ScaleFlux

ScaleFlux builds silicon solutions that move data quickly and efficiently across datacenter and AI infrastructure — eliminating storage and memory bottlenecks.

ScaleFlux co-founder and CEO Hao Zhong says its controller chips target the external memory and storage tiers that increasingly constrain AI data pipelines. The company unveiled controllers for PCIe Gen 6 SSDs and Compute Express Link memory, supporting the second and third stages behind processors with high-bandwidth memory. Zhong says the shift from training toward inference and agentic workloads increases demand for nearby data capacity, bandwidth and performance.

Oct 2026 · SiliconANGLENews article

ScaleFlux uses intelligent SSD and CXL memory controllers to help AI systems treat NAND flash as active memory for workloads such as key-value cache offload. Vice President of Products and Marketing Pichai Balaji says GPUs can be left waiting as data moves through slower memory and storage tiers. The company reports more than 200 flexible data placement streams and 7 to 10 drive writes per day on standard triple-level-cell NAND.

Oct 2026 · SiliconANGLENews article

SiliconANGLE’s theCUBE provides coverage of the AI Data Pipeline Forum on Oct. 13, with John Furrier hosting conversations about infrastructure for dense AI clusters and distributed inference. The forum focuses on system-level challenges including data movement, storage, memory, networking, power and cooling. ScaleFlux develops storage and memory controller technologies, while Dell advances modular rack-scale architectures to meet growing compute density and operational demands.

Oct 2026 · SiliconANGLENews article