#Phison [#SSD](<https://hardwarebrief.com/hashtag/ssd>) [#ArtificialIntelligence](<https://hardwarebrief.com/hashtag/artificialintelligence>) #PCIe6 #DataStorage



# Phison Unveils AI-Powered Computational SSDs

2026-11-21T15:33:24-05:00

> The new storage architecture integrates an NPU and local memory to process data directly on the drive, reducing transfer bottlenecks.

Phison is reimagining the role of the SSD by moving processing power directly onto the storage device. The company recently detailed a new computational-storage architecture that integrates a Neural Processing Unit (NPU) and dedicated local memory alongside NAND flash. By bringing compute to the data, Phison claims the design can slash data transfer requirements by 10 to 100 times for specific workloads, effectively bypassing the traditional bottlenecks of moving massive datasets to a central CPU or GPU.

## Processing power at the edge of storage

The heart of this architecture is an NPU capable of 36,800 DMIPS, supported by 64GB of attached memory featuring 130GB/s of bandwidth. This setup is designed to handle modern AI and machine learning frameworks natively, with Phison targeting support for PyTorch, vLLM, OpenVINO, and llama.cpp. While the computational SSD remains in development without a formal product name or release window, it signals a shift toward autonomous storage components that can filter and process information before it ever leaves the drive.

## Next-generation speeds and mixed-media efficiency

Alongside its AI-focused research, Phison is pushing the limits of raw throughput with its upcoming PCIe 6.0 controller, dubbed the X3 or PS5303. Leveraging the 64 GT/s lane transfer rate of the Gen6 standard, the X3 aims for sequential speeds up to 28GB/s and 6.8 million IOPS. To support this ecosystem, the company is also developing a suite of infrastructure hardware, including redrivers and retimers.

Phison is further optimizing storage density through a "Mixed Media" design. This approach combines high-speed pSLC flash and high-density QLC flash behind a single 16-channel controller. By integrating fast cache and dense storage within the same drive, the design seeks to balance performance and capacity for data-heavy environments. While these technologies represent the future of the company's roadmap, Phison has yet to provide specific sampling schedules for the new hardware.



## Related news

#MooreThreads #LushanGPU [#ArtificialIntelligence](<https://hardwarebrief.com/hashtag/artificialintelligence>) [#PCGaming](<https://hardwarebrief.com/hashtag/pcgaming>) [#TechNews](<https://hardwarebrief.com/hashtag/technews>)

2026-11-21T17:51:32-05:00

## [Moore Threads Teases Record-Breaking Lushan GPU](<https://hardwarebrief.com/article/moore-threads-teases-record-breaking-lushan-gpu-b3bd8407>)

> The upcoming desktop graphics card promises the largest consumer VRAM capacity ever to power high-end gaming and local AI workloads.

[#Asus](<https://hardwarebrief.com/hashtag/asus>) #ProArtGR1X [#Nvidia](<https://hardwarebrief.com/hashtag/nvidia>) [#ArtificialIntelligence](<https://hardwarebrief.com/hashtag/artificialintelligence>) [#Workstation](<https://hardwarebrief.com/hashtag/workstation>)

2026-11-21T11:23:12-05:00

## [ASUS Opens Inquiries for ProArt GR1X AI Desktop](<https://hardwarebrief.com/article/asus-opens-inquiries-for-proart-gr1x-ai-desktop-43e8fd27>)

> The compact workstation features NVIDIA Blackwell GPUs and Grace CPUs to deliver 1 PFLOP of local AI performance for enterprise users.

[#SpaceX](<https://hardwarebrief.com/hashtag/spacex>) [#Nvidia](<https://hardwarebrief.com/hashtag/nvidia>) #ProjectSaturn [#ArtificialIntelligence](<https://hardwarebrief.com/hashtag/artificialintelligence>) #ApolloGlobalManagement

2026-11-21T10:48:25-05:00

## [SpaceX Secures $40B Lease for Nvidia AI Hardware](<https://hardwarebrief.com/article/spacex-secures-40b-lease-for-nvidia-ai-hardware-754eaed2>)

> The aerospace firm is utilizing a specialized financing trust to scale its terrestrial data centers with billions in high-end processors.

[#Nvidia](<https://hardwarebrief.com/hashtag/nvidia>) [#GeForce](<https://hardwarebrief.com/hashtag/geforce>) [#SSD](<https://hardwarebrief.com/hashtag/ssd>) [#Hardware](<https://hardwarebrief.com/hashtag/hardware>) [#TechNews](<https://hardwarebrief.com/hashtag/technews>)

2026-11-19T12:15:32-05:00

## [NVIDIA Drivers Reportedly Causing Excessive SSD Wear](<https://hardwarebrief.com/article/nvidia-drivers-reportedly-causing-excessive-ssd-wear-83870917>)

> A runaway logging process in recent GeForce drivers is generating up to 30GB of disk writes per hour, potentially damaging user hardware.

#SSSTC [#Kioxia](<https://hardwarebrief.com/hashtag/kioxia>) #QLC [#SSD](<https://hardwarebrief.com/hashtag/ssd>) #IndustrialTech

2026-11-17T16:19:27-05:00

## [SSSTC Unveils PQ2 Series Industrial QLC SSDs](<https://hardwarebrief.com/article/ssstc-unveils-pq2-series-industrial-qlc-ssds-44f29b04>)

> SSSTC launches its first industrial-grade QLC storage solution featuring Kioxia BiCS FLASH 8 and sequential read speeds up to 7,000 MB/s.

#ASRockSkyOC #IntelOverclocking #HardwareModding [#ArtificialIntelligence](<https://hardwarebrief.com/hashtag/artificialintelligence>) [#PCGaming](<https://hardwarebrief.com/hashtag/pcgaming>)

2026-11-17T15:48:46-05:00

## [AI Used to Revive ASRock SkyOC Overclocking Tech](<https://hardwarebrief.com/article/ai-used-to-revive-asrock-skyoc-overclocking-tech-24b24a3a>)

> A hardware enthusiast utilized AI to decompile legacy firmware and identify code that once bypassed Intel's non-K processor restrictions.

#IntelD270 [#CrescentIsland](<https://hardwarebrief.com/hashtag/crescentisland>) #SC26 [#GPU](<https://hardwarebrief.com/hashtag/gpu>) [#ArtificialIntelligence](<https://hardwarebrief.com/hashtag/artificialintelligence>)

2026-11-17T06:30:44-05:00

## [Intel Unveils D270 Crescent Island GPU at SC26](<https://hardwarebrief.com/article/intel-unveils-d270-crescent-island-gpu-at-sc26-cc1d48bb>)

> The new Xe3P-based hardware outperforms NVIDIA's Blackwell Server Edition in financial and AI workloads while using significantly less power.

[#Kioxia](<https://hardwarebrief.com/hashtag/kioxia>) [#Dell](<https://hardwarebrief.com/hashtag/dell>) #DataStorage #AIDataCenter #NVMe

2026-11-12T11:14:08-05:00

## [Kioxia and Dell Pack 157PB Into Single Storage Rack](<https://hardwarebrief.com/article/kioxia-and-dell-pack-157pb-into-single-storage-rack-4465f92f>)

> Kioxia and Dell have developed a high-density storage solution for AI data centers, fitting 157 petabytes into a single 42U rack cabinet.

[#Microsoft](<https://hardwarebrief.com/hashtag/microsoft>) [#AI](<https://hardwarebrief.com/hashtag/ai>) #SatyaNadella [#ArtificialIntelligence](<https://hardwarebrief.com/hashtag/artificialintelligence>) [#TechNews](<https://hardwarebrief.com/hashtag/technews>)

2026-11-11T16:57:28-05:00

## [Microsoft CEO Outlines Vision for 20 Million AI Agents](<https://hardwarebrief.com/article/microsoft-ceo-outlines-vision-for-20-million-ai-agents-db3c58c4>)

> Satya Nadella details plans for massive AI agent deployment at Microsoft while addressing security risks and recent workforce changes.

#AnthropicAI #ClaudeFable [#ArtificialIntelligence](<https://hardwarebrief.com/hashtag/artificialintelligence>) [#TechNews](<https://hardwarebrief.com/hashtag/technews>) #MachineLearning

2026-11-11T15:38:41-05:00

## [Anthropic Launches Claude Fable 5.2 AI Model](<https://hardwarebrief.com/article/anthropic-launches-claude-fable-5-2-ai-model-4c2038e5>)

> Anthropic's new Claude Fable 5.2 outperforms GPT-6 Astra and Gemini 4 Carbon, offering 30 percent faster speeds and record-breaking reasoning.
