AI Compute Shifts to Post-Training
This shift signals a fundamental change in AI resource allocation, moving from raw model creation to optimization and deployment. For developers and businesses in ASEAN, this means more accessible, refined AI models are on the horizon, potentially reducing costs for implementing advanced AI solutions compared to training new models from scratch.
Frontier labs allocated 67% of compute to pre-training in Q1 2024.
By Q4 2026, this will drop to just 7%.Post-training and reinforcement learning will grow from 4% to 55%, while inference will account for 38%.
AI 算力战争已经换了战场。 2024年Q1,前沿实验室67%的算力用于预训练;到2026年Q4,预计只剩7%。 与此同时,后训练与RL从4%升至55%,推理占38%。约90%的算力,正在被后训练和推理消耗。 真正持续吞噬算力的,可能不再是训练下一个大模型,而是让现有模型变得更聪明,并服务更多用户。📊 2 ⚡
This report is categorized under AITOP's "Industry" desk, focusing on China's AI industry and domestic foundation model ecosystem. Within our cross-regional analytical framework, it directly maps to AI developer toolchains, coding agents, and engineering workflows. Key entities and technical topics under continuous tracking include frontier labs, compute allocation, pretraining, post-training.
Source attribution: curated from AI Will (@FinanceYF5).