𝗧𝗵𝗲 𝗠𝗶𝗹𝘃𝘂𝘀 𝟯.𝟬 𝗥𝗼𝗮𝗱𝗺𝗮𝗽 𝗶𝘀 𝗵𝗲𝗿𝗲 — and we’re incredibly excited to share what’...
- Milvus 3.0 扩展至三大方向：原地数据搜索、更丰富的语义查询引擎、Lakebase工作流基础。
- 新增功能包括外部集合、Iceberg等格式支持、语义排序与聚合、FAISS索引家族等。
- Zilliz Cloud Lakebase集成流处理、导入导出、在线查询等功能，促进AI数据基础设施循环。
结构提纲
按章节快速跳转。
- §引言
介绍 Milvus 3.0 路线图发布，强调对AI应用的支持升级。
阐述Milvus 3.0在数据原地搜索、语义查询引擎和Lakebase工作流上的主要扩展。
介绍如何使数据湖中的数据无需迁移即可进行搜索。
展示超越最近邻搜索的功能，如排序、聚合等。
介绍Zilliz Cloud Lakebase如何整合服务与发现于统一数据层。
思维导图
用一张图看清主题之间的关系。
查看大纲文本（无障碍 / 无 JS 友好）
- Milvus 3.0 路线图
- 核心扩展方向
- 原地数据搜索
- 语义查询引擎增强
- Lakebase工作流基础
- 目标用户
- AI团队
- 平台特色
- Zilliz Cloud
- 数据基础设施循环
金句 / Highlights
值得收藏与分享的关键句。
Milvus 3.0 不仅限于在线服务路径，正成为帮助AI团队服务、发现、改进再服务的数据基础设施环的一部分。
新增External Collection、Iceberg支持等特性，让数据湖中的数据可直接搜索。
Zilliz Cloud Lakebase集成多种功能，如流处理、在线查询，支持去重、异常检测和聚类等生产任务。
For years, Milvus has been known for production vector search: high QPS, low latency, and 100B+ scale for online AI applications.
But AI teams are now asking for more https://t.co/vxQZrMwV8H" / X
Milvus on X: "𝗧𝗵𝗲 𝗠𝗶𝗹𝘃𝘂𝘀 𝟯.𝟬 𝗥𝗼𝗮𝗱𝗺𝗮𝗽 𝗶𝘀 𝗵𝗲𝗿𝗲 — and we’re incredibly excited to share what’s next. For years, Milvus has been known for production vector search: high QPS, low latency, and 100B+ scale for online AI applications. But AI teams are now asking for more https://t.co/vxQZrMwV8H" / X
JavaScript is not available.
We’ve detected that JavaScript is disabled in this browser. Please enable JavaScript or switch to a supported browser to continue using x.com. You can see a list of supported browsers in our Help Center.
Terms of ServicePrivacy PolicyCookie PolicyImprintAds info © 2026 X Corp.
Don’t miss what’s happening
People on X are the first to know.
Post
See new posts
Conversation
𝗧𝗵𝗲 𝗠𝗶𝗹𝘃𝘂𝘀 𝟯.𝟬 𝗥𝗼𝗮𝗱𝗺𝗮𝗽 𝗶𝘀 𝗵𝗲𝗿𝗲 — and we’re incredibly excited to share what’s next. For years, Milvus has been known for production vector search: high QPS, low latency, and 100B+ scale for online AI applications. But AI teams are now asking for more than serving. They still need fast vector search in production. They also need to explore, search, deduplicate, cluster, and analyze massive embedding datasets that already live in data lakes — without copying everything into another system first. That is the next chapter Milvus 3.0 is moving toward. This roadmap expands Milvus in three major directions: 𝟭. 𝗦𝗲𝗮𝗿𝗰𝗵𝗶𝗻𝗴 𝗱𝗮𝘁𝗮 𝘄𝗵𝗲𝗿𝗲 𝗶𝘁 𝗮𝗹𝗿𝗲𝗮𝗱𝘆 𝗹𝗶𝘃𝗲𝘀 External Collection, open-format support with Iceberg, Vortex, Lance, and Parquet, Snapshot, Spark Connector, Flexible Schema, and TEXT LOB all move Milvus closer to a world where lake-resident data can become searchable, semantic data without heavy migration. 𝟮. 𝗔 𝗿𝗶𝗰𝗵𝗲𝗿 𝘀𝗲𝗺𝗮𝗻𝘁𝗶𝗰 𝗾𝘂𝗲𝗿𝘆 𝗲𝗻𝗴𝗶𝗻𝗲 Order By, Aggregation, StructList, FAISS family indexes, MinHash, Null Vector, Entity TTL, RLS, and related features extend Milvus beyond nearest-neighbor search into a more complete execution engine for semantic search workloads. 𝟯. 𝗔 𝗳𝗼𝘂𝗻𝗱𝗮𝘁𝗶𝗼𝗻 𝗳𝗼𝗿 𝗟𝗮𝗸𝗲𝗯𝗮𝘀𝗲 𝘄𝗼𝗿𝗸𝗳𝗹𝗼𝘄𝘀 𝗼𝗻 𝗭𝗶𝗹𝗹𝗶𝘇 𝗖𝗹𝗼𝘂𝗱 (𝗺𝗮𝗻𝗮𝗴𝗲𝗱 𝗠𝗶𝗹𝘃𝘂𝘀) With streaming, import, external tables, online queries, on-demand queries, and production jobs such as deduplication, anomaly detection, and clustering, Zilliz Cloud Lakebase connects serving and discovery on a shared data layer. This is the bigger shift behind Milvus 3.0: Vector search is no longer just an online serving path. It is becoming part of the data infrastructure loop that helps AI teams serve, discover, improve, and serve again. Milvus 3.0 and Lakebase are coming soon. Stay tuned — we can’t wait to share more with the Milvus community.
·
2
2
New to X?
Sign up now to get your own personalized timeline!
Sign up with Apple
By signing up, you agree to the Terms of Service and Privacy Policy, including Cookie Use.
Relevant people
-  Milvus @milvusio Follow Click to Follow milvusio The most widely adopted open source vector database for #AI#OpenSource#VectorSearch!Image 4: 💬Discord: https://milvus.io/discord!Image 5: 🔗Find us: https://linktr.ee/Milvusio
Trending now
What’s happening
Entertainment · Trending
#GreysAnatomy
Only on X · Trending
#911onABC
Trending in United States
Celtics
Trending in United States
The Devil Wears Prada 2
|
|
|
|
|
More
© 2026 X Corp.
问问这篇内容
回答仅基于本篇材料Skill 包
领域模板，一键产出结构化笔记投融资雷达包
把一条融资 / 创投新闻整理成投资人视角的雷达卡：交易要点、判断、竞争格局、风险、尽调清单。
- · 交易要点（公司 / 轮次 / 金额 / 投资人 / 估值，材料未明示则写 “未披露”）
- · 投资 thesis（这家公司为什么值得关注）
- · 竞争格局与替代方案