Unverified50% confidenceFactExact time
M系列芯片在20-40瓦范围内即可完成相当水平的推理任务，而NVIDIA GPU动辄数百瓦功耗
1
Sources
50%
Confidence
Long-term
Relevance
8/27/2026
First Seen
Sources
Related Entities
Related Claims
Unverified英伟达H100 GPU单卡功耗高达700瓦，一个配备5万块H100的超算集群总功耗可达35兆瓦（MW）77% similarVerifiedConsumer-grade GPUs such as the RTX 4090 have 24 GB of VRAM and can run quantized models with billions of parameters.76% similarUnverifiedDeploying an open-source model with 70 billion parameters running at FP16 precision requires at least four A100/H100 GPUs with 80 GB of VRAM each, with hardware procurement costs alone exceeding one million yuan.73% similarUnverified测试者用双4090（48GB显存）工作站运行20.7B稠密模型，每秒生成27.6个Token，仅比矮星略快，但参数量小10倍以上69% similarUnverified承载2万亿参数的GPT-5.6 Sol用NVIDIA芯片只需14颗B300，而用Cerebras则需90颗WSE3芯片69% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/810128API
curl https://kongchang.com/api/v1/knowledge/claims/810128MCP
get_claim(id=810128)