Amazon AWS (AI Cloud + Advertising Dual Engine)

📌 Background

The global cloud computing market entered a second high-growth phase driven by AI in 2026, as enterprises migrate large volumes of traditional workloads to the cloud to run AI inference and training tasks. Amazon AWS, with its existing IaaS/PaaS customer base and extensive data center coverage, has become the primary beneficiary of this migration wave. At the same time, the traffic and user data accumulated by its e-commerce platform provide a natural monetization channel for its advertising business, creating a landscape in which cloud and advertising advance side by side.

👤 Target Customers

Mainly targets technology companies and startups that need to train and run inference on AI models, enterprise customers of all kinds looking to migrate legacy systems to the cloud, and brand advertisers seeking to monetize traffic through the e-commerce platform. The payers are corporate budgets and venture capital funding for AI startups; on the advertising side, payments come from brands and sellers.

💰 Revenue Streams

Revenue comes from three areas: First, AWS charges based on actual usage of resources such as compute, storage, and databases, and locks in long-term enterprise contracts through reserved instances and committed-use discounts; second, it uses its in-house Trainium chips and hosted model services such as Anthropic/OpenAI as selling points to drive new AI compute consumption and generate incremental charges; third, the advertising business contributes approximately $19.8B in quarterly revenue, generated through natural exposure in e-commerce and streaming scenarios.

🧮 Cost Structure

The main cost is AI-driven hyperscale capital expenditure, with full-year 2026 guidance as high as $220B, used to expand data centers, procure network infrastructure, and develop in-house AI chips. In addition, it must cover compensation for chip R&D talent, the compute provisioning costs of hosting large models, and investment in Prime membership content and operation and maintenance of its advertising technology platform.

🛡️ Moat

It has the industry's broadest portfolio of cloud services and a large existing base of enterprise customers; its massive market scale makes reserved instances and discount strategies more attractive. Its in-house Trainium chips directly lower the cost structure of AI inference and training, creating a closed loop between hardware and cloud services. At the same time, e-commerce traffic and user profile data provide advertisers with high-conversion placement scenarios, making it difficult for competitors to replicate the same scale in a short period.

🔑 Keys to Success

• Compute + model bundling agreements with Anthropic/OpenAI
• Cost advantage of in-house Trainium chips
• Conversion of existing enterprise workloads and legacy system migration to the cloud

⚠️ Risks

• Whether the AI capital expenditure of as much as $220B can generate corresponding revenue returns remains under pressure to be validated
• Azure/GCP are capturing AI cloud share at a faster growth rate, creating potential share loss
• Slowing growth in the core e-commerce business (+20%) may weigh on overall margins and reinvestment capacity

🏢 Cases

• Amazon Web Services (Q2 2026 $42.2B, +37%, run rate $169B)
• Amazon Ads ($19.8B, +26%)

📊 SWOT Analysis