VAST Data AI OS 5.5 integrates in-place analytics and vector search

The release shifts the data paradigm from moving data to compute toward executing compute where data resides. For streaming video and high-scale media companies, this reduces the latency and cost overhead of complex data pipelines previously necessary for real-time analytics. By merging vector retrieval with standard SQL operations, VAST enables more sophisticated AI-assisted metadata investigations without the fragility of multi-system integration. Watch for the addition of GROUP BY and JOIN functions in upcoming iterations to see if VAST can fully displace traditional data warehouse layers in media tech stacks.

The release of AI OS 5.5 follows VAST Data’s significant expansion in the enterprise infrastructure market. Per Frontier Enterprise in April 2026, the company closed a Series F funding round that valued the business at $30 billion, more than tripling its previous $9.1 billion valuation from late 2023. This capital influx, led by Drive Capital and involving NVIDIA, underscores the market's appetite for infrastructure that collapses the traditional separation between storage and AI compute.

Historically, VAST has focused on its Disaggregated Shared-Everything (DASE) architecture to differentiate itself from incumbents like Dell PowerScale and Pure Storage. According to company reports from December 2024, VAST reached $200 million in Annual Recurring Revenue (ARR) and achieved free cash flow positive status. The new analytical features in version 5.5 are designed to compete directly with specialized data lakehouses. Per VAST's internal benchmarks in September 2025, their database architecture can deliver up to 20x faster analytics compared to standard Apache Iceberg stacks.

In the broader ecosystem, VAST has deepened its ties with AI hardware providers. Per digitalocean.com in September 2025, VAST infrastructure is increasingly being paired with NVIDIA Blackwell systems to support massive GPU clusters. This integration is critical for media-heavy workloads, where real-time video analysis and embedding generation require high-throughput object storage that can simultaneously handle complex metadata queries. Competitors such as WEKA and DDN have also raised significant capital to address this space, with WEKA securing $140 million in 2024 to scale its own AI-optimized file system.