In a recent episode of the Datastorage.com podcast, John Kosturos, Founder and CEO at SpringDB and Datastorage.com, sat down with our Co-Founder, President, and Chief Technology Officer, Sunny Smith. As an independent media and research platform, Datastorage.com helps founders, investors, and IT leaders cut through complexity to understand the real economics behind their infrastructure choices. This episode shines a light on the realities of scaling AI infrastructure and what it actually takes to build a true AI company.
Catch the full conversation here and go inside the GPU carrier layer with Sunny Smith.
During the podcast, Sunny shared the foundational story of Massed Compute and detailed his vision for evolving from an IT infrastructure company into an AI company. A big takeaway from the chat was Sunny’s insights on the current neocloud space. He explained that most “neoclouds” act as resellers, layering their own software over third-party hardware rather than completely owning their infrastructure. Massed Compute eliminates the middleman by owning and operating our GPUs. This direct-ownership model is the foundation of Massed Compute LocalMetal, our on-prem infrastructure solution.
Massed Compute LocalMetal is ideal for enterprises wanting cloud-like flexibility with physical hardware security and control.
Another interesting discussion point was how raw compute power is nothing without the right architecture. Sunny highlighted storage as the ultimate Achilles’ heel for enterprises today. To achieve optimal AI workload performance, storage must be ultra-fast and physically adjacent to the compute stack. Sunny discussed Massed Compute’s partnership with PacketFabric, which uniquely pairs on-demand GPU compute with high-performance cloud networking.
John and Sunny also talked about the financial friction points holding enterprises back. Massed Compute’s goal is to provide clients with transparent pricing and zero ingress or egress fees.
In the end, Sunny revealed that most companies leave massive amounts of performance on the table, stretching from BIOS settings all the way up to Layer 7. In high-performance computing, idle time is the enemy, making full-stack optimization the only real path forward.
Optimize Your AI Infrastructure With Massed Compute
Email us at [email protected] or send us a message today to get started with AI infrastructure built for speed, transparency, and raw power.
