The summer of 2025 has been a whirlwind of innovation at AWS, with groundbreaking announcements across AI, security, and developer experience — many unveiled during the AWS Summit New York event. From Amazon S3 Vectors promising 90% cost reduction for vector storage to DynamoDB’s multi-region strong consistency and Lambda’s cloud debugging capabilities, AWS continues to push boundaries. This summer saw particular emphasis on AI infrastructure scaling with EKS supporting 100K nodes per cluster and new tools like Bedrock AgentCore and Kiro IDE transforming how we build intelligent applications. We’ve gathered the most impactful announcements to help you navigate this rapidly evolving cloud landscape.
We all know, that reading long lists of technical announcements might be boring. Therefore, this time we made the announcements more, should we say, interesting.
Table of contents
- AWS Certificate Manager introduces public certificates you can use anywhere
- IAM Access Analyzer now identifies who in your AWS organization can access your AWS resources
- AWS WAF announces general availability of Resource-level DDoS protection for Application Load Balancers (ALB)
- Amazon Bedrock Guardrails announces tiers for content filters and denied topics
- Amazon S3 Tables now supports Model Context Protocol (MCP) Server
- Introducing Amazon S3 Vectors: First cloud storage with native vector support at scale (preview)
- Introducing Amazon Bedrock AgentCore: Securely deploy and operate AI agents at any scale (preview)
- Amazon EKS enables ultra scale AI/ML workloads with support for 100K nodes per cluster
- Announcing Amazon Nova customization in Amazon SageMaker AI
- Now in GA: Accelerate troubleshooting with Amazon CloudWatch investigations
- Monitor and debug event-driven applications with new Amazon EventBridge logging
- Accelerate safe software releases with new built-in blue/green deployments in Amazon ECS
- AWS Lambda enables developers to debug functions running in the cloud from VS Code IDE
- Introducing Kiro
- Amazon QuickSight launches Trusted Identity Propagation (TIP) for Athena Direct Query
- Amazon QuickSight supports 2B row SPICE dataset
- Amazon S3 Metadata now supports metadata for all your S3 objects
- Streamline the path from data to insights with new Amazon SageMaker capabilities
Security
AWS Certificate Manager introduces public certificates you can use anywhere
ACM certificates finally got passports! For $15 ($149 for wildcards), your certs can now travel beyond AWS to secure any workload. No more CloudFront-only parties—these 395-day wonders work everywhere, even that server in your closet!
Posted on: Jun 17, 2025 | Read more
IAM Access Analyzer now identifies who in your AWS organization can access your AWS resources
IAM Access Analyzer becomes the ultimate resource stalker! Now it knows exactly who’s eyeing your S3, DynamoDB, and RDS treasures. This digital detective scans policies daily, creating a 360-degree view of your access landscape. Security teams get a fancy dashboard to spot and fix suspicious access—or just automate alerts and let developers take the blame. Available in all commercial regions for your compliance team’s next PowerPoint presentation!
Posted on: Jun 17, 2025 | Read more
AWS WAF announces general availability of Resource-level DDoS protection for Application Load Balancers (ALB)
ALBs just got their own superhero cape! AWS WAF’s new DDoS protection acts like an on-site bouncer, spotting troublemakers in seconds while your legitimate traffic parties on. It even checks IDs in those sneaky X-Forwarded-For headers! Set it to “always on” or “only when we’re popular” mode—either way, your load balancers can finally relax and stop sweating about those pesky DDoS villains.
Posted on: Jun 26, 2025 | Read more
AI
Amazon Bedrock Guardrails announces tiers for content filters and denied topics
Bedrock Guardrails just got its PhD in linguistics! The new Standard tier can now spot naughty content in 60 languages (yes, even Finnish!), catch typos from sneaky users, and tell the difference between actual jailbreaks and harmless jokes. It’s like having a bouncer who studied psychology—smart enough to know when your AI is being manipulated but cool enough to let the fun conversations through. Just remember to check the “let my data travel” box if you want the premium experience!
Posted on: Jun 24, 2025 | Read more
Amazon S3 Tables now supports Model Context Protocol (MCP) Server
S3 Tables and LLMs are now best friends forever! With MCP Server, your AI assistant finally understands database-speak without you playing translator. Just say “Hey, find me all customers who spent over $1000 last month” and watch your AI do the heavy lifting while SQL queries become ancient history. It’s like giving your AI assistant the ultimate S3 Tables cheat sheet—no more awkward “I don’t know how to do that” responses. Your data projects just got a serious speed boost!
Posted on: Jul 15, 2025 | Read more
Introducing Amazon S3 Vectors: First cloud storage with native vector support at scale (preview)
Vector storage just got a 90% discount! S3 Vectors is like Marie Kondo for your embeddings—organizing billions of vectors without breaking a sweat or your bank account. No more awkward vector database setups or performance anxiety when scaling. Just toss your vectors in, ask them questions in milliseconds, and watch your AI apps suddenly perform like they’ve been drinking premium coffee. Perfect for when your RAG application needs to remember everything it’s ever read without charging you for the privilege!
Posted on: Jul 15, 2025 | Read more
Introducing Amazon Bedrock AgentCore: Securely deploy and operate AI agents at any scale (preview)
AWS just built a luxury apartment complex for your AI agents! AgentCore gives your digital minions secure IDs, long-term memory, and a toolbox that would make Batman jealous. Deploy them anywhere—from tiny IoT devices to massive server farms—without the usual security nightmares or integration headaches. It’s like a universal translator for AI frameworks that lets Claude, Llama, and Anthropic models all play nicely together. Finally, enterprise-grade AI without needing a PhD in prompt engineering!
Posted on: Jul 16, 2025 | Read more
Amazon EKS enables ultra scale AI/ML workloads with support for 100K nodes per cluster
EKS just went from bodybuilder to Godzilla! With 100,000 nodes per cluster, it’s like giving your AI workloads an entire digital country to live in. Need 1.6 MILLION Trainium chips or 800K NVIDIA GPUs? No problem—EKS doesn’t even break a sweat. Your massive language models can now party across a compute landscape the size of Manhattan while your Kubernetes tools keep working like nothing extraordinary happened. This isn’t just scaling up—it’s scaling to “yes, we can train Skynet now” territory!
Posted on: Jul 16, 2025 | Read more
Announcing Amazon Nova customization in Amazon SageMaker AI
Nova models just got their own personal stylist! SageMaker now lets you dress up these foundation models in your company’s unique fashion sense—from basic alterations to complete makeovers. With ready-made “recipes” that work like AI cookbooks, you can season your models with industry-specific flavors without becoming a gourmet AI chef. Pre-training, post-training, fine-tuning—it’s like sending your AI to finishing school where it learns your business lingo and company culture. Your competitors will wonder how your AI got so smart about your specific industry!
Posted on: Jul 16, 2025 | Read more
Operational Excellence
Now in GA: Accelerate troubleshooting with Amazon CloudWatch investigations
CloudWatch just got its detective license! This AI-powered Sherlock Holmes hunts down your AWS problems while you grab coffee. It spots suspicious anomalies, connects seemingly unrelated clues, and presents you with a “whodunit” report complete with remediation recommendations. No more late-night troubleshooting sessions or frantic Slack messages—CloudWatch Investigations has already assembled the suspects and practically solved the case. Your MTTR just went from “panic mode” to “I’ll fix that after lunch.”
Posted on: Jun 24, 2025 | Read more
Monitor and debug event-driven applications with new Amazon EventBridge logging
EventBridge just became the ultimate event paparazzi! Now it tracks your events' entire life story—from birth (published) to finding their soulmate (rule matching) to delivery and even their dramatic failures. No more event mysteries or “what happened to that notification?” questions. It’s like installing security cameras throughout your event-driven architecture that capture every juicy detail. Your events can run, but they can’t hide—every move is now documented for your debugging pleasure!
Posted on: Jul 15, 2025 | Read more
Accelerate safe software releases with new built-in blue/green deployments in Amazon ECS
ECS just got deployment training wheels that actually look cool! Blue/green deployments are now built right in—no more cobbling together custom scripts or praying to the deployment gods. Ship new code with the confidence of someone who always has a backup plan, because you do! If your new version starts acting weird, just flip the switch and instantly roll back like nothing happened. It’s like having an “oops” button for your containers that doesn’t make you look bad in front of your boss. Ship more often, sleep better at night!
Posted on: Jul 17, 2025 | Read more
Developer Experience
AWS Lambda enables developers to debug functions running in the cloud from VS Code IDE
Lambda functions can no longer hide their bugs in the cloud! VS Code now reaches directly into AWS to catch your functions red-handed when they misbehave. No more “works on my machine” excuses or deploying 17 times with added console.log statements. Just set your breakpoints, hit debug, and watch as your Lambda spills all its secrets without leaving your favorite editor. It’s like having X-ray vision into the cloud—finally making “serverless” feel a little less “codeless” when things go wrong!
Posted on: Jul 17, 2025 | Read more
Introducing Kiro
Meet your new coding buddy with superpowers! Kiro isn’t just another IDE—it’s an AI pair programmer that actually understands what you’re trying to build. It writes code while you sip coffee, debugs your mistakes before you make them, and even helps architect entire systems from vague requirements. Think of it as having a senior developer who never sleeps, never complains about your variable names, and costs less than a daily latte. From “I have an idea” to “it’s in production,” Kiro is the sidekick that makes you look like the coding hero you always knew you could be!
Posted on: July 14,2025 | Read more | Check video
Databases
Amazon DynamoDB global tables with multi-Region strong consistency is now generally available
DynamoDB just achieved the impossible—global tables that are both everywhere AND in sync! It’s like having identical twins who can read each other’s minds across continents. Your data now enjoys first-class travel between regions with zero jet lag (RPO of zero). Need to survive a regional apocalypse? No problem—your database will keep humming along without missing a beat. The “eventually” in “eventually consistent” just got upgraded to “right now, thank you very much.” Physics professors are scratching their heads while your global applications are celebrating!
Posted on: Jun 30, 2025 | Read more
Data and analytics
Amazon QuickSight launches Trusted Identity Propagation (TIP) for Athena Direct Query
QuickSight just got a bouncer at the data door! With Trusted Identity Propagation, your dashboards now know exactly who’s asking for what data and can enforce VIP-only access to your sensitive insights. It’s like your Lake Formation rules are now bouncers checking IDs before letting anyone see those exclusive analytics. No more “one dashboard fits all” compromises or building seventeen different versions for different teams. Your data governance team can finally sleep at night while your business users still get all the insights they need—just not the ones they shouldn’t have!
Posted on: Jul 1, 2025 | Read more
Amazon QuickSight supports 2B row SPICE dataset
QuickSight just doubled its appetite! SPICE can now gobble up a whopping 2 BILLION rows of data without getting indigestion or slowing down. It’s like upgrading from a compact car to a monster truck while somehow keeping the same gas mileage. Your dashboards can now handle twice the historical data or double the business categories without making you wait any longer for insights. Perfect for when your boss asks, “Can we see the last five years instead of just two?” and you want to say “yes” without sweating about performance. Big data just got twice as big but stayed just as speedy!
Posted on: Jul 2, 2025 | Read more
Amazon S3 Metadata now supports metadata for all your S3 objects
S3 just got X-ray vision for your objects! The new Metadata service is like having a meticulous librarian who catalogs every single file in your buckets and updates their index cards hourly. Now you can SQL-query your way through billions of objects without waking them up or paying retrieval fees. It’s like being able to search through your storage attic without actually opening all those dusty boxes. Perfect for when you need to answer “what do we actually have in there?” questions from management without spinning up a data engineering project. Your objects can’t hide their secrets anymore!
Posted on: Jul 15, 2025 | Read more
Streamline the path from data to insights with new Amazon SageMaker capabilities
SageMaker just became the Switzerland of your data world—neutral territory where all your data types can finally get along! It’s now the ultimate matchmaker connecting your structured data, unstructured files, and analytics tools in one harmonious ecosystem. QuickSight dashboards? Connected. S3 documents and media? Cataloged. Lakehouse data? Automatically onboarded. It’s like your data finally joined the same social network and started sharing updates with each other. No more awkward data blind dates or playing telephone between systems—your insights pipeline just went from “it’s complicated” to “in a relationship.”
Posted on: Jul 15, 2025 | Read more
Everything else
AWS Free Tier update: New customers can get started and explore AWS with up to $200 in credits
AWS is now paying you to learn their platform! The Free Tier just got a $200 upgrade—half just for showing up and half for actually trying things out. It’s like getting paid to test drive a Ferrari, with $100 just for sitting in the driver’s seat and another $100 for taking it around the block. Play with EC2, dabble in Bedrock AI, or set up a budget (ironically, to track the free money they gave you). Perfect for cloud-curious developers who’ve been hesitating to jump in. Your wallet and your resume will thank you!
Posted on: Jul 15, 2025 | Read more