NVIDIA Set to Launch Next-Gen R100 AI GPUs by Late 2025, Featuring Advanced Power Efficiency

The upcoming R100 GPUs, named after astronomer Vera Rubin, will utilize TSMC's 3nm technology and focus on reducing power consumption in data centers.

• NVIDIA's R100 GPUs, entering mass production in Q4 2025, aim to enhance AI performance while managing high energy demands.
• The new GPUs will feature a 4x reticle design and incorporate eight HBM4 memory stacks, indicating significant advancements in hardware architecture.
• Analyst Ming-Chi Kuo highlights the shift towards power efficiency, a critical factor for the sustainability of data center operations.
• NVIDIA plans to integrate these GPUs into system and rack solutions by the first half of 2026, marking a significant upgrade over current models.
• The development reflects NVIDIA's strategic adoption of proven process technologies to balance performance and power efficiency.

3 Articles

Tom's Hardware Wccftech TweakTown