ORICO Introduces PCIe 5.0 FlashMem FM9800 SSD for Local AI Workloads

As more artificial intelligence workloads move from cloud platforms to personal computers, hardware manufacturers are exploring new ways to handle the growing memory requirements of local AI models. Larger language models, extended context windows, and multi-agent applications can put considerable pressure on system RAM and GPU memory. ORICO is addressing part of this challenge with its upcoming FlashMem FM9800, an AI-focused SSD that combines conventional storage with dedicated cache space for local inference workloads.

The ORICO FlashMem FM9800 uses a PCIe 5.0 interface and an M.2 2280 form factor. The company plans to offer the drive in 1TB and 2TB capacities, with its storage divided into two areas: Storage OS and AI Cache. The first is intended for the operating system, applications, and everyday files, while the second is designed to hold selected AI model weights, context data, and inference cache.

The drive is also designed to work with AI storage middleware and ORICO’s Herdsman local inference engine. Through firmware policies and host-side software, the system coordinates data between GPU memory, system RAM, and the dedicated AI cache. This approach adds another storage tier to compatible AI systems, allowing selected data to remain on the SSD rather than occupying high-speed memory at all times.

However, the FM9800 is not intended to replace system DRAM or GPU VRAM. NAND flash storage has different performance characteristics, and its usefulness for AI inference will depend on how effectively the software manages data movement between memory tiers. Instead, ORICO is targeting users who want to experiment with larger local AI models on systems with limited memory capacity.

Potential applications include local document processing, software development, research, and handling sensitive business information without relying entirely on cloud-based AI services. Compatibility will remain an important consideration, as the FM9800 requires supported hardware, operating systems, inference engines, and AI model versions.

With the FlashMem FM9800, ORICO is extending its storage portfolio into the emerging local AI hardware market. The company continues to develop conventional storage products, including portable SSDs, USB flash drives, and memory cards, while its USB4 and Thunderbolt 5 products target high-performance storage workflows.

The FM9800’s market reception will likely depend on software compatibility, sustained performance, and measurable improvements in supported AI applications. Pricing, sequential read and write speeds, and availability have not yet been specified in the supplied announcement. Nevertheless, ORICO’s new SSD reflects a growing interest in using storage as another component of the local AI data-management pipeline rather than simply a place to keep files and applications.

SpecificationORICO FlashMem FM9800
Product NameFlashMem FM9800
Product TypeAI SSD
InterfacePCIe 5.0
Form FactorM.2 2280
NAND Flash3D TLC NAND
AI Cache TechnologypSLC
Capacity Options1TB / 2TB
1TB AI Cache85GB
1TB Storage OS768GB
2TB AI Cache128GB
2TB Storage OS1,638GB
Storage ArchitectureDual-partition design
AI Cache UsageModel weights, context data, inference data
Software SupportAI Storage Middleware
Inference EngineHerdsman
Memory CoordinationGPU VRAM, System RAM, AI Cache
Cache ManagementIndependent wear-management and read/write scheduling
CoolingGraphene thermal pad
Operating SystemSupported OS versions (not specified)
Sequential Read SpeedNot announced
Sequential Write SpeedNot announced
Endurance (TBW)Not announced
WarrantyNot announced
PricingNot announced
AvailabilityNot announced
Jani Dushman
Jani Dushman

I'm Jani, a dedicated Tech Writer and Reviewer at Xiaomitoday. With a passion for exploring and dissecting the latest in technology, my mission is to bring you insightful and comprehensive reviews that empower your decision-making in the fast-evolving world of gadgets and tech.

We will be happy to hear your thoughts

      Leave a reply

      XiaomiToday
      Logo