← Back to the AI Glossary 📖 Infrastructure & Ops

GPU (Graphics Processing Unit)

Specialised computer hardware, originally built for video game graphics, that turned out to be extremely well suited to the parallel maths AI training and inference require.

The global shortage and high cost of top-tier GPUs (like NVIDIA's) is a major reason AI training is so expensive and why access to compute has become a competitive advantage.
One of 60 free AI glossary terms
Plain-language definitions for the AI jargon you'll actually run into — no email needed, ever, for this section.
Browse the Full Glossary →
More in Infrastructure & Ops
Model Serving
The infrastructure and software responsible for actually running a trained model and handl...
Tokens Per Second (TPS)
A measure of how fast an AI model generates output, expressed as the number of tokens it p...
Throughput
The total volume of requests or tokens an AI system can process in a given amount of time,...
On-Device AI
Running an AI model directly on a phone, laptop, or other local device, rather than sendin...