QNAP Unveils the New QAI-M100 and QAI-U100 Edge AI Accelerators

[GFreeman]

QNAP Systems, Inc., a leading computing, networking, and storage solutions innovator, has released the new edge AI accelerators QAI-M100 and QAI-U100 for QNAP NAS. With plug-and-play installation and flexible configuration, these accelerators provide small-and-medium-sized businesses and individual users with an affordable solution to enhance AI applications such as facial recognition and object recognition on NAS.

"With the rapid growth of edge AI, the demand for real-time data processing and analysis at the edge has increased," said Andy Chung, Product Manager of QNAP, adding "The combination of QNAP NAS and the QAI edge AI accelerators not only provides a high-performance storage platform but also significantly boosts the computing efficiency of AI applications, improving overall cost-effectiveness."



When performing AI computations with the QAI-M100 and QAI-U100, the system operates with lower power consumption and higher efficiency. Users can install two QAI accelerators on a single NAS or application, enhancing the computing performance even further.

• QAI-M100: Uses an M.2 2280 B+M key interface and has a 3 TOPS NPU. It requires a NAS with an M.2 2280 PCIe slot or a PCIe expansion slot to install a QNAP QM2 expansion card. Built-in 1 GB memory and heat dissipation cooling.
• QAI-U100: Uses a USB 3.2 Gen 1 Type-A interface and has a 3 TOPS NPU. It requires a NAS with a USB 3.2 Gen 1 Type-A port. Built-in 1 GB memory, heat dissipation cooling, and LED status light.

According to QNAP Lab testing, adding one QAI edge AI accelerator to a QNAP NAS can increase the AI object recognition speed of QuMagie Smart Album by up to 36% and facial recognition speed by up to 22%.

• Note: All performance improvement data is based on a comparison of the TS-673A NAS with and without using a QAI accelerator.

System Requirements

• NAS Operating System: QTS 5.2.1.2930 build 20241025 (or later), QuTS hero h5.2.1.2929 build 20241025 (or later)
• Compatible Software: QNAP AI Core v3.5.0 (or later), Multimedia Console v2.7.0 (or later), QuMagie v1.5.1 (or later)

View at TechPowerUp Main Site | Source

 

[Chaitanya]

Quite weak when compared to competing M.2 form solutions.

 

[kondamin]

Why not just pick a Soc with a decent npu.
3tops isn’t worth opening the rack

 

[Alan Smithee]

This was exciting for a moment, as QNAP becomes the first mainstream NAS manufacturer to ship an AI accelerator. However, 3 TOPS is just pitiful and lower performance than a Coral. It's higher priced too, at $60 for M.2 (vs. $25) and $70 for USB (vs. $60).

Even by QNAP's press release the performance is underwhelming, with the new "accelerators" improving still-image face recognition speed by a whopping 22% vs. a Ryzen V1500B.

This announcement implies QNAP won't deliver any real AI for another year or two at least, which is a real shame.

 

[TheLostSwede]

It's Rockchip RK1808 based, so it's technically a system on an M.2 module, which could most likely be used to run something else on. It's kind of pricey for what it is though.