NVIDIA L4 24GB - an energy-efficient accelerator for AI inference and video in the data center
The L4 is built on the Ada Lovelace architecture (AD104 graphics processor) and carries 7,680 CUDA cores, 240 fourth-generation Tensor cores and 60 third-generation RT cores. The FP8 Transformer Engine delivers up to 485 TFLOPS FP8/INT8 and 242 TFLOPS FP16/BF16 - that is enough for generative models, recommender systems, NLP and computer vision without excessive power consumption.
24 GB of GDDR6 memory with ECC enabled by default, a 192-bit bus and 300 GB/s of bandwidth keep larger batches and longer contexts in memory. The 2x NVENC, 4x NVDEC and 4x JPEG video engines with hardware AV1 cover mass transcoding, AI video effects and video stream analytics.
- Ada Lovelace: 7,680 CUDA, 240 Tensor (4th gen), 60 RT (3rd gen)
- 24 GB GDDR6 with ECC, 192-bit bus, 300 GB/s
- 72 W, single-slot low-profile (HHHL), passive cooling, no auxiliary power
- PCIe Gen4 x16 (compatible with Gen4 x8 / Gen3 x16)
- NVIDIA vGPU 15.2+, SR-IOV (up to 256 VF), Secure Boot, NEBS-3
The single-slot 72 W format powered from the PCIe slot allows the L4 to be installed in dense edge servers and platforms with no free space for display outputs - the card is headless, without display connectors, and is aimed at vGPU and graphics virtualization. ETE.UA is an official PNY/NVIDIA supplier in Ukraine. Check availability and B2B terms: +380 93,594 00 77.
Frequently Asked Questions
What power supply does this card need?
Auxiliary power connector: 1x16-pin. With two of these cards in one system, size the supply with headroom and check the circuit rating.
What model size fits into 24 GB of memory?
Roughly a 13B model at 4-bit, or a 7B model with a long context. The exact figure depends on context length: longer conversations consume more cache. Memory: 24 GB GDDR6, 384 bit bus.
Which GPU does this card use?
NVIDIA L4. The GPU generation decides which compute formats are accelerated in hardware, and that affects inference speed more than raw memory capacity does.
What warranty is provided and what is in the box?
Warranty: 36 months. Ships in the manufacturer's original packaging. In the box: 1x low-profile bracket.
Not sure this card fits your system?
We will check compatibility with your platform, calculate the power draw for your configuration and tell you whether the memory is enough for your models.
The model you need is out of stock? ETE is listed among official PNY partners, so we bring professional NVIDIA cards to order with the manufacturer warranty.
Message us on Telegram or leave a request.
