 |
Alongside launching a new serverless region on AWS ap-southeast-1, Pinecone also unveiled the Nexus knowledge engine, KnowQL query language, Marketplace, Builder plan, and native full-text search capabilities
SingaporeMay 5, 2026 /PRNewswire/ — Pinecone, a knowledge infrastructure provider for AI applications at scale, today announced the launch of its first serverless region in Asia on AWS Asia Pacific (Singapore) (ap-southeast-1), bringing its full suite of serverless vector databases and knowledge infrastructure to the Asia-Pacific market. This new region enables enterprises in Southeast Asia, Australia, and across the broader Asia-Pacific region to run AI workloads with low-latency performance while meeting local data residency compliance requirements.
Alongside the Singapore region launch, Pinecone also announced a series of major product releases, including: Pinecone Nexus—a knowledge engine for AI agents; KnowQL—a declarative query language for agent retrieval; Pinecone Marketplace—featuring over 90 ready-to-use knowledge applications; a new Builder plan at $20 per month; native full-text search capabilities for the Pinecone core database; and dedicated read nodes—designed for high-throughput, low-latency production workloads, reducing costs by up to 97% at scale.
Ash Ashutosh, CEO of Pinecone, said: “High-quality knowledge infrastructure should be accessible to all developers, in all regions, worldwide. Launching our first serverless region in Asia is a significant milestone for Pinecone. Enterprises in the Asia-Pacific region can now use the same infrastructure trusted by over 9,000 customers globally, while meeting the data residency, low-latency, and proximity requirements of businesses in this region. Combined with today’s simultaneous releases of Nexus, KnowQL, Marketplace, and the new Builder plan, we are building the most comprehensive AI knowledge infrastructure in the industry, removing all barriers to AI development with scalable capabilities and affordable pricing.”
Pinecone Nexus: A Knowledge Engine for Agents
As AI evolves from assistive applications to autonomous agents, the demand for knowledge infrastructure is undergoing a fundamental shift. Currently, agents spend approximately 85% of their workload on context retrieval, resulting in task completion rates of only 50-60%, unpredictable task durations, and high token costs. Pinecone Nexus addresses this pain point by moving reasoning upstream—from “retrieval” to the “knowledge compilation” stage. Its context compiler transforms raw data into standardized, task-adapted components ready for direct agent use; composable retrievers fetch relevant components with low latency, support field-level citation tracing, and resolve data conflicts through deterministic mechanisms. Early test results show token usage reduced by up to 90%, task completion rates exceeding 90%, and task speeds improved by 30 times.
At the core of Nexus is KnowQL, a declarative query language that provides a unified knowledge access interface for agents. Without the need for custom tool definitions or dedicated glue code, a single call specifies output format, citation requirements, and latency thresholds, returning trustworthy, structured, and verifiable knowledge.
Pinecone Marketplace
Pinecone Marketplace is a catalog of fully functional, ready-to-use knowledge applications that users can deploy, customize, and run immediately. At launch, it offers over 90 mature solutions spanning categories such as sales and revenue, insurance, real estate, legal and compliance, human resources, and customer support. These applications, built by Pinecone and its partners, require no underlying infrastructure setup by enterprises. The Marketplace is free to use at launch, with partner-built commercial solutions to follow.
New Pricing and Infrastructure Upgrades
Pinecone’s new Builder plan, at just $20 per month, provides full access to its production-grade infrastructure with complimentary technical support. Dedicated read nodes offer provisioned read capacity, using a warm data path and fixed hourly billing, targeting sustained high-throughput business scenarios, saving 77% to 97% in costs at scale. For enterprises with data residency or compliance requirements, Bring Your Own Cloud (BYOC) enables deployment of Pinecone’s fully managed service within the customer’s own cloud environment. Additionally, Pinecone has integrated native full-text search into its core database, now in public preview. This feature supports hybrid retrieval, balancing the breadth of semantic search with the precision of exact matching, to handle query traffic at agent scale.
The Singapore region is now officially available for use. This is Pinecone’s first serverless region in Asia, following previous serverless region deployments on AWS in the US and Europe, with plans to add more regional nodes in the future.
Visit here to get started. For more details on today’s announcements, please refer to the blog.