Prime Day 2026 Scale and the Infrastructure That Absorbed 2.3 Quadrillion Requests
Amazon Prime Day 2026, powered by AWS, processed massive structural metrics in June, highlighted by Amazon ElastiCache peaking at 2.3 quadrillion daily requests, as reported by aws.amazon.com. Over the four-day event, Amazon DynamoDB handled over 59 trillion requests with single-digit millisecond response times while peaking at 192 million requests per second. AWS Lambda recorded over 2.3 trillion daily invocations, and Amazon EBS crossed 24.8 trillion I/O operations, moving over an exabyte of data daily. Furthermore, Amazon ECS launched an average of 158.3 million tasks per day on AWS Fargate—a 47.7 percent increase over the previous year—while AWS Graviton powered up to 49 percent of the Amazon EC2 compute utilized by Amazon.com, according to aws.amazon.com data.
Behind those throughput figures lay thousands of deliberate stress tests. The engineering team executed more than 44,000 AWS Fault Injection Service experiments prior to the event, marking a sixfold increase compared to 2025.
Claude Haiku 5.5 and Nova 2.5 Sonic Expand AI Tooling on AWS
Shifting toward smaller, cost-optimized deployments, AWS introduced Claude Haiku 5.5 on Amazon Bedrock and Claude Platform, priced at roughly 75 percent less than Haiku 4.5 for most tasks. It represents the first Haiku model equipped with effort controls, allowing engineers to balance token cost against intelligence per task. It pairs with Claude Opus 5.5 for complex planning while running parallel subagents for routing, classification, and minor code edits.
Concurrently, Amazon Nova 2.5 Sonic launched to support real-time speech-to-speech voice agents featuring a 256K context window, controllable turn-taking, and expressive voices across seven languages. It integrates with Strands Bidi Agents, which reached general availability to streamline multi-step conversational workflows.
GuardDuty RDS Protection and Kafka OAuth Governance
Security updates arrived with Amazon GuardDuty RDS Protection expanding behavioral baseline monitoring to Amazon Aurora PostgreSQL and Amazon RDS for PostgreSQL. By applying machine learning anomaly detection directly to incoming queries, GuardDuty identifies data exfiltration and destruction attempts without requiring agents or database configuration changes. Additionally, AWS Lambda supports OAuth for self-managed Apache Kafka event sources.
