SC26: Phison Unveils Three Next-Generation Storage Innovations for the New Compute Era
Advancements address critical data movement, processing and storage challenges for AI and HPC infrastructure
This is a Press Release edited by StorageNewsletter.com on November 17, 2026 at 2:01 pmPhison Electronics, a reference in NAND flash storage solutions, announced technology innovations in PCIe Gen6, Computational Storage and Mixed Media architectures at Supercomputing 2026 (SC26).
As AI and HPC workloads push the limits of traditional infrastructure, Phison is removing bottlenecks by introducing new technologies built to accelerate data movement, bring compute closer to storage and optimize flash performance and capacity.
The latest Phison technology highlights include:
- PCIe Gen6: Breaking the Data Movement Barrier As AI workloads demand faster movement of increasingly massive datasets, PCIe Gen5 is approaching its bandwidth limits. PCIe Gen6 doubles the per-lane transfer rate to 64 GT/s, opening the door to faster storage performance while introducing new signal integrity challenges
- Doubling the potential of enterprise storage: Phison’s next-gen X3 (PS5303) enterprise SSD controller targets up to 28GB /s sequential throughput and 6.8 million IO/s on the PCIe Gen6 x4, NVMe 2.3 interface, approaching twice the throughput of PCIe Gen5 SSDs while enabling greater bandwidth per drive and improved throughput per watt
- Delivering end-to-end PCIe Gen6 innovation: Phison is developing SSD controllers, redrivers and retimers together to maintain signal integrity at higher transfer speeds. The Phison PS7161 is the first PCIe 6.0 redriver on the PCI-SIG Integrators List, continuing the company’s legacy in leading next-gen connectivity
- Computational Storage: Bringing Compute to the Data Traditional storage architectures require massive datasets to move between SSDs, memory and CPUs, creating bandwidth bottlenecks that faster processors alone cannot solve. Phison’s computational storage technology reimagines this approach by moving processing capabilities directly into the drive, where the data already resides
- Reducing data movement by up to 100 times: By enabling drives to scan and process data in parallel, Phison’s Computational Storage architecture returns relevant results instead of transferring entire datasets, potentially reducing data transfers by 10 to 100 times for applicable workloads while freeing host CPU resources
- Integrating AI-ready compute within storage: Built on Phison’s decades of NAND expertise, the architecture combines an integrated NPU that delivers 36,800 DMIPS of compute performance, and 64GB of memory with 130GB /s bandwidth. Support for familiar frameworks including PyTorch, vLLM, OpenVINO and llama.cpp enables developers to build upon the technology without specialized programming tools
- Mixed Media: Eliminating the TradeoQ Between Performance and Capacity Traditional enterprise storage systems rely on separate high-performance cache drives and high-capacity SSDs, forcing organizations to balance speed, endurance and capacity while sacrificing valuable drive bays. Phison Mixed Media technology challenges this conventional architecture by combining both storage tiers within every drive
- Combining performance and capacity in a single SSD: Phison’s Mixed Media architecture integrates high-endurance pSLC and high-capacity QLC flash behind a single controller across 16 channels, eliminating the need for dedicated cache drives and allowing every drive bay to contribute usable storage capacity
- Enabling scalable, resilient storage: By distributing cache across drives, the architecture allows performance and capacity to scale together while reducing the impact of individual drive failures. This means losing one drive out of 24 would result in a 4% cache loss instead of a 25% loss. Built on standard NVMe interfaces, the technology is designed to integrate with existing infrastructure without requiring specialized host software
“The next era of computing won’t be defined by processing power alone, but by how eficiently we can move, process and store the data that fuels it,” said Michael Wu, GM and president, Phison US. “At Phison, we’re challenging the limitations of traditional storage architectures with innovations that double bandwidth, bring intelligence directly to the data and eliminate longstanding tradeofs between performance and capacity. These advancements reflect our vision for a future where storage is no longer a bottleneck, but a catalyst for AI and HPC innovation.”
Where: Supercomputing 2026 (SC26), booth 5413
When: November 17 – 19, 2026













