Verified65% confidenceFactExact time
Research data from Epoch AI shows that the training compute for frontier AI models roughly doubles every 6–10 months, far exceeding the growth rate of Moore's Law.
3
Sources
65%
Confidence
Medium-term (~90 days)
Relevance
8/15/2026
First Seen
Valid until: 11/13/2026
Sources
Related Claims
UnverifiedResearch data from Epoch AI shows that the computing power used for AI training has grown by more than 300,000 times from 2012 to the present.80% similarVerifiedThe total energy consumption of the AI inference phase has gradually surpassed that of the training phase, because inference runs continuously 24×7.74% similarUnverifiedTogether AI has reduced training prices for select models.73% similarUnverifiedFrom NVIDIA's A100 to H100 to B200, each generation provides approximately 1.5 to 3 times the AI training compute at the same power consumption.72% similarUnverifiedThe paper projects that if AI R&D is fully automated, the research return rate holds within its historical range, and no new bottlenecks emerge, the pace of AI progress could increase tenfold in approximately one and a half years.71% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/753184API
curl https://kongchang.com/api/v1/knowledge/claims/753184MCP
get_claim(id=753184)