MatrixOne Hyper-Converged AI Database

MatrixOrigin was founded in 2021 and operates through Matrix Origin (Shenzhen) Information Technology Co., Ltd. It builds a unified data foundation for enterprise AI, aimed at companies that run separate systems for transactions, analytics, full-text search and vector search and struggle to govern data used by AI agents.

MatrixOne is a cloud-native, MySQL-compatible database that handles OLTP, OLAP, vector and full-text workloads in one engine, with storage-compute separation on S3-style object storage and built-in IVF and HNSW vector indexes. Its Git for Data features add snapshots, branches, merges, time travel and rollback so AI agents can work on isolated copies of production data. It powers MatrixOne Intelligence, a multimodal RAG and agent platform, and serves more than 100 enterprises such as China Mobile, Lenovo, Amway and Jiangxi Copper.

• HTAP engine: OLTP, OLAP, vector and full-text workloads with workload isolation.
• Storage-compute separation: compute, transaction and S3-based storage layers with auto-scaling.
• Vector search: built-in IVF and HNSW indexes with hybrid vector and full-text queries.
• Git for Data: snapshot, branch, diff, merge, time travel and rollback through SQL extensions.
• MySQL compatibility: existing MySQL tools and applications connect without rewrites.
• Flexible editions: free Community Edition on GitHub and Enterprise Edition with SLAs, on edge, private or public cloud.

• Manufacturers in Vietnam combining MES transactions, sensor time series and document search in one database.
• Enterprises replacing MySQL plus Elasticsearch plus a vector store for internal RAG assistants.
• Energy and engineering firms automating bid and contract review on private infrastructure.

• Replaces separate transactional, analytical, Elasticsearch and vector stores with one MySQL-compatible engine.
• Branching on SF100 data was 4.1x to 13.7x faster than DoltDB for data cleaning in the vendor benchmark.
• StoneCastle cut complex aggregations from 10-30 minutes to seconds; ETAO Tech merged three databases into one in under a week.
• Safe agent experimentation on production data via branches, audit trail and rollback.

Enterprises adopting AI agents need data platforms that combine transactions, analytics and vector retrieval while keeping data inside their own boundaries for sovereign AI. In Vietnam and Southeast Asia, manufacturers and public bodies with data localisation needs favour converged databases they can run on private infrastructure.

Founded in 2021; legal entity Matrix Origin (Shenzhen) Information Technology Co., Ltd. Series A of over 10 million USD in August 2026, after a Pre-A+ round in 2025. Earlier backers include 5Y Capital, K2VC, Eastern Bell Capital and VNET Group. More than 100 enterprise customers.