VAST Data has taken a significant leap forward in the AI and data infrastructure space by adding vector search capabilities to its database and strengthening its integration with Google Cloud Platform (GCP).
This move underscores VAST’s commitment to enabling real-time AI, scalable data processing, and unified access across multi-cloud and on-premises environments.
At the heart of VAST's architecture is its DASE (Disaggregated Shared Everything) platform, which supports a rich data ecosystem consisting of the Data Catalog, DataSpace, DataStore, and the recently enhanced DataBase and InsightEngine.
Vector Search
Now, with vector search becoming a native part of the DataBase, users can store and retrieve vector embeddings—multi-dimensional numerical representations used in AI and semantic search—alongside traditional metadata and unstructured content.
-
Eliminates the need for external orchestration layers, indexes, or ETL pipelines.
-
Supports seamless hybrid queries combining SQL, vectors, and unstructured data.
-
Vector search is evolving into a key tool for real-time context retrieval, memory, and reasoning in AI agents.
-
Delivers sub-second performance at trillion-vector scale.
-
Uses sorted projections, precomputed materializations, and smart CPU fallback to accelerate search.
All indexes are co-located with data, which allows compute nodes to perform parallel vector comparisons across modalities like text, images, and audio without latency or data sprawl.
AI Pipelines & Cloud-Edge Flexibility
-
VAST has expanded its Google Cloud integration beyond its April 2024 launch.
-
Provides a unified platform for AI model training, inference, RAG, and analytics across cloud, edge, and on-premises.
-
Users can now deploy VAST clusters directly on Google Cloud for improved scalability and reduced complexity.
New features include:
-
InsightEngine for native, data-centric AI pipelines at the storage layer.
-
DataSpace for unified, global data access across environments.
-
Multi-protocol support including NFS, SMB, S3, block, and database access.
-
Allows AI teams to streamline infrastructure and run large-scale data workloads on one high-performance platform.
VAST CEO Renen Hallak described this as a "leap forward," emphasizing how the collaboration with Google Cloud empowers developers and researchers to move faster, innovate smarter, and scale without limits.
In addition to its presence on Google Cloud, VAST's platform is also available on AWS via the Marketplace (with version 5.2 currently live and 5.3 as the latest), and it offers limited availability on Microsoft Azure via Directed Availability virtual appliances.
As VAST continues to push the boundaries of AI infrastructure and data management, it positions itself as a formidable player in the evolving space of intelligent, scalable, and unified AI data platforms.
