Aiserver Manufacturer
1 / 5
The Tesla L40 is a data center GPU based on the Tesla Ada Lovelace architecture designed to deliver superior visual computing performance.
Powered by the advanced TSMC 4nm process, it features 18,176 CUDA cores, 568 fourth-generation Tensor Cores, and 142 third-generation RT Cores. It delivers up to 91.6 TFLOPS of FP32 computing power.
Equipped with 48GB of GDDR6 memory with 384-bit width and up to 864GB/s bandwidth, easily handling data science, 3D modeling, and high-resolution rendering.
Rendering and 3D Graphics: Accelerates professional 3D visualization, allowing artists to iterate faster and process complex geometries in real-time.
AI Training and Data Science: Provides massive throughput for single-GPU AI training and development, including full FP8 support for reduced training times.
Omniverse Enterprise: Delivers powerful RTX and AI capabilities for digital twins and photorealistic 3D synthetic data generation.
Streaming and Video: Features three encoding and three decoding engines with AV1 support for efficient broadcast and production workflows.
| Form Factor | Full-height, full-length dual-slot PCI-Express Gen4 (267mm x 111mm) |
| Power | 1x 16-pin power connector, 300W Max Consumption |
| Display Out | 4x DisplayPort 1.4a (Supports 4x 5K/60Hz, 2x 8K/60Hz, or 4x 4K/120Hz) |
Zero Trust architecture with silicon-based root of trust and Multi-Factor Authentication.
Autonomous collaboration with OpenManage systems management for simplified server control.
Designed with recycled materials and energy-efficient options to reduce carbon footprint.
The Tesla L40 is designed for 24x7 enterprise data center operations. It features energy-efficient hardware optimized for large-scale deployment. With support for Secure Boot and meeting NEBS Level 3 standards, it provides an extra layer of security for modern infrastructure.
