On-demand mode for Pinecone BYOC: scale without sizing every index

Starting today, Pinecone Bring Your Own Cloud (BYOC) supports on-demand capacity mode, so you can create indexes without sizing each one. 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. You start with a small pool and add capacity as your data grows.

Pinecone BYOC runs Pinecone inside your own AWS, Google Cloud, or Azure account, where your data stays under your control while Pinecone operates the database. Teams in legal, healthcare, and financial services often choose it to keep their data in their own cloud.

How on-demand mode works

If you run semantic or lexical search for your company, most systems make you choose the number of shards for every index when you create it. If you choose too few, you rebuild the index when it outgrows them. If you choose too many, you pay for hardware that sits idle. Adding shards later brings its own problem. Existing data has to move to the new shards, and some shards still end up much fuller than others.

On-demand mode uses dynamic sharding, which tracks where each part of every index lives instead of tying an index to a fixed set of shards. As your data grows, you can add shards without rebuilding anything or moving existing data. Pinecone sends new data to the shards with the most room to keep your data spread evenly across the pool.

If one namespace gets far more queries than the others, you can move it to its own index on dedicated read nodes in the same BYOC cluster. Your other namespaces keep sharing the on-demand pool.

Security and support

Scaling your on-demand pool keeps the zero-access BYOC model. Your vectors, metadata, and queries stay in your cloud account. Pinecone adds capacity without SSH, VPN, inbound network access, or a standing cross-account role. Pinecone monitors the cluster's health through operational metrics like CPU and latency, without seeing your data. When you need help, you share logs with Pinecone support so we can troubleshoot with you.

Get started

On-demand mode is available now for every BYOC customer. If you're new to BYOC, Pinecone's setup wizard generates a deployment project that you run in your cloud account to create your cluster. If you already run BYOC, your cluster is ready. To create an on-demand index, pick your BYOC cluster as the location in the Pinecone console, API, or SDKs and choose on-demand as the capacity mode. To plan your cluster, talk with your Pinecone account team or contact us.

For more detail, the BYOC guide covers the architecture, how data is secured, and the BYOC FAQ.