New! RTX 3090, RTX 3080, RTX 3070 pre-order

GPU workstation for deep learning

Up to 4x NVIDIA GPUs, including RTX 3090, RTX 3080, RTX 3070, RTX 2080 Ti, Titan RTX, Quadro RTX 5000, RTX 6000, and RTX 8000. Pre-installed with Ubuntu, TensorFlow, PyTorch, Keras, CUDA, and cuDNN.

Customize now

1000+ research groups trust Lambda

TSMC is a customer MEDIATEK is a customer  is a customer  is a customer

Zero Setup, Easy Updates

GPU2020 Stack comes free with your computer. Machine Learning libraries work out-of-the box and can be updated automatically.

Technical specifications

Up to 4x NVIDIA GPUs
Choose from RTX 30 series (Ampere), Titan RTX, Quadro RTX 8000, and RTX 6000
AMD Ryzen or Intel Core i9
Configurable up to 64 cores, 128 threads, and 256 MB cache
Up to 256 GB
Fits up to eight 32 GB DIMMs at 3200 MHz
OS drive
Up to 2 TB
3,200 MB/s seq. read and 2,000 MB/s seq. write
Extra storage
Up to 61 TB
Fits up to eight 7.68 TB SATA SSDs.
Power supply
Up to 1600 watts
of maximum continuous power at voltages between 100 and 240V
Size & weight
Width: 13.1" (332 mm)
Height: 16.3" (415 mm)
Depth: 18.4" (458 mm)
Weight: 38 pounds (17.2 kg)

Frequently asked questions

Not seeing the answer to your question? Contact sales

What if I want a specific GPU, CPU, storage setup, networking config, [...] that's not available on the website?
We offer many options for GPUs, CPUs, storage, and networking that aren't available on the website. If you don't see the option you want, contact our sales team by email at or by phone at +(86) 18677555856. Please note that customized machines are subject to our Custom Orders Return Policy.
What if my workstation develops a hardware issue?
Lambda offers perpetual support from our engineers and up to 3 years of warranty. If you experience issues with your workstation, our engineers are available by email at and by phone at +1 (855) 882-6011. We will immediately replace any faulty components (e.g. GPU, CPU, power supply), which you may return at your convenience. If your workstation is still under warranty, all replacements will be free of charge. Solutions to technical issues may also be found within our community, Deep Talk.
Are discounts available to academics, students, non-profits, public agencies, or start-ups?
Yes, discounts are available. Your discount will depend on industry, project size, and institution. For more details, please contact us by email at or by phone at +(86) 18677555856.
When will my purchase be delivered?
Your workstation typically ships within 1-3 business days of purchase. Highly customized orders may have a longer lead times. For the most accurate delivery estimates, please contact our sales team.
Does GPU2020 offer shipping to countries outside the United States?
Yes, GPU2020 ships globally. Please note if you're outside the United States, your order may be subject to duties and tariffs. For more information, please contact our sales team.
What operating system options are available?
The most popular options are Ubuntu 20.04, Ubuntu 18.04 and Windows 10 Pro. Alternative operating systems are also available. Please contact us for details.
Can I dual boot Windows and Linux?
Yes, we can dual boot your machine. Please contact us for details.
How loud is this workstation?
Approximately 10 decibels above ambient noise. It is designed to be used at your desk. You can use it at home, at your office, in your lab, or access it remotely. For lowest noise levels, install the machine on a stable surface that doesn't vibrate. The machine exhausts heat via airflow from it's rear, right, and top panels. To ensure proper cooling and low fan noise, keep the machine at least 18 inches from any wall and avoid placing objects near the machine that may obstruct this airflow.
What payment methods are accepted?
We accept wire transfers, ACH transfers, and credit cards. Please note that credit card transactions are subject to an additional 3% charge.
Show all FAQs Show fewer FAQs

Explore our research

GPU2020's research papers have been accepted into the top machine learning and graphics conferences, including ICCV, SIGGRAPH Asia, NeurIPS, and ACM Transactions on Graphics (TOG).

HoloGAN: Unsupervised Learning of 3D Representations from Natural Images
We propose a novel generative adversarial network (GAN) for the task of unsupervised learning of 3D representations from natural images. Learn More
RenderNet: A Deep Conv. Network for Differentiable Rendering from 3D Shapes
Traditional computer graphics rendering pipelines are designed for procedurally generating 2D images from 3D shapes with high performance. Learn More
Adversarial Monte Carlo Denoising with Conditioned Aux. Feature Modulation
Denoising Monte Carlo rendering with a very low sample rate remains a major challenge in the photo-realistic rendering research. Learn More