US 
Neousys GT-92GC Intel 14th/13th/12th Gen Core 2U Fanless Rackmount GPU Computer with RTX 2000 Ada
Intel
Linux
Windows
Dual Display Support
PoE Support
USB3
1x M.2
2x mPCIe
2x SIM
4G
5G
Wi-Fi
RAID
NVMe SSD
Ignition Control
Fanless
CE
EN50155
FCCGT-92GC Overview
Multi-camera inspection vehicles, robotic cells and autonomous machines often need GPU inference in a rack, yet a fan-cooled GPU server rarely copes with dust, vibration and a vehicle battery supply. The Neousys GT-92GC is a 2U 19-inch rackmount GPU computer that cools an NVIDIA RTX 2000 Ada card and an Intel 14th/13th/12th Gen Core processor passively, so the whole unit runs fanless from -25 °C to 55 °C with 35 W CPUs.
Compute comes from socketed LGA1700 processors up to the Core i9-14900 on the R680E chipset, with two DDR5-4800 SODIMM slots that take ECC or non-ECC memory, Intel vPro/AMT 16.0 and dTPM 2.0. The RTX 2000 Ada handles deep-learning inference across many camera streams, while HDMI and DisplayPort outputs drive local displays. Storage pairs an M.2 2280 NVMe socket at PCIe Gen4 x4 with two hot-swappable 2.5-inch trays that support RAID 0 or 1, so recorded data can be pulled from the GT-92GC without opening it.
Camera connectivity is the main reason to choose this rackmount GPU computer. Eight Intel I350-AM4 GbE ports on X-coded M12 connectors deliver 802.3at PoE+ to GigE or IP cameras, with a 70 W budget on a 12 V vehicle supply and 100 W on 24 V, plus a separate M12 GbE uplink with Wake-on-LAN. Four USB 3.2 Gen2 ports add USB3 cameras, and two isolated CAN 2.0 ports with SocketCAN, two isolated RS-232/422/485 ports and 4-channel isolated DI and DO connect the vehicle or machine controller. Two full-size mini PCIe sockets with SIM slots take wireless modules.
Power arrives at 8 to 48 V DC on an M12 L-coded connector, with built-in ignition control for vehicle installs; Neousys recommends 18 to 48 V for 65 W processors. The GT-92GC measures 440 x 250 x 88 mm, weighs 7.7 kg and mounts in a 19-inch rack or on a wall with the optional kits. Vibration and shock are tested to EN 50155:2021 and IEC 61373 Category I Class B, with EN 50121 EMC, which suits rolling-stock, agricultural and inspection-vehicle deployments.
GT-92GC Specifications
| Overview | |
|---|---|
| Brand & Model | Neousys GT-92GC |
| Features | Embedded Fanless GPU Industrial M12 Connectors Rugged |
| Performance | |
| Processor Series | Intel Pentium G Series Alder Lake Intel Core 12th Gen Alder Lake Intel Core 13th Gen Raptor Lake Intel Celeron G Series Alder Lake |
| Processor Models | Intel Pentium Gold G7400E: 3.6GHz, 2 Cores, 4 Threads, 6MB Cache, 46W Intel Pentium Gold G7400TE: 3.0GHz, 2 Cores, 4 Threads, 6MB Cache, 35W Intel Core i3-14100: 3.5~4.7GHz, 4 Cores, 8 Threads, 12MB Cache, 60W Intel Core i3-14100T: 2.7~4.4GHz, 4 Cores, 8 Threads, 12MB Cache, 35W Intel Core i5-13400E: 2.4~4.6GHz, 10 Cores, 16 Threads, 20MB Cache, 65W Intel Core i5-14400: 2.5~4.7GHz, 10 Cores, 16 Threads, 20MB Cache, 65W Intel Core i5-14500: 2.6~5.0GHz, 14 Cores, 20 Threads, 24MB Cache, 65W Intel Core i5-14500T: 1.7~4.8GHz, 14 Cores, 20 Threads, 24MB Cache, 35W Intel Core i7-14700: 2.1~5.4GHz, 20 Cores, 28 Threads, 33MB Cache, 65W Intel Core i7-14700T: 1.3~5.2GHz, 20 Cores, 28 Threads, 33MB Cache, 35W Intel Core i9-14900: 2.0~5.8GHz, 24 Cores, 32 Threads, 36MB Cache, 65W Intel Core i9-14900T: 1.1~5.5GHz, 24 Cores, 32 Threads, 36MB Cache, 35W Intel Celeron G6900E: 3.0GHz, 2 Cores, 2 Threads, 4MB Cache, 46W Intel Celeron G6900TE: 2.4GHz, 2 Cores, 2 Threads, 4MB Cache, 35W Intel Core i3-12100E: 3.2~4.2GHz, 4 Cores, 8 Threads, 12MB Cache, 60W Intel Core i3-12100TE: 2.1~4.0GHz, 4 Cores, 8 Threads, 12MB Cache, 35W Intel Core i3-13100E: 3.3~4.4GHz, 4 Cores, 8 Threads, 12MB Cache, 60W Intel Core i3-13100TE: 2.4~4.1GHz, 4 Cores, 8 Threads, 12MB Cache, 35W Intel Core i5-12500E: 2.9~4.5GHz, 6 Cores, 12 Threads, 18MB Cache, 65W Intel Core i5-12500TE: 1.9~4.3GHz, 6 Cores, 12 Threads, 18MB Cache, 35W Intel Core i5-13500E: 2.4~4.6GHz, 14 Cores, 20 Threads, 24MB Cache, 65W Intel Core i5-13500TE: 1.3~4.5GHz, 14 Cores, 20 Threads, 24MB Cache, 35W Intel Core i7-12700E: 2.1~4.8GHz, 12 Cores, 20 Threads, 25MB Cache, 65W Intel Core i7-12700TE: 1.4~4.6GHz, 12 Cores, 20 Threads, 25MB Cache, 35W Intel Core i7-13700E: 1.9~5.1GHz, 16 Cores, 24 Threads, 30MB Cache, 65W Intel Core i7-13700TE: 1.1~4.8GHz, 16 Cores, 24 Threads, 30MB Cache, 35W Intel Core i9-12900E: 2.3~5.0GHz, 16 Cores, 24 Threads, 30MB Cache, 65W Intel Core i9-12900TE: 1.1~4.8GHz, 16 Cores, 24 Threads, 30MB Cache, 35W Intel Core i9-13900E: 1.8~5.2GHz, 24 Cores, 32 Threads, 36MB Cache, 65W Intel Core i9-13900TE: 1.0~5.0GHz, 24 Cores, 32 Threads, 36MB Cache, 35W |
| Performance Cores | 2 Cores 4 Cores 6 Cores 10 Cores 12 Cores 14 Cores 16 Cores 20 Cores 24 Cores |
| Supported GPUs | NVIDIA RTX 2000 Ada Generation: 2816 CUDA Cores, 88 Tensor Cores, 16GB GDDR6 ECC, 70W |
| AI Performance | 192 TOPS |
| Number of GPUs | Single GPU Support |
| GPU Family | NVIDIA RTX Ada |
| GPU Architecture | NVIDIA Ada Lovelace |
| System GPU Power Budget | 70W |
| Memory | 2x DDR5 4800MHz SO-DIMM Slot, 128GB |
| Chipset | Intel R680E |
| PCIe Gen Support | PCIe 4.0 |
| BIOS | UEFI |
| Secure Crypto Processor | TPM 2.0 (Trusted Platform Module) |
| Operating System | Linux Windows 10 Windows 11 |
| Graphics | |
| Integrated GPU | Intel UHD Series |
| Display Standards | HDMI 1.4b |
| Multi Display Support | Dual Display Support |
| Display Outputs | 1x Display Port 1x HDMI |
| Audio | |
| Interface | Line-Out Mic-In |
| I/O | |
| Ethernet | 9x Ethernet Ports |
| 8x GbE LAN [M12] [PoE] 1x GbE LAN [M12] | |
| LAN Controllers | Intel I219LM Intel i350AM4 |
| PoE Standard | PoE+ [IEEE 802.3at] |
| USB | 4x USB Ports |
| 4x USB 3.2 Gen 2 [Type A] | |
| CANbus | 2x CANbus Ports |
| 2x Isolated CANbus [M12] | |
| Digital I/O | Isolated DIO [4-in, 4-out] |
| Remote Management | Intel AMT Intel vPro |
| Expansion | |
| M.2 | 1x M.2 Expansion Slots |
| 1x M.2 M-Key [PCIe x4 Gen4] [2280] | |
| mPCIe | 2x mPCIe Slots |
| 2x mPCIe [PCIe/SIM/USB] | |
| SIM Slots | 2x SIM Slots |
| Wireless Connectivity | Optional 4G Modem Optional 5G Modem Optional WiFi/Bluetooth |
| Storage | |
| Storage | 1x M.2 SSD [M-Key 2280, PCIe Gen 4 x4] 2x 2.5" HDD/SSD [SATA] |
| NVMe | NVMe SSD Support |
| Front Accessible Storage | Front Accessible Storage |
| Hot Swap SSD Support | Hot Swappable SSD Support |
| RAID | RAID 0 (Striping) RAID 1 (Mirroring) |
| Power | |
| Power Input | 8-48V DC [M12] |
| Power Support Modes | AT: PWR on Input ATX: Press Button to Power |
| Ignition Sensing | Supports Ignition Control |
| Environmental Conditions | |
| Thermal Dissipation | Fanless Passively Cooled Chassis |
| Operating Temperature | -25°C to 55°C [-13°F to 131°F] |
| Storage Temperature | -40°C to 85°C [-40°F to 185°F] |
| Humidity | 10-90% (Non-condensing) |
| Shock | IEC 61373 |
| Vibration | IEC 61373 |
| EMC | EN55032 EN55035 EN50121 |
| Certification | CE EN50121-3-2 EN50155 FCC |
| Mechanical Properties | |
| Weight | 7.7 Kg (16.94 Lbs) |
| Chassis Material | Aluminum |
| Dimensions | 440mm x 250mm x 88mm (17.32″ x 9.84″ x 3.46″) |
| Mounting Options | 19" Rack Mount Wall Mount |
Official Neousys USA Distributor
Things Embedded is an official Neousys USA distributor, and the GT-92GC is the Neousys answer for GPU inference in a standard 2U rack. Where the installation is washed down or fully exposed, the SEMIL-2000GC moves to an IP69K rack chassis with an NVIDIA L4 GPU and M12 PoE+. For defense vehicles that must meet MIL-STD-461G and MIL-STD-1275D, the SEMIL-2200GC is the closer fit, and the GT-92RL-H is the railway build of the GT-92 family. Each of these shares the M12 connector approach, so cabling stays sealed from camera to computer.
Across our range, the GT-92GC sits with our rugged GPU computers and the wider set of industrial PCs with GPU, where its fanless RTX 2000 Ada card and wide-temperature rating set it apart. The eight PoE+ M12 ports also place it among our PoE camera computers, and the ignition-controlled 8 to 48 V input suits it to in-vehicle computers on inspection and agricultural vehicles. Rack and wall kits are optional, so the same unit can go into a cabinet or onto a bulkhead.
Consult. Design. Integrate.
Things Embedded is a leading USA supplier of industrial computing and edge AI systems. We combine technical expertise with a collaborative approach to deliver optimal solutions for your application. Whether it’s an off-the-shelf platform or a custom design, we stay at the forefront of the latest technologies to ensure your success at the edge. Our team works closely with you from concept through deployment, providing guidance, engineering support, and integration expertise every step of the way.
Connect with our solutions experts today to review your requirements and find the ideal solution fit for your application.