NVIDIA 900-2G193-0000-000 L4 24GB GDDR6 192bit Graphics card
The NVIDIA L4 (part number 900-2G193-0000-000) is a low-profile, single-slot datacenter GPU built on the Ada Lovelace architecture, positioned for energy-efficient AI inference, video processing, and virtual desktop infrastructure (VDI). With 24GB of GDDR6 memory on a 192-bit interface, it targets buyers deploying GPU density in cloud, edge, and industrial platforms where power and space budgets are constrained. Practical note: it uses passive cooling, so it requires a chassis with directed airflow rather than a workstation without server-grade airflow.
Key specifications (table)
| Parameter | Specification |
|---|---|
| Part Number | 900-2G193-0000-000 |
| GPU Family | NVIDIA L4 (Ada Lovelace architecture) |
| Memory Capacity | 24GB GDDR6 |
| Memory Interface | 192-bit |
| Form Factor | Compact, low-profile design for small-form-factor systems and edge devices |
| Cooling | Passive (no fans; requires chassis airflow — typical for server deployment) |
| Power Profile | Energy-efficient, low power usage (power-draw suited to AI acceleration workloads) |
| Target Use Cases | AI inference, video processing, VDI, cloud, edge, and industrial platforms |
| Lifecycle | Designed for long lifecycle support in enterprise environments |
Good to know
- Because cooling is passive, verify your server or edge chassis provides sufficient directed airflow across the card before deployment; it is intended for rack and industrial enclosures, not desktop towers without airflow channels.
- Within NVIDIA’s datacenter lineup, the L4 sits below the larger L40S/Hopper-class accelerators as a cost- and power-optimized inference card, offering a balance of memory capacity and efficiency rather than maximum training throughput.
- Confirm host platform PCIe slot clearance (low-profile/small-form-factor) and server thermal validation with your integrator before ordering, and request current stock availability and lead times from our sales team.
For tray and reel pricing, current stock by part number and date code, and volume discounts on bulk orders, request a bulk quotation.
The NVIDIA L4 (part number 900-2G193-0000-000) is a low-profile, single-slot datacenter GPU built on the Ada Lovelace architecture, positioned for energy-efficient AI inference, video processing, and virtual desktop infrastructure (VDI). With 24GB of GDDR6 memory on a 192-bit interface, it targets buyers deploying GPU density in cloud, edge, and industrial platforms where power and space budgets are constrained. Practical note: it uses passive cooling, so it requires a chassis with directed airflow rather than a workstation without server-grade airflow.
Key specifications (table)
| Parameter | Specification |
|---|---|
| Part Number | 900-2G193-0000-000 |
| GPU Family | NVIDIA L4 (Ada Lovelace architecture) |
| Memory Capacity | 24GB GDDR6 |
| Memory Interface | 192-bit |
| Form Factor | Compact, low-profile design for small-form-factor systems and edge devices |
| Cooling | Passive (no fans; requires chassis airflow — typical for server deployment) |
| Power Profile | Energy-efficient, low power usage (power-draw suited to AI acceleration workloads) |
| Target Use Cases | AI inference, video processing, VDI, cloud, edge, and industrial platforms |
| Lifecycle | Designed for long lifecycle support in enterprise environments |
Good to know
- Because cooling is passive, verify your server or edge chassis provides sufficient directed airflow across the card before deployment; it is intended for rack and industrial enclosures, not desktop towers without airflow channels.
- Within NVIDIA’s datacenter lineup, the L4 sits below the larger L40S/Hopper-class accelerators as a cost- and power-optimized inference card, offering a balance of memory capacity and efficiency rather than maximum training throughput.
- Confirm host platform PCIe slot clearance (low-profile/small-form-factor) and server thermal validation with your integrator before ordering, and request current stock availability and lead times from our sales team.
For tray and reel pricing, current stock by part number and date code, and volume discounts on bulk orders, request a bulk quotation.
