TiDB, powered by PingCAP

Manus launched in March 2025 and went viral overnight. Two million waitlist signups. Agentic workloads generating thousands of stateful iterations per task. And a monolithic database that wasn't built for any of it. They migrated to TiDB Cloud in two weeks. The result: a successful public launch, zero re-architecture of application code, and a data layer elastic enough to support Manus 1.5's agent swarms exploring thousands of parallel solution paths simultaneously. The lesson? When agentic AI scales faster than expected, your database either scales with it or becomes the bottleneck. Read the full case study to see how they pulled it off. 🔗 https://ow.ly/EQ1a50YO6zf #AgenticAI #TiDB #AIInfrastructure #DistributedSQL

That's a wrap at the Hong Kong Web3 Festival 2026! It was an incredible week of high-intensity conversations. Our big takeaway? The future of Web3 depends on a data layer that is truly AI-ready and built for massive scale. A huge thank you to everyone who stopped by! From the teams at HashKey to the academic researchers and builders who shared their insights—your energy made this event a success. If you’re building the next generation of data-intensive dApps, let’s keep the conversation going: https://pingcap.com #HKWeb3Festival #TiDB #Web3Infrastructure #HongKongTech

Exciting news! TiDB is heading to Montreal for the 42nd IEEE International Conference on Data Engineering (ICDE 2026)! As the world shifts toward autonomous systems, the data layer needs to keep up. We’re proud to sponsor this year’s event and showcase how TiDB—the distributed SQL database built for AI agents—is bridging the gap between academic research and enterprise reality. Visit the TiDB booth to: - Meet our experts to talk high-concurrency and low-latency architecture. - Discuss scaling agentic workflows and handling unpredictable data bursts. - Connect with the OSS community behind TiKV. See you in Montreal from May 4th–8th! #ICDE2026 #DistributedSQL #AI #DataEngineering #TiDB #Database

There's a type of architectural debt that doesn't announce itself. It builds in the gap between how your database was designed and how your application actually runs today, especially once AI workloads enter the picture. Unpredictable query patterns. Concurrency profiles that don't match the original assumptions. Scale that arrived faster than the infrastructure could adapt. The engineers solving this aren't waiting for it to become a crisis. They're the people coming to TiDB SCaiLE Europe 2026. Thursday, 4 June | Epicenter Stockholm. Distributed SQL, agentic AI data architecture, and production-grade case studies from teams operating at real scale. Spots are limited. Agenda dropping soon. Register → https://ow.ly/knaW50YNV14

Great insights from Vitalii Ionov! We had a blast hosting this workshop for the GMI Startup Program. TiDB is built to help AI startups scale effortlessly, and it’s a pleasure to support founders who are pushing the boundaries of what’s possible. Modern AI startups need infrastructure that grows as fast as their user base, and it's a privilege to provide the backbone for that journey with TiDB.

234 databases. 6 engineers. Zero downtime. We're bringing Plaid's Platform Architecture Lead Zander Hill on for a live webinar on May 21 to break down exactly how Plaid migrated 234 databases across 100 services from Amazon Aurora MySQL to self-hosted TiDB. In 2.5 years, with a small team, and three zero-downtime TiDB upgrades along the way. If you're evaluating Aurora alternatives or planning a large-scale migration, this is the real-world playbook you won't find in documentation. 📅 May 21 | 10-11am PST Register: https://ow.ly/Zx8b50YNvhK #TiDB #DatabaseMigration #AuroraMigration #DistributedSQL #PlatformEngineering #MySQLMigration

The industry is moving from experimental chatbots to fully autonomous agentic workflows — and the data layer has to move with it. TiDB is headed to the AI Agent Conference in New York (May 4–5, New York Hilton Midtown). Come find us at the expo hall to talk agent memory architecture, high-concurrency state management, and what it actually takes to build scalable, agent-first applications in production. If you're in New York and building in the agentic AI space, let's connect: https://ow.ly/l8kO50YNurP #AIAgentConf #GenerativeAI #TiDB #NYC #AIInfrastructure #DistributedSQL

65% infrastructure cost reduction. 10x fewer LLM tokens. 200,000+ IoT devices on a real-time multi-agent platform. Bernard Kavanagh will show you how at Data Innovation Summit - and the architecture is replicable. Tickets are going fast. 🎟️ Link in the post below.

Data Platform & Architecture Stage on DIS 2026 – What to Expect? https://hubs.li/Q04cJkpj0 As we near the 11th edition of the Data Innovation Summit in Stockholm, the focus for technology leaders has intensified on the structural integrity of the modern enterprise – a necessity when it comes to industrial-scale execution. The Data Platform & Architecture Stage serves as arena for this discussion, moving beyond the surface level of applications to the core systems that make the autonomous enterprise possible. The success of any AI strategy is dictated by the strength of its underlying architecture; here, we connect legacy reliability and high-performance, agentic ecosystems of the future, ensuring that platforms will not be just a repository, but a scalable engine for industrial innovation. Secure Your Spot: Check the full agenda and reserve one of the remaining tickets today. https://hubs.li/Q04cJfQ-0 See you in Stockholm! #hyperight #DISummit2030 #DIS2026x1

Most engineers struggle to explain complex ideas without losing half the room. If you're doing technical talks, developer relations, or teaching anything to engineers - you know the challenge. How do you keep people engaged while actually making sense? We built an AI voice journal that learns from your talk transcripts and becomes a tutor. It knows your explanations, your analogies, your style. Ask it questions later and it answers like you would. ✔️ Stores talk context in vector embeddings so the AI remembers your actual words ✔️ Scales to hundreds of transcripts without slowing down 💡 Your unique voice stays consistent - it's not generic AI advice The technical part? We used TiDB's vector search to make retrieval fast enough that the AI feels responsive, not clunky. This isn't about replacing speakers. It's about giving your best explanations a second life - helping people understand on their own time. What's your biggest challenge when trying to make technical concepts stick with your audience? 🔗 https://ow.ly/snz650YHKF0 #AI #ArtificialIntelligence #GenAI #TiDB #DeveloperRelations #AI #VectorSearch #TechTalks