VAST Data

Texas Advanced Computing Center (TACC) is advancing HPC and AI infrastructure with its new Horizon system, real-time data assimilation, and high-performance storage.

Warehouses and lakehouses were built for analytical workloads AI doesn't have. VAST DataBase unifies transactional and analytical data to deliver 210x faster updates than Iceberg without file compaction taxes.

Moving data between fragmented storage systems and separate Kubernetes compute clusters is a tax on AI pipeline velocity. With VAST Native Compute in 5.5, containerized functions run directly alongside files, tables, event streams, and vector embeddings in the VAST DataEngine. By bringing managed compute into the data layer, organizations eliminate cluster management overhead and execute real-time processing where data actually lives. Exploring the operational shift behind VAST Native Compute 👇

Scaling Agentic AI isn't just a compute challenge—it's quickly becoming a memory and retrieval bottleneck. As multi-turn agents reason across longer context windows, storing and retrieving Key-Value (KV) cache data efficiently is critical to avoiding heavy recomputation costs and keeping GPUs saturated. Premiering Tuesday, August 25th during the Supermicro Open Storage Summit, I’ll be joining experts from #Solidigm and #Supermicro for our session: Breaking the Context Wall: Storage for Scalable Agentic AI. We unpack: 🚀 How to balance query reprocessing against token retrieval 🚀 The emergence of a dedicated "context memory" tier between local GPU SSDs and network storage 🚀 Key architecture trade-offs when persisting tokens across large-scale storage arrays If you’re building or scaling long-context AI pipelines, tune in on August 18th! Join the session here: https://lnkd.in/dMBgSN65 #StorageSummit2026 #OpenStorageSummit #SupermicroStorage2026

The number of Shared Everything listeners has grown so much over the last year and Nicole Hemsoth Prickett wanted to make sure some of the first episodes don’t get lost in the mix. Today she's featuring an encore episode with the Executive Director of the Texas Advanced Computing Center (TACC) Dan Stanzione and VAST Data's Don Schulte. They discuss the evolution of HPC, the rise of AI infrastructure, and what's next for one of the world's leading supercomputing centers. Enjoy and see you next week for a new episode!

ANZ is one of the fastest-moving AI infrastructure markets in the world right now, and VAST Data is at the center of it. Growth like this needs people that can match it, making sure ANZ customers get the depth of expertise this market now demands. That's why I'm glad to introduce @Paul Armstrong, our new regional sales director for ANZ. Paul joins us with more than 20 years in enterprise infrastructure, including a run at @Lenovo's Infrastructure Solutions Group where he helped drive Lenovo to the #1 x86 server revenue position in Australia and New Zealand. He has deep relationships across the NVIDIA partner ecosystem and a track record of closing complex, large-scale infrastructure deals. Connect with Paul, connect with our team, and reach out if you are scaling AI infrastructure in the region. VAST is open for business in ANZ. 🇦🇺 Paul Armstrong Ashish BhojaniSunil ChavanRene TyhouseJack CrookesPaul Bruton Oxana Plis

Everyone loves the gooey center of the cake. For the world's largest AI clouds, that center is VAST Data. On stage at RAISE Summit, VAST Data Co-Founder Jeff Denworth explains how VAST serves as the connective tissue behind modern AI infrastructure.

Collecting physical AI data is straightforward. Reasoning over it in real time across thousands of autonomous machines is where legacy architectures collapse. NCS Group CEO Sam Liew reveals how VAST and NCS are solving the real-time data foundation challenge behind Physical AI. By combining VAST’s unified architecture with NVIDIA GPUDirect Storage and NVIDIA Cosmos Reason, NCS has built the data fabric powering the next generation of humanoid systems, drones, and autonomous mobile fleets.

Analytical execution belongs at the data layer, not above it. Data should not have to move to be analyzed. When querying massive datasets, sending raw data over the network to external compute nodes is a tax on both time and infrastructure. The VAST Native Query Engine brings over 50 in-place SQL functions directly to the storage tier—evaluating regressions, percentiles, and conditional aggregations natively on flash. By returning only finalized result sets, VAST eliminates network I/O bottlenecks and lowers compute overhead across your entire data stack. Read Colleen Quinn's blog to see how in-place execution works 👇