re:Invent Recap
It’s been a while since I last attended re:Invent… long enough that I’d almost forgotten how expensive a bottle of water can be in a Vegas hotel room. This time was different. Instead of just attending, I wore many hats: audio-visual tech, salesperson, technical support, friendly ear, and booth rep. re:Invent is an experience that’s hard to explain to anyone outside tech. Picture 65,000 people converging in Las Vegas: DJ booths thumping beside deep-dive technical sessions, competitions running nonstop, and enough swag to fill a million Christmas stockings. Only then do you start to grasp what it’s really like. Needless to say, having the privilege to fly halfway across the globe, stay in a bougie hotel, and help host the impressive ScyllaDB booth was a fitting way to finish the year on a high. This year was ScyllaDB’s biggest re:Invent presence yet… a full-scale booth designed to show what predictable performance at extreme scale really looks like. The booth was buzzing from open to close, packed with data engineers, developers, and decision-makers exploring how ScyllaDB handles millions of operations per second with single-digit P99 millisecond latency. Some of the standout moments for me included: Customer sessions at the Content Hub featuring Freshworks and SAS, both showcasing how ScyllaDB powers their mission-critical AI workloads. In-booth talks from Freshworks, SAS, Sprig, and Revinate. Real users sharing real production stories. There’s nothing better than hearing how our customers are conquering performance challenges at scale. Technical deep dives exploring everything from linear scalability to real-time AI pipelines. ScyllaDB X Cloud linear-scale demonstration, a live visualization of throughput scaling predictably with every additional node. Watching tablets rebalance automatically linearly never gets old. High-impact in-booth videos driving home ScyllaDB’s key differentiators. I’m particularly proud of our DB Guy parody along with the ScyllaDB monster on the Vegas Sphere (yes, we fooled many with that one) For many visitors, this was their first time seeing ScyllaDB X Cloud and Vector Search in action. Our demos made it clear what we mean by performance at scale: serving billions of vectors or millions of events per second, all while keeping tail latency comfortably under 5 ms and cost behavior entirely predictable. Developers that I chatted to loved that ScyllaDB drops neatly into existing Cassandra or DynamoDB environments while delivering much better performance and a lower TCO. Architects zeroed in on our flexibility across EC2 instance families (especially i8g) and hybrid deployment models. The ability to bring your own AWS (or GCP) account sparked plenty of conversations around performance, security, and data sovereignty. What really stood out this year was the shift in mindset. re:Invent 2025 confirmed that the future of extreme scale database engineering belongs to real-time systems … from AI inference to IoT telemetry, where low latency and linear scale are essential for success. ScyllaDB sits right at that intersection: a database built to scale fearlessly, delivering the control of bare metal with the simplicity of managed cloud. If you missed us in Vegas, don’t worry … you can still catch the highlights. Watch our customer sessions and the full X Cloud demo, and see why predictable performance at extreme scale isn’t just our tagline. It’s what we do every day. Catch the re:Invent videos
