Intel® Core™ Processors (14th Gen) with Discrete GPU Options and High-Bandwidth 10GbE Edge AI Computer
• Intel® Core™ Processor (14th Gen)
• NVIDIA RTX™ Embedded and Intel® Arc™ GPU Options
for Scalable AI Accelerating Solution
• 2 x Marvell 10GbE for High-Bandwidth Uplink and Aggregated Downlink Connection
• 4 x 2.5GbE with PoE+ Option Provides Simplied Connection to Edge Devices
• Front Removeable RTC Battery Design for Easy Maintenance
• Intel® Core™ (14th gen) i9-14900T 24C/32T processor
• Intel® Core™ (14th gen) i7-14700T 20C/28T processor
• Intel® Core™ (14th gen) i5-14500T 14C/20T processor
• Processor-integrated GPU – Intel® UHD Graphics 770
• Optional discrete GPU – NVIDIA RTX™ Embedded 2000 Ada / 3500 Ada / 5000 Ada Intel® Arc™ A370E
R680E
2 x DDR5-5600 SO-DIMM up to 96GB
(ECC support with ECC DRAM)
TPM 2.0, dTPM
Auto reset for unresponsive system
(custom 1-255 sec/min settings in BIOS)
• Without discrete GPU – 2 x HDMI® up to 3840 x 2160 @ 60Hz as specied in HDMI 2.0b
• With discrete GPU – Additional 2 x DP++ up to 7680 x 4320 @ 60Hz
1 x Line-out, 1 x Line-in, 1 x Mic-in
• 2 x 10GbE (Marvell AQC113) via RJ-45 (optional M12 X-code)
• 4 x 2.5GbE (Intel I226) via RJ-45 (optional PoE+ and M12 X-code)
* ETH3 supports Intel® AMT
** All Ethernet ports support PXE boot and Wake-on-LAN
*** PoE supports 25.5W per port with total power budget: 120W
8 x USB 3.2 Gen 2
8 x DI (DC 5-60V), 8 x DO (DC 5V/100mA)
2 × RS-232/422/485
• 1 x M.2 3042/3052 Key B for WWAN with dual SIM support
• 1 x M.2 2230 Key E for Wi-Fi/BT
* Thermal solution is required for M.2 5G module
2 x Mini PCIe (full-size)
• 1 x M.2 2280 Key M for NVMe/SATA SSD
• 2 x M.2 2280 Key M for NVMe SSD
• 2 x Removable tray for 2.5″ SATA SSD
* RAID 0/1/5 support via Intel® VMD.
• DC 9-60V input (nominal power input DC 12V/24V/48V) via 5-pin terminal block
* Nominal power input of DC 24V or above is required for system with discrete GPU and PoE
OCP, OVP, surge protection, reversed polarity protection
Ignition detection, Smart Power Management
Front removeable tray for RTC coin cell battery
• Optional battery kit for up to 10 minutes emergency backup time
(Battery charging Temp. 0°C~45°C and discharging Temp. -10°C~60°C)
* BBU backup time varies depending on actual overall system power consumption
and battery is required to be charged before being used for system power backup
** Operating temperature will be limited within -10°C and 60°C with the battery kit installed.
Patent No.: M447854 (Built-in battery)
• Windows 11 IoT Enterprise LTSC
• Ubuntu 22.04 LTS
• -40°C ~ 70°C (-40°F ~ 158°F) with 0.6 m/s airow
* Operating temperature varies by accessories installed.
-40°C ~ 85°C (-40°F ~ 176°F)
10% RH ~ 90% RH (non-condensing)
• IEC 60068-2-64, random, 2.5G (Z) 0.21G (Y) 0.76 (Z)@5~500Hz, 1hr/axis
• MIL-STD-810H, Method 514.8, Category 4
• EN 61373:2010, Clause 8
• MIL-STD-810H, Method 516.8, 20G (11ms)
• EN 61373:2010, Clause 10
CE, FCC Class A, UKCA, E-Mark, EN 50155, EN 45545-2 (R25), IEC 62368-1
Aluminum Alloy
7 x SMA connector mounting hole
Wall mounting
5.70 kg (12.57 lb)
290 x 250 x 98 mm (11.42 x 9.84 x 3.86 in.) with mounting brackets
IP40
ABOX-5221(P)(G)(-M12X)-zz (P= w/ PoE, G=w/ Discrete GPU, M12X=w/ M12 X-coded connectors,
zz=i9=i9-14900T, zz=i7=i7-14700T, zz=i5=i5-14500T
Intel® Core™ Processors (14th Gen) with Discrete GPU Options and High-Bandwidth 10GbE Edge AI Computer
Made in Taiwan
| Discrete GPU Graphic Cores | Memory | Type/Size | AI TOPS* | System Operating Temp. |
|
NVIDIA RTX™ Embedded 5000 Ada Generation |
9728 CUDA / 304 Tensor |
GDDR6 16GB |
682 |
0°C ~ 60°C |
|
NVIDIA RTX™ Embedded 3500 Ada Generation |
5120 CUDA / 160 Tensor |
GDDR6 12GB |
369 |
0°C ~ 60°C |
|
NVIDIA RTX™ Embedded 2000 Ada Generation |
3072 CUDA / 96 Tensor |
GDDR6 8GB |
232 |
0°C ~ 70°C |
|
Intel® Arc™ A370E |
8 Xᵉ-cores / 128 Execution Units |
GDDR6 4GB |
56 | 0°C ~ 70°C |
*AI TOPS Precision – FP8 for NVIDIA RTX™ Embedded Ada Generation, INT8 for Intel® Arc™.
DDR5-5600 SO-DIMM, 8GB to 48GB, -40°C ~ 85°C (ECC option available)
M.2 2280 Key M NVMe SSD w/ thermal kit, -40°C ~ 85°C
M.2 2280 Key B-M SATA SSD w/ thermal kit, -40°C ~ 85°C
2.5″ SATA SSD, -40°C ~ 85°C
M.2 2230 Key A-E Wi-Fi module, -40°C ~ 85°C
M.2 3042/3052 Key B WWAN module w/ thermal kit, -40°C ~ 85°C
M.2 2242 Key B GNSS module, -40°C ~ 85°C
Mini PCIe GNSS module, -40°C ~ 85°C
Mini PCIe CAN module, -40°C ~ 85°C
BAT-5200 battery kit, charging Temp. 0°C ~ 45°C, discharging Temp. -10°C ~ 60°C
AC/DC 100-240V/24V 280W C14 bare wire power adapter
Nuremberg, Germany — March 9th, 2026 — SINTRONES, a global provider of rugged embedded and Edge AI computing solutions, will present its latest innovations at Embedded World 2026 (Hall 3, Booth 3-458), featuring the ThermoSiphon™ Edge AI Computer alongside a world-leading lightweight compact Edge AI system. Developed under IEC 62443-4-1 certified secure development processes, SINTRONES’ Edge AI solutions are designed for mission-critical applications, including industrial automation, transportation mobility, and military and defence, where real-time perception, local decision-making, long-term operational reliability, and cyber resilience are essential. Demonstrations will include practical AI scenarios such as dangerous material and weapon detection, illustrating how on-device inference enhances situational awareness and response time in safety and security sensitive environments. Patented ThermoSiphon™ Edge AI Computer: Advanced Passive Cooling A central element of the showcase is SINTRONES’ ThermoSiphon™ Edge AI Computer, ABOX-5221, engineered for sustained high-performance Edge AI computing under extreme thermal conditions. Its patented self-circulating, phase-change cooling architecture enables continuous operation without active cooling components such as fans or pumps. By eliminating mechanically driven cooling elements, this design reduces maintenance requirements and supports stable AI inference over extended operating periods. The ThermoSiphon™ platform is intended for long-term stationary deployments across industrial automation and military and defence environments, where thermal stability, system reliability, and continuous uptime are critical. World-Leading Lightweight Edge AI for Autonomous Mobility Also featured is the IBOX-604-G2, positioned as a world-leading lightweight compact Edge AI system, about 0.3kg only, purpose-built for autonomous mobility in transportation systems and mobile defence applications. Powered by NVIDIA® Jetson™ platforms, the IBOX-604-G2 enables AI inference directly on autonomous vehicles, unmanned systems, and mobile robotic platforms, where size, weight, and power efficiency are key constraints. By processing sensor and vision...
Read more