M5Stack LLM‑8850 card is an M.2 M-Key 2242 AI acceleration module powered by an Axera AX8850 SoC delivering 24 TOPS ( INT8) of performance, and suitable for host devices such as Raspberry Pi 5, Rockchip RK3588 SBCs, and even x86 PCs like mini PCs with a spare M.2 Key-M socket. The card ships with 8GB RAM, a 32Mbit SPI NOR flash, and also supports H.265/H.264 8Kp30 video encoding and 8Kp60 video decoding, with up to 16 channels for 1080p videos. It is also equipped with an active cooling system to maintain stable temperatures and prevent thermal degradation inside enclosures. M5Stack LLM‑8850 card specifications: SoC – Axera AX8850 CPU – Octa-core Cortex‑A55 processor at 1.7 GHz NPU – 24 TOPS @ INT8 VPU Video Encoder – 8K @ 30 fps H.264/H.265 encoding, supports scaling / cropping Video Decoder – 8K @ 60 fps H.264/H.265 decoding, supports 16 channels 1080p parallel decoding, supports scaling / cropping Memory – 8GB 64‑bit LPDDR4x @ 4266 Mbps Storage – 32Mbit QSPI NOR […]
Axelera Metis M.2 Max Edge AI module doubles LLM and VLM processing speed
Axelera AI’s Metis M.2 Max is an M.2 module based on an upgraded Metis AI processor unit (AIPU) delivering twice the memory bandwidth of the current Metis M.2 module for compute-intensive Edge AI inference applications such as large language models (LLMs) and vision language models (VLMs). The new Metis M.2 Max also offers a slimmer profile, advanced thermal management features, and additional security capabilities. It is equipped with up to 16 GB of memory, and versions for both a standard operating temperature range (-20°C to +70°C) and an extended operating temperature range (-40°C to +85°C) will be offered. These enhancements make Metis M.2 Max ideal for applications in industrial manufacturing, retail, security, healthcare, and public safety. Axelera AI Metis M.2 Max specifications and host requirements: Accelerator – Metis AIPU’ System Memory – 1GB, 4GB, 8GB, or 16GB memory Host Interface – M.2 2280 M-key edge connector with PCIe Gen. 3.0 […]
Realtek RTL8127 10GbE PCIe cards and M.2 modules are starting to show up for $35 and up
Realtek introduced the RTL8127 10GbE to PCIe NIC at Computex 2025 last May as a low-power, low-cost 10Gbps Ethernet solution for PCIe and M.2 networking cards with an estimated price of around $30 to $40. The good news is that 10GbE PCIe and M.2 cards are now available for sale, just not always at the expected price points. Let’s have a look at some of the options available. Auvidea M20E M.2 10GbE module The first RTL8127 device that came to our attention last month was the Auvidea M20E M.2 10GbE module. Auvidea M20E specifications: Chipset- Realtek RTL8127 Ethernet – 10-pin PicoBlade connector for RJ45 or M12 Xcode adapter Host interface – M.2 Key-E (PCIe Gen4 x1) edge connector Compatible with NVIDIA Jetson Orin NX/AGX and others Power Supply Via edge connector Optional PoE via 4-pin JST-GH Power Consumption – < 1.5W Dimensions – 22 x 42mm; suitable for 2242 or […]
Pironman 5 Mini is a cheaper Tower PC case for the Raspberry Pi 5
If you found the Pironman 5 and Pironman 5 Max Tower PC cases for the Raspberry Pi 5 pretty cool, but had a hard time justifying paying much more for the enclosure than for the board itself, SunFounder has an update for you. The company has now introduced the Pironman 5 Mini Tower PC enclosure with a similar, yet smaller design, fewer features, and easier assembly, bringing the price down to $45.99. That means the OLED information display, most RGB LEDs, the oversized tower cooler, the micro HDMI to HDMI adapter board, and more are all gone. Pironman 5 Mini features: Compatible SBC – Raspberry Pi 5 only. The Raspberry Pi 4 won’t work, and I can’t think of another SBC with two micro HDMI ports and the exact same layout as the Pi 5. Storage MicroSD card slot Optional NVMe SSD via M.2 M-Key (2230/2242/2260/2280) PCIe 2.0/3.0 socket Video […]
Qualcomm IPQ5322 embedded boards target WiFi 7 and 10GbE networking applications
There are already some consumer-grade WiFi 7 routers based on Qualcomm IPQ5322 SoC like Xiaomi Router BE6500 Pro and TP-Link Deco BE65, but if you plan on developing your own embedded solution, Compex and Wallys offer separate embedded boards based on the quad-core Cortex-A53 router processor with support for 10GbE, 2.5GbE, and up to tri-band WiFi 7. Compex AP.MI01.2 AP.MI01.2 specifications: SoC – Qualcomm IPQ5322 ‘Miami’ Series quad-core Arm Cortex-A53 processor @ 1.5GHz System Memory – 1GB, DDR4 16-bit (1×16-bit) interface Storage – 128MB NAND Flash, 4MB NOR Flash, 8GB eMMC flash Networking 4x 2.5Gbps Gigabit Ethernet RJ45 ports (note: the photo below shows 4x Gigabit Ethernet (1GbE) ports, but the rest of the documentation mentions 2.5GbE) 1x SFP+ receptacle (20 positions) Wireless On-board 2×2 2.4GHz MU-MIMO 802.11b/g/n/ax/be, max 25dBm per chain 2x U.FL connectors Frequency Range – 2.412~2.484 GHz Modulation – OFDMA: BPSK, QPSK, 16-QAM, 64-QAM, 256-QAM, 1024-QAM WiFi 7 […]
n-Fuse ZBC-M2-1630 – An affordable Zigbee, Thread, and Bluetooth M.2 1630 module
n-Fuse ZBC-M2-1630 is an M.2 1630 module based on Silicon Labs’ MGM260P module with an EFR32MG26 Bluetooth, Thread, and Zigbee multi-protocol Cortex-M33 SoC. While adding Zigbee, Thread, and/or Bluetooth through a USB dongle is common, it’s harder to find M.2 or mPCIe cards. Mixtile used to sell a 2-in-1 Zigbee & Z-Wave mPCIe module based on Silicon Labs MG23 and MG24 chips, but as noted in our Mixtile Edge 2 kit review, it’s now sold to businesses in 50 or 100-piece bundles, or integrated into the company’s gateway. The ZBC-M2-1630 (aka ZBC26-M2-1630) Key A/E M.2 module offers an alternative available to both individuals and companies. n-Fuse ZBC-M2-1630 specifications: Wireless Module – Silicon Labs MGM260P SoC – EFR32MG26 SoC MCU core – Arm Cortex-M33 @ 78 MHz with DSP instructions and floating-point unit AI/ML accelerator – Matrix Vector Processor that is 8x faster and 6x lower power than Cortex-M33 processing, and […]
QNAP QAI-M100 and QAI-U100 Edge AI accelerators for NAS systems improve image recognition performance
NAS used to store your files in your local network for easy access, but new models like QNAP TS-216G, TS-AI642, and TS-133 feature an Arm processor (Rockchip RK3566 or RK3588) with a built-in NPU to accelerate image recognition, so that users can find their photos with a simple text search like “Disneyland Mickey Mouse”. However, if you own an older NAS that lacks an AI accelerator, QNAP has launched the QAI-M100 (M.2 2280) and QAI-U100 (USB) equipped with a Rockchip RK1808 3 TOPS AI accelerator designed to speed up image and facial recognition by up to 36%. QNAP QAI-M100/QAI-U100 specifications: SoC – Rockchip RK1808 CPU – Dual-core Arm Cortex-A35 @ 1.6 GHz NPU – 3 TOPS NPU Memory – 1GB RAM Host interface QAI-M100 – M.2 2280 PCIe Gen2 x1 (M+B key) edge connector QAI-U100 – USB 3.2 Gen1 male port Accessories – Installation guide and heatsink (QAI-M100 only) Dimensions / […]
BentoIO CMX1 – A low-profile Raspberry Pi CM5 carrier board that’s just 11mm thick
The BentoIO CMX1 is a compact, low-profile Raspberry Pi CM5 carrier board, developed by Terioto in Germany. With an 85 x 85 mm footprint and a height of just 11 mm, it’s designed for space-constrained applications like embedded systems, industrial control, and compact edge devices. After the official launch of the Raspberry Pi CM5 module and its carrier board, we’ve seen boards like Waveshare CM5-NANO-B and Orange Pi CM5 “Tablet” Base Board get released. There is no doubt that they are compact, but they are not as slim as this board. While it’s slim, that doesn’t mean it lacks I/Os; it includes a 40-pin Raspberry Pi-compatible GPIO header, two USB 3.0 ports, two internal USB 2.0 headers, a MIPI DSI connector and an optional HDMI port for video output, optional gigabit Ethernet, a PCIe M.2 M-Key for expansion, and it uses a USB-C for power. Other features include an RTC battery connector, […]









