Pinecone

It's 2026. Are LLM hallucinations solved? Not quite. Models only know what they were trained on, and new data never stops arriving. The fix: give your model somewhere to look things up. Learn more: pinecone.io

Pinecone BYOC customers can now create indexes in their own cloud account without sizing each one. On-demand mode for BYOC is available today. All your on-demand indexes share one capacity pool that can hold millions of namespaces, one for each tenant of your app or each user of your AI agent. The pool starts small, so you don't pay for idle hardware up front. Pinecone uses dynamic sharding to keep your data spread evenly across the pool, and as your data grows, you add shards without moving existing data. If one tenant gets far more queries than the rest, you can move it to its own index on dedicated read nodes in the same environment. Read the blog: https://lnkd.in/eXmDv_FA

Building search across text, images, audio, and video shouldn't mean building a separate search experience for every data type. Multimodal search uses vector embeddings to enable retrieval across different formats based on meaning, not just keywords or metadata. For developers, that means less reliance on manual tagging and more flexible ways for users to discover relevant content. Go deeper: https://lnkd.in/gYQdu4Fg

We’re at AI Engineer New York! Come find the Pinecone team at Booth 55. Bring your AI questions, tell us what you’re building, and let’s talk about your next steps. #AIEngineerNYC #Pinecone

See you tomorrow, New York! The Pinecone team will be at Booth 55 for AI Engineer New York. Stop by to meet the team and talk through your AI use case. The expo opens Monday at 4 PM ET, with the welcome reception on the expo floor from 5–7 PM. #AIEngineerNYC #Pinecone

What a great night at #SFTechWeek! We had an incredible time bringing together AI builders, engineers, and founders for our Agentic Knowledge Happy Hour! The evening was filled with great conversations, new connections, and a live demo from Jörg Schad , Ph.D., our VP of Engineering, showcasing how to build knowledge into AI agents at scale. A huge thank you to everyone who joined us! We love seeing this community come together.

The Pinecone team will be setting up shop at AI Engineer NYC next week - Booth 55! Swing by the Sheraton New York Times Square. Whether you're building agents or wanna talk search & recommendation - stop by for a chat & demo of Nexus. Nexus turns documents, data, and business rules into reusable knowledge, helping agents answer complex questions more accurately with less repeated retrieval and fewer tokens. Swing by booth 55 or DM me and we'll grab a coffee ☕

With namespace aliases, now available in Pinecone, swapping the namespace your app queries takes one API call instead of a code change and a redeploy. Point your app at an alias once. From then on, load and test new data in a fresh namespace while the app keeps querying the current one, then repoint the alias when it's ready. Keep the old namespace around and it's just as easy to roll back if necessary. Learn more 🔗 https://lnkd.in/gvjU2JjR

Heading to AI Engineer New York? Come meet the Pinecone team at Booth 55. Let’s talk about what you’re building with AI, your retrieval challenges, and how Pinecone can help. October 12–14 at the Sheraton New York Times Square. #AIEngineerNYC #Pinecone

LA Tech Week starts Monday, and our monthly LA Agentic AI Meetup is on day two: Tuesday, October 13, 5 to 7pm at Gulp in Playa Vista. Community member PJ Kedilaya is demoing ModelVantage, which tracks uptime and latency for every major LLM, region by region. RSVP: https://lnkd.in/gguh9qve TECH WEEK by a16z