BiCS10 Has Moved From Roadmap Talk To Samples
As of Friday, 10 July 2026, BiCS10 is no longer just a next-generation NAND concept. Sandisk announced sampling of its BiCS10 1Tb TLC 3D NAND on 2 July 2026, followed by Kioxia announcing sample shipments of its own 10th-generation BiCS FLASH devices on 3 July 2026. The technology was developed through the companies’ long-running NAND partnership, and the first wave is clearly aimed at data-intensive storage rather than ordinary consumer SSDs. The headline figure is easy to spot: 332 layers. But the more useful story is how Kioxia and Sandisk are combining vertical scaling with lateral density improvements, faster NAND signaling, and lower-power operation to prepare for the next class of enterprise drives. (sandisk.com)
Article contains affiliate links, commission may be earned.
The Core Specs: 1Tb TLC, 332 Layers, 4.8 Gb/s Interface
The first announced BiCS10 part is a 1Tb TLC 3D NAND flash memory device. Sandisk says the new generation uses 332 layers and adds Toggle DDR6.0, SCA protocol, and PI-LTT technology to support high-speed, low-power operation. Kioxia lists a 4.8 Gb/s NAND interface speed, which it describes as a 33% improvement over its 8th-generation BiCS FLASH. Kioxia also says bit density has improved by 59% compared with the 8th generation, thanks not only to the higher layer count but also to better lateral density. In other words, BiCS10 is not relying on stacking alone; it is also making more efficient use of die area. (sandisk.com)
See SANDISK Optimus GX PRO 8100 PCIe 5 SSD price
Why Layer Count Is Only Part Of The SSD Story
It is tempting to compare NAND generations by layer count alone, but that can miss the important engineering details. A taller NAND stack can help capacity, yet SSD performance also depends on how quickly data moves between NAND and controller, how efficiently the die is laid out, and how much power is spent during reads and writes. Tom’s Hardware described the current BiCS10 samples as reaching more than 29 Gb/mm² of areal density and framed them as among the densest 3D NAND devices now in sampling. That matters because denser flash can help SSD makers place more capacity into the same physical footprint, but the interface and controller side still has to keep up before those gains turn into real enterprise products. (tomshardware.com)
View Crucial T705 13 GB/s NVMe SSD on partner website
A Better Fit For PCIe 5.0 And PCIe 6.0 Enterprise Drives
The most obvious destination for BiCS10 is the next wave of high-capacity data-center SSDs, especially as PCIe 5.0 drives mature and PCIe 6.0 platforms move closer to broader deployment. AI training, retrieval-augmented generation, checkpointing, analytics, and high-throughput object storage all put pressure on SSD capacity, latency, bandwidth, and energy efficiency. BiCS10’s faster NAND interface could help future SSD controllers feed more data through each channel, while the density gain could support larger drive capacities without simply increasing the number of packages. Still, this is an early sampling stage, so it would be too soon to claim final SSD capacities, endurance ratings, or drive-level performance numbers. Those will depend on controller design, firmware, packaging, thermal limits, and the product targets chosen by each SSD vendor.
Buy WD_BLACK 2TB SN8100 NVMe SSD here
What To Watch Next
For storage buyers and infrastructure teams, BiCS10 is best viewed as a signal of where enterprise flash is heading rather than a product to order today. The key questions now are practical ones: when volume production scales, which controller partners are ready, what capacities appear first, and how SSD makers balance performance against power draw. Kioxia has also described a dual-axis strategy that keeps its 9th-generation solutions in play while advancing 10th-generation BiCS for higher capacity and performance. That makes sense for data centers, where not every workload needs the newest NAND immediately. For AI-heavy environments, though, BiCS10 shows how NAND makers are preparing SSDs to carry more data, move it faster, and do it with tighter efficiency targets. (kioxia.com)
Explore NVIDIA DGX Spark Personal AI Desktop Supercomputer for AI workloads
Comments
No comments yet. Be the first to share your thoughts.