The AI Wire

Nunchaku Lite Finally Lets You Run Big Diffusion Models on Consumer GPUs—Without the Usual Headaches

Nunchaku Lite integrates SVDQuant quantization into Diffusers, enabling W4A4 inference on consumer GPUs with half the VRAM and up to 1.8x speedup. It supports pre-quantized checkpoints via standard `from_pretrained()` calls, with NVFP4 for Blackwell and INT4 for older GPUs. While slightly slower than the native engine, it offers model-agnostic support and a straightforward quantization toolkit.

openai ▸ industry· filed 24 Jul 2026

OpenAI's Project Camellia: A Template for AI Infrastructure Negotiations?

OpenAI's Project Camellia in Georgia exemplifies how AI firms are structuring physical expansion with pledges on energy, community investment, and tool access. The initiative highlights a trend of AI companies seeking favorable locations, while local policymakers face balancing economic development with environmental concerns. This voluntary framework may set a benchmark for future negotiations, though questions remain about its sufficiency compared to potential regulations.