MongoDB President and CEO @dittycheria recaps the largest expansion of our platform yet from the #MongoDBlocal NYC stage. We introduced the foundation: MongoDB 9.0, the fastest version of MongoDB we’ve ever built, designed for applications and agents that act on live data. Elasticity at any scale: Atlas Infinite, which scales compute and storage independently as demand changes. Agents built where your data lives: Atlas Agent Engine provides the memory, runtime, and governance to run agents in production. It sits on the data foundation customers already trust and works with any model, framework, or cloud. Watch the full keynote: mongodb.social/6017BGIi19

MongoDB Atlas is now a first-class data layer for agentic coding in @SpaceXSI. Powered by the new MongoDB Atlas Managed MCP Server, agents can provision clusters, load data, and stand up Vector Search all from natural-language prompts, with human approval built in. “Developers should be able to focus on what they want to build, not get slowed down by the tools behind it. Our collaboration brings together SpaceXSI’s AI-assisted development experience and MongoDB’s trusted data platform, helping teams move more smoothly from idea to production. Together, we’re giving organizations more choice in how they build and the confidence to scale what they create.” - Toni Adams, Head of Global Partnerships, SpaceXSI See it in action 👉 mongodb.social/6013BHnISv

Can your data architecture keep up with your business? Our 3️⃣ new Tech Leadership & Architecture Skill Badges help you answer that. Start learning: mongodb.social/6019BHpI05

Paginated search has a quiet failure mode: results duplicate or vanish between pages because relevance scores drift across replicas. We fixed it with StableTfl, a new Lucene scoring function that computes relevance from the query and document alone. Read how it works: mongodb.social/6016BHrgwA

Most teams scaling in the cloud pay for more than their workload needs. ☁️ 🚇 Principal Technologist, @Dev_TheAnalyst, on why Atlas Infinite, with independent storage and compute scaling, deserves your attention.

From searching data to understanding the connections that make it meaningful. For @LexisNexisRisk Solutions, where decisions worth millions of pounds depend on data accuracy, AI needs to understand relationships, business context and time, because in financial markets, two seemingly contradictory facts can both be true at different points in time. To tackle this challenge, LexisNexis’s three-person Maverick team set out to go beyond the limitations of traditional RAG and build a time-aware reasoning system. Using MongoDB Atlas, they brought ontologies, knowledge graphs, vectors and transactional data together in one place, creating a unified foundation for context-aware AI. As James Melvin, Principal Innovation Software Engineer at LexisNexis Risk Solutions, explains: “MongoDB helps us to run one central data store that holds everything… it’s literally the store of our AI.” By unifying its AI stack on MongoDB Atlas, the team eliminated the complexity of synchronising data across multiple systems and built a foundation designed to deliver more accurate, contextual answers for high-stakes financial decisions. Discover how LexisNexis Risk Solutions is rethinking enterprise AI search with MongoDB: mongodb.social/6019BHJQyF

MongoDB 9.0 is out, and it includes three observability updates for Atlas. 1️⃣ Atlas events now integrate with OpenTelemetry, helping you correlate configuration changes, access updates, and scaling events with metrics and logs in your existing observability tools. 2️⃣ The redesigned Metrics UI provides a unified, cluster-level view of database health and performance, with node selectors to help you focus your investigation. 3️⃣ Query Shape Insights now covers inserts, updates, and deletes, helping you identify inefficient write operations and uncover opportunities to optimize performance and resource utilization. Read the details: mongodb.social/6015BH1SEB

Your AI agent works in the demo. Can you govern, observe, and evaluate it in production? Three new MongoDB Skill Badges show you how 👉 mongodb.social/6012BG7VFg

4 to 6 hours. That's how long our sellers spent gathering context per account, per deal cycle. So we built Holly, an AI workspace on Atlas Agent Engine that brings research, meeting prep, and outreach into one place. Reps using it cut prospecting research time by 95%. Check it out: mongodb.social/6014BG7VnG

We asked builders at MongoDB.local one question: light mode or dark mode? 🌙☀️ Tell us where you land.

From system of record to system of action. Here's what that means, straight from #MongoDBlocal NYC: ✨ Atlas Agent Engine: Build, deploy, and govern agents in production, with built-in memory and one control plane to verify every agent action. 🚀 Atlas Infinite: Compute and storage scale independently to handle unpredictable workloads with greater elasticity and responsiveness. Check out all the updates: mongodb.social/6017BG442T

Keep your clusters responsive and predictable, even under overload. Introducing Intelligent Workload Management (IWM), which is on by default for eligible dedicated MongoDB 9.0 Atlas clusters. It comes at no additional cost and requires zero application changes. Learn how IWM enhances MongoDB Atlas availability: mongodb.social/6016BGLwus

Your database used to just remember things. Now it has to make decisions at 2am without you. New IDC research says the shift is already here: 🤖 30% of organizations already run agentic AI in production 💰 Another 50% are investing behind it ⌛ 64.6% of use cases now require real-time or near-real-time delivery Four in five enterprises are past deciding if agents belong in their operations. They're working out how to run them safely. Read the IDC Whitepaper, sponsored by MongoDB: mongodb.social/6014BG255A

Introducing Atlas Agent Engine .🚀 Announced at #MongoDBlocal NYC, Atlas Agent Engine is MongoDB's platform for building AI agents directly on your data. Get a secure way to build, deploy, and govern agents in production. Agents act on the business context you already have, right where your data lives, with the framework and model of your choice. It is governed by default, with memory and guardrails built in. Watch the keynote highlights: mongodb.social/6015BGLZQH