NVIDIA Nemotron Bolt 3.5 74B releases Oct. 15

A 74B-parameter, 9B-active MoE model compressed to 24GB via 2-bit quantization and quantization-aware training, aiming to balance speed and intelligence. A Nemotron 3.5 Bolt E4B 4GB version releases Oct. 23.

• Trump announces $2.4B super intelligence initiative2026-10-14
• Microsoft has a new machine learning tool to improve weather forecasts2026-10-14
• Jensen Huang 'optimistic' on AI risks, sovereignty and jobs2026-10-14
• Analyst warns AI's biggest risk is debt-fueled growth2026-10-14
• TechCrunch Disrupt 2026 Day 2 Focuses on AI2026-10-14