GPU
A Graphics Processing Unit (GPU) is a specialized electronic circuit designed to rapidly manipulate and alter memory. Because training neural networks involves massive matrix multiplication, the parallel processing power of GPUs is critical for modern AI workloads.
Frequently Asked Questions
Why are GPUs better than CPUs for AI?▼
CPUs have a few powerful cores optimized for sequential tasks. GPUs have thousands of smaller cores designed to perform simple operations (like matrix multiplication) simultaneously.
What is VRAM?▼
Video RAM (VRAM) is the fast dedicated memory on the GPU. Modern AI models require massive VRAM (e.g. 24GB to 80GB) to store parameters during training and inference.
Quick Facts
- CategoryHardware & Infrastructure
- Key ApplicationModel training compute, high-throughput inference hosting, and heavy image synthesis
Coverage Trend12 Weeks
Related AI Terms
GPU Media Coverage & Intelligence
Choosing the Right NVIDIA Platform for Running Inference on CoreWeave
Explore the ideal NVIDIA GPUs for running inference on CoreWeave-optimize latency, reduce token cost, and match your model to the ideal GPU for real-time performance.
5 Misunderstandings About Enterprise AI Training Infrastructure
Enterprise AI leaders often think more GPUs equals faster time-to-market. The reality is more complex. Read on for five misconceptions about enterprise AI training that inflate TCO and slow delivery.
CoreWeave Trains DeepSeek-V3 Benchmark in Two Minutes
CoreWeave's MLPerf® Training v6.0 results set new records, demonstrating how customers can train frontier AI models faster, scale more efficiently, and get more value from every GPU deployed.
AI ready data is the missing link keeping enterprise AI stuck in pilot mode
Enterprises have poured billions into artificial intelligence infrastructure - GPUs, cloud capacity, model tooling - yet most deployments remain mired in experimentation rather than generating measurable business value. The bottleneck is not compute. It is AI ready data. The gap between owning data
AI coding agents taught robots how to install GPUs and cut zip ties
Nvidia's self-improvement program for robots enlists teams of AI coding agents.
Rackspace signs definitive agreement to deploy 30MW of AMD AI compute
Rackspace and AMD have completed a deal to roll out massive GPU infrastructure globally.
QumulusAI and the shift from GPU scarcity to GPU efficiency
Neocloud provider QumulusAI announced today that it has secured more than $124 million in customer subscriptions for three-year terms with Hyperbolic and another leading artificial intelligence inference platform. These agreements cover deployments totaling 1,280 Nvidia Corp. Blackwell GPUs, deliver
What AI benchmarks miss about real-world performance
Presented by F5 Enterprise AI teams have spent years solving for compute, securing GPU allocations, negotiating cloud capacity, and benchmarking training throug
NVIDIA Confidential Computing to Help Expand Apple's Private Cloud Compute
NVIDIA GPUs with Confidential Computing are now used for confidential inference in Apple's Private Cloud Compute (PCC), as it expands beyond Apple's data center
Accelerate LLM model loading and increase context windows with GPUDirect on Amazon FSx for Lustre and TurboQuant
If you're iterating on deploying large language models (LLMs) on AWS GPU instances, you've probably noticed the larger the model to be loaded into GPU High Band