firmly above the upper-half threshold and signals investment is paced, not front-loaded.

Resolved positively

Trainium 3 customer logos — No named non-Anthropic customers were disclosed, but Jassy stated Trainium 3 is "nearly fully subscribed" and Trainium 4 has "much already reserved" 18 months from broad availability. The aggregate $225B Trainium commitment pool and Trainium chip business at $20B+ annual run rate is stronger evidence of breadth than logo disclosure would have been.

Resolved positively

FY2026 capex variance from $200B — Maintained at ~$200B with no revision. Management framed this as committed-and-justified by the $225B Trainium commitments and $364B AWS backlog.

Continue monitoring

International operating margin against Q4's 2.1% — International margin expanded to 3.6%, well above the sub-2% downside risk flagged last quarter. The "sharper prices" 2026 investment did not compress the line; reported revenue growth accelerated to +19% YoY (though ex-FX was flat at +11%).

Resolved positively

What to watch into next quarter

AWS growth sustaining 26%+ — does the 20.2% → 24% → 28% trajectory hold a fourth quarter, or does the easy-comp period from the 2024 capacity throttle finally end? A Q2 print of 26%+ would extend the multi-quarter reacceleration; a print back toward 24% would suggest 28% included a release of constrained capacity.

AWS operating margin recovering toward the 39.5% YoY mark — Q1's 37.7% reversed sequential compression but remains 180bps below Q1 2025. Watch whether Trainium economics close the YoY gap or whether depreciation from the $200B build keeps the line below the prior-year peak.

Q2 operating income inside the $20–24B band — upper half ($22–24B) would confirm operating leverage from chip economics is durable; lower half would signal Q1's $23.9B included one-time items or that depreciation acceleration is now flowing through.

AWS backlog growth and Anthropic disclosure integration — does the $364B figure (ex-Anthropic) keep growing, and does management start consolidating the Anthropic $100B+ into the headline backlog? A reported figure of $500B+ in Q2 would be the cleanest evidence the demand base broadened structurally.

Trainium revenue run rate trajectory — first cross-quarter data point on the newly disclosed $20B+ chip run rate. A Q2 print of $25B+ run rate would mark accelerating customer adoption; flat would suggest Q1's number was the steady state.

Leo commercial milestones — any named consumer-tier pricing, additional enterprise contracts, or firm Delta in-flight date would advance Leo from disclosed-strategic-vector toward earnings-driver status.

Sources

Amazon Q1 FY2026 earnings press release (SEC filing): https://www.sec.gov/Archives/edgar/data/1018724/000101872426000012/amzn-20260331xex991.htm
Q1 FY2026 earnings call Q&A.
Tapebrief Q4 FY2025, Q3 FY2025, and Q2 FY2025 AMZN briefs (for cross-quarter context).

Get the next brief, free.

We publish analyst-grade earnings briefs the same day or morning after every call — headline numbers, segment KPIs, Q&A highlights, and tone analysis. Free during beta.

This is not investment advice.