ScaleFlux CEO to Deliver FMS 2026 Keynote with NVIDIA on Memory Solutions to Scale AI Data Pipeline

ScaleFlux experts will lead seven presentations, including a keynote with NVIDIA, covering AI data infrastructure, CXL memory, SSD architecture, storage security, and the evolving role of flash memory in AI Systems.

MILPITAS, Calif., July 16, 2026 /PRNewswire/ -- ScaleFlux today announced that CEO & Co-founder Hao Zhong will deliver a keynote at Future of Memory and Storage (FMS) 2026 alongside Jason Hardy, VP of Storage Technology at NVIDIA, highlighting the industry's evolving approach to AI infrastructure and next-generation memory and storage technologies.

The keynote headlines a strong presence for ScaleFlux at FMS 2026, where six technical sessions featuring ScaleFlux experts have been selected for the conference program. Together, they will cover AI infrastructure, flash storage, CXL memory, SSD architecture, and storage security, showcasing the company's leadership across the technologies shaping the future of data infrastructure.

Across these sessions, ScaleFlux's experts will tackle some of the most significant infrastructure challenges facing large-scale AI deployments, including memory scalability, storage efficiency, reliability, security, and the rising cost of AI infrastructure.

"AI is forcing the industry to rethink the traditional boundaries between memory and storage," said Hao Zhong, CEO & Co-founder of ScaleFlux. "At FMS, we'll share practical architectures and real-world engineering approaches that make AI infrastructure more scalable, more efficient, and ultimately more economical, from memory expansion and tiers of SSD solutions."

With seven appearances across three days, ScaleFlux will be one of the conference's most visible technical contributors.

ScaleFlux Sessions at FMS 2026

Keynote: Memory Solutions to Scale AI Data Pipeline
Wednesday, August 5, 1:20pm PT
Hao Zhong, CEO & Co-founder of ScaleFlux, and Jason Hardy, VP of Storage Technology at NVIDIA
As AI models continue to grow, memory capacity has become one of the industry's biggest bottlenecks. In this keynote, Hao Zhong and Jason Hardy will examine how flash is emerging as a new memory tier for AI inference, enabling larger models, improved GPU utilization, lower infrastructure costs, and more scalable AI deployments.

When AI Meets NAND: Rethinking SSD Controller Design for Heterogeneous Inference State 
Tuesday, August 4, 10:20am PT
Yang Liu, VP of Hardware Engineering
As AI inference workloads evolve, SSD controllers must manage increasingly diverse data types and access patterns. This session explores new controller architectures designed to optimize performance and efficiency for heterogeneous AI inference states.

Reinventing CXL Memory for RAS and TCO 
Tuesday, August 4, 10:35am PT
Prof. Tong Zhang, Chief Scientist
This presentation examines how next-generation CXL memory solutions can improve reliability, availability, and serviceability (RAS) while reducing total cost of ownership for AI and enterprise deployments.

Managing Write Amplification Without Data Placement 
Tuesday, August 4, 10:35am PT
Keith McKay, Sr. Director, Solutions Architecture and Technical Partnerships
Write amplification remains one of the most significant factors impacting SSD endurance and efficiency. This session discusses innovative approaches to reducing write amplification without relying on traditional data placement techniques.

Chat with the Experts: KV Storage, Computational Storage, Storage Security and Form Factors/Low Power
Wednesday, August 5, 7:15pm PT
Pichai Balaji (VP of Products & Marketing), Keith McKay (Sr. Director, Solutions Architecture), Ben Reese (Senior Solutions Architect), Junichi Hyodo (Sr. Principal SSD Solutions Architect)

Breaking the HBM Cost Wall for AI Inference 
Thursday, August 6, 8:35am PT
Prof. Tong Zhang, Chief Scientist at ScaleFlux
High-bandwidth memory (HBM) is critical for AI performance but presents significant cost challenges. This session explores alternative architectures and technologies that can help scale AI inference economically.

When Flash Becomes Memory: Rethinking the Five-Minute Rule for AI Infrastructure
Thursday, August 6, 10:05am PT
Prof. Tong Zhang, Chief Scientist at ScaleFlux, and Vikram Sharma Mailthody, Sr Research Scientist, Nvidia
Co-presented with NVIDIA, this session revisits the classic Five-Minute Rule in the context of modern AI infrastructure and examines how advances in flash technology are reshaping memory hierarchy decisions.

Can We Trust the Black Box? What It Takes to Secure an SSD in the AI Era 
Thursday, August 6, 1:45pm PT
Ben Reese, Senior Solutions Architect
As storage devices become increasingly intelligent, ensuring trust, transparency, and security becomes more critical. This presentation examines the architectural and operational requirements for securing SSDs in AI-driven environments.

Meet ScaleFlux at FMS 2026 (booth #519)

Attendees are encouraged to meet with the ScaleFlux team throughout FMS 2026 to discuss AI infrastructure challenges, explore emerging storage and memory technologies, and learn how ScaleFlux solutions help organizations build more efficient and scalable AI systems.

For more information about ScaleFlux at FMS 2026, visit: https://scaleflux.com/on-the-road/

About ScaleFlux

ScaleFlux is a semiconductor silicon solutions company delivering advanced storage and memory solutions designed to transform data infrastructure in a highly scalable and sustainable manner. Founded in 2014 and headquartered in Milpitas, California, ScaleFlux develops innovative storage and memory controller technologies with breakthrough performance, efficiency, security & scalability that power AI, cloud computing, data center, enterprise, and edge applications.

For more information, visit www.scaleflux.com.

SOURCE ScaleFlux, Inc.