BU BÖLÜM HAKKINDA
AI workloads are changing how organizations need to think about data.
In this episode of AI Cloud Essentials, Alon Horev, co-founder and CEO of VAST Data, joins the conversation to discuss how VAST Data and CoreWeave are rethinking data infrastructure for the AI era.
As AI workloads grow in scale, organizations need more than raw compute. They need fast access to massive datasets, real-time data streaming, efficient storage, strong multi-tenancy, and infrastructure designed specifically for modern AI training and inference.
Alon explains how VAST Data is building an AI operating system designed to support the data infrastructure behind AI applications, and how the company’s partnership with CoreWeave has evolved around a shared goal: building highly scalable, efficient AI infrastructure.
The conversation explores why traditional infrastructure architectures are being challenged by workloads that can span hundreds of thousands of GPUs and access hundreds of petabytes of data at once.
In this episode, you’ll learn:
• Why AI workloads require a new approach to data infrastructure
• How CoreWeave and VAST Data are co-engineering infrastructure for AI
• Why efficiency matters as AI compute and power consumption continue to scale
• How modern AI workloads are changing storage, networking, and data access requirements
• Why multi-tenancy and flexible consumption models matter for AI clouds
• How organizations can improve infrastructure utilization and reduce AI costs
• Why real-time data streaming is becoming increasingly important for agentic AI
• How data requirements differ across training, inference, fine-tuning, and reinforcement learning
• Where AI adoption is accelerating across industries like healthcare, autonomous driving, farming, and animation
Alon also explains the scale of today’s infrastructure challenge, including AI workloads running across as many as 200,000 GPUs while accessing hundreds of petabytes of information.
As AI moves from model training toward inference, autonomous agents, event streaming, vector search, and increasingly complex applications, the underlying data architecture has to evolve with it.
For AI infrastructure leaders, developers, data architects, and organizations scaling AI workloads, this conversation offers a practical look at how the data layer is being rebuilt for the next generation of AI.