AI Distillation grips Silicon Valley and DC as firms chase model efficiency gains in 2026 tech push.

According to Sundar Pichai, Gemma 3 introduces groundbreaking efficiency in AI models, with its largest 27B model operating on just one H100 GPU. This represents a significant leap, requiring at least 10 times less computational power for similar performance compared to other models, potentially reducing operational costs and energy consumption for traders and developers.