UBS: Dissecting AWS AI Backlog Orders, Maintaining Amazon Buy Rating

Overview

This research report is issued by UBS (UBS), focusing on the analysis of Amazon (AMZN.US) cloud computing department (AWS) backlog order structure, particularly dissecting AI business operations (Bedrock platform and partnerships with Anthropic, OpenAI). Due to previous overestimation of AWS income forecast for 1Q26, analysts revised their methodology, breaking down total backlog into core non-AI sections, Bedrock, and recently signed large AI lab contracts. The report conclusion states that AI business will become the core engine driving AWS growth, pushing 2026 AWS revenue growth rate to 36%, significantly higher than market consensus. Based on this, UBS maintains a "Buy" rating and $333 target price for Amazon, believing the current stock price remains undervalued relative to its status as a quality asset.

Core views

AWS Growth Dynamics Restructuring and AI Share Surge: UBS points out that AI-driven computing revenue share in AWS total revenue is rising rapidly. According to CEO Andy Jassy disclosed AI income run rate of $15 billion for 1Q26, it was calculated that Bedrock quarterly income was approx. $1.7 billion. Estimated that by end of 2026, AI-related revenue will account for 26% of AWS total revenue, reaching 30% in 2027. This structural change is key to supporting AWS accelerated growth. Backlog Fine-grained Decomposition: The report splits AWS Performance Obligations (PO) into three main parts: 1) Core non-AI business; 2) Bedrock platform, backlog approx. $35 billion as of 1Q26 end, up nearly 340% YoY; 3) Large AI lab contracts, including agreements signed with OpenAI and Anthropic. Among them, OpenAI's second $100 billion commitment agreement is expected to partially convert to revenue starting 2Q26, and fully convert in 3Q26; Anthropic's approx. $100 billion agreement is expected to start converting from 4Q26. Revenue Forecast Upside and Profit Potential: Based on the new backlog accumulation model, UBS forecasts 2026 AWS revenue will reach $175.9 billion, up 36% YoY, while market consensus is only $166.6 billion (up 29%). This difference will be further compounded in 2027, leading to UBS estimate of 2027 operating profit being approx. 40% higher than market consensus. Although quarterly forecasts may fluctuate due to linear conversion assumptions, annual trend judgments are more reliable. Synergistic Growth of E-commerce and Advertising Businesses: Besides AWS, the report also highlights other business segments of Amazon. E-commerce business sees faster margin expansion driven by unit economic improvement as unit growth continues to exceed cost growth. Meanwhile, Prime Video introducing advertising, leveraging partnerships and global scale of live sports content, is expected to bring higher margin revenue potential. In addition, expansion of Same-Day Delivery service and deepening of Grocery business will also promote faster GMV growth and market share gain.

Analysis framework

Bottom-up Backlog Decomposition Method: The institution abandoned the previous practice of using total Performance Obligations (PO) and average contract term to derive revenue, instead adopting a finer "Stacking Methodology". Specific steps include: 1) Identify and isolate independent backlog orders for Bedrock platform; 2) Stack separately announced large contracts for OpenAI and Anthropic; 3) Remaining part belongs to core non-AI business. This method captures contract term differences among different business lines (especially high-growth AI business) more accurately (e.g., OpenAI initial agreement amortized over 7 years, subsequent agreements over 8 years). Revenue Conversion Linear Assumption and Calibration: When converting backlog to actual revenue, the institution adopted linear conversion assumption, combined with specific contract start times obtained through communication with the company (e.g., OpenAI second agreement starts partially in 2Q26). While acknowledging linear assumption may have deviations at quarterly level, it believes annual dimension provides more reliable revenue guidance. Free Cash Flow Valuation Framework: At valuation level, the institution insists on using Price/Free Cash Flow (P/FCF) multiple method, because Amazon management has historically emphasized free cash flow generation capability. By forecasting free cash flow for next 12 months (to 2Q28E), and applying 30x valuation multiple, the target price is derived. This method avoids distortion from solely relying on net profit, better aligning with Amazon's cash return logic after heavy capital investment.

Methodology notes

• Industry/Industrial Analysis FrameworkVolume-Price Split

  Splitting AWS total backlog into core non-AI, Bedrock, and major AI lab contracts

  By breaking down overall business data into sub-items with different drivers (e.g., traditional cloud vs. AI cloud), one can more clearly identify growth sources and quality, avoiding structural changes being masked by aggregate data.

• Valuation MethodFCFF/FCFE Free Cash Flow

  Valuation method based on free cash flow (FCF)

  For tech companies with massive capital expenditure and in expansion phase, free cash flow reflects true cash-generating ability and shareholder return potential better than net profit, thus often used as core valuation anchor.

• Company Fundamentals and Financial Framework

  Model converting Performance Obligations to Revenue

  Using reported unfulfilled performance obligations (i.e., backlog) combined with average contract terms to build an "Order-to-Revenue" conversion waterfall model for forecasting future medium-to-long term revenue visibility.

Asset mapping & comparison

Structured mapping from thesis to named assets (strengths, weaknesses, peers, risks).

• Amazon.com (AMZN.US)

  Direct beneficiary, core carrier for AWS AI business growth and e-commerce margin improvement

  Strengths

  Possesses large Bedrock backlog and top-tier AI lab partners (OpenAI, Anthropic); E-commerce unit economics continuously improving; Prime Video advertising monetization potential is high.

  Weaknesses

  High capital expenditure intensity, potentially impacting free cash flow in short term; Quarterly revenue forecast subject to fluctuation risk due to contract conversion rhythm.

  Comparison

  Compared to market consensus, UBS forecasts are more optimistic for AWS medium-to-long term revenue and profit, believing market has not fully priced in structural growth brought by AI.

  Risks

  Intensifying competition; Deterioration of consumer sentiment; Capital expenditure exceeds expectations.

Key data

• Bedrock Backlog Orders (End of 1Q26)$35 billionYoY Growth approx. 340%
• 2026 AWS Revenue Forecast$175.9 billionYoY growth 36%, higher than market consensus of $166.6 billion (+29%)
• AI Revenue % of AWS (2026E Exit Rate)26%Approx. 10% in 1Q26, forecasted to reach 30% in 2027
• 2027 Operating Profit Forecast DifferenceApprox. 40% above consensusDue to higher AWS revenue forecast and operating leverage
• Target Price$333.00Based on 30x FCF for trailing 12 months ending 2Q28E
• Estimated Free Cash Flow (2Q27E-2Q28E)$121.8 billionBasis for valuation calculation

Impact & implications

Correction of Market Expectations: The report argues that the market currently underestimates AWS growth potential, especially incremental impact brought by AI business. As large deals like OpenAI and Anthropic are gradually confirmed and converted to revenue, Wall Street will have to upgrade income and profit expectations for Amazon 2026-2027 upwardly. Revaluation Logic: Amazon currently trades based on 2027 estimated GAAP P/E of merely 18x. The report considers that as a company possessing high-quality AI assets and high-growth cloud business, it should not trade at multiples below market average. With the release of the "coiled spring thesis", there is space for stock price to outperform broader market further. Business Synergy Effects: AI drives not only AWS but also forms positive cycle across the whole company by optimizing e-commerce logistics efficiency (unit economic improvement) and enhancing Prime Video content attractiveness (advertising monetization).

Risks

• Intensity of competition increases from online and offline retail competitors
• Deterioration of consumer sentiment affecting trading volume and speed
• Capital expenditure intensity for e-commerce platform or AWS higher than expected, dragging down free cash flow forecast

What to watch

• Scale and composition of AWS new backlog orders (especially AI-related portion)
• Specific rhythm of revenue conversion from large contracts with OpenAI and Anthropic
• Sustainability of improvement in e-commerce business unit economics
• Progress of Prime Video advertising business scaling