Unverified50% confidenceFactExact time
According to estimates, training GPT-3 (175 billion parameters) consumed approximately 1,287 MWh of electricity, equivalent to roughly 550 tonnes of CO₂ emissions.
1
Sources
50%
Confidence
Long-term
Relevance
8/31/2026
First Seen
Sources
Related Entities
Related Claims
Unverified同一模型在不同数据中心训练碳排放量可能相差40倍以上，冰岛碳强度约18 gCO₂eq/kWh，依赖煤电地区可能超过800 gCO₂eq/kWh61% similarUnverified美国能源信息署数据显示，火力发电厂每发1兆瓦时电力平均消耗约1500-2700升水，核电站约2000-3400升/兆瓦时，天然气联合循环电厂约700-1100升/兆瓦时59% similarUnverified根据IPCC评估，核电全生命周期碳排放强度约为12克二氧化碳当量/千瓦时，远低于天然气发电（约490克）和煤电（约820克）59% similarUnverifiedAccording to the IEA 2024 Electricity report, data center electricity consumption reached approximately 460 TWh in 2023, and is expected to more than double to over 1,000 TWh by 2026.57% similarUnverified美国整个数据中心行业在2023年的总用电量约为176TWh57% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/832000API
curl https://kongchang.com/api/v1/knowledge/claims/832000MCP
get_claim(id=832000)