Glossary / Compute

Compute

The processing power used to train and run AI models, supplied by chips such as GPUs in large data centers. Access to compute is one of the main limits on AI progress.

In the news

• Oracle Introduces Fusion Claw Runtime for Enterprise AI Agents14h ago
• GitHub rebuilds Git storage architecture from scratch to handle AI traffic surge19h ago
• Mathematicians Urge OpenAI Boycott After AI Proofs Flood Field1d ago
• GlobalFoundries and TSMC Ink $2B Deal for US Interposer Production1d ago
• OpenAI Reports Lower Annualized Revenue Than Earlier Projections1d ago
• OpenAI Publishes 722 Manuscripts Containing Math Problem Solutions3d ago
• South Korea Proposes $3.49 Billion AI Funding Plan3d ago
• Sam Altman-Backed Volantis Aims to Build AI Accelerator Using Optical Interposer4d ago
• Reflection debuts Beam, an open-weight AI model to rival Chinese competitors4d ago
• Google Researchers Propose New Method to Prevent AI Agent Test Memorization5d ago
• Harvard Professor Builds BootLoops Harness for AI Scientific Calculations6d ago
• Ai2 Releases Olmo-Core 3 Framework for Efficient Mixture-of-Experts AI TrainingOct 2