Solidigm Joins Bloomberg Intelligence Podcast: How Storage Disrupts the AI Value and Economics Stack
Co-CEO Xin Guo talks NAND supply dynamics, the rise of KV cache, and why storage is a day-zero design decision for AI infrastructure
Solidigm, a pioneer in enterprise data storage, recently joined Bloomberg Intelligence for a discussion on its Tech Disruptors podcast with semiconductor analyst Jake Silverman to unpack the forces transforming the NAND industry.
From AI-driven demand and severe supply constraints to the emergence of key value (KV) cache as a mission-critical workload, they unpacked Solidigm's expanding role in the AI hardware stack and how the company is driving infrastructure breakthroughs with industry leaders.
During the discussion, Guo explained that the AI-driven demand on storage and memory has been bigger than most anticipated and that the industry’s supply constraints go beyond underinvesting during the last downturn.
It’s a fundamental shift from a human- or interaction-driven infrastructure to a machine-driven infrastructure. And the magnitude of the demand multiplies quickly.Xin Guo, Co-CEO, Solidigm
Guo described KV cache as the model's working memory, or the tokenized history and context of everything a model has ingested and produced. Essential for multi-turn inference, as workloads move from single chat sessions to agentic AI, that footprint explodes.
Minutes prior to recording the podcast, Guo also presented at the Bloomberg Intelligence Generative AI Forum alongside Akamai Technologies' Robert Blumofe and DDN's Alex Bouzari. When panelists covered what happens when AI puts increased pressure on the data layer, Guo spoke to how inference is shattering traditional storage hierarchy.
"Storage is no longer an afterthought in your AI infrastructure. It needs to be a day-zero design consideration, because storage has become mission-critical for your entire infrastructure efficiency, performance, and output, which turns into AI economics and your return," Guo explained.
The full episode is published on the Bloomberg Tech Disruptors podcast series, and the Bloomberg Generative AI Forum stream is now available.