AITOPAI热报
IndustryMulti-source

AWS 每月收货数十亿美元 Nvidia 服务器，因缺电囤置仓库

Amazon mua GPU AI trị giá hàng tỷ USD rồi để trong kho vì chưa có điện để bật

Genk
Blog/Media· 9 sources
Read the original
Why we picked it

AWS 每月收几十亿美元的 Nvidia 服务器却开不了机，因为没电。算力大战已经卷到电力供应这一步了。

AWS 每月接收价值数十亿美元的 Nvidia GPU 服务器。部分数据中心尚未完工或电力供应不足，导致这批 GPU 暂时无法投入使用，只能存放在仓库。此事反映电力供应已成为 AI 算力扩张的主要瓶颈。

Original · Genk

Amazon mua GPU AI trị giá hàng tỷ USD rồi để trong kho vì chưa có điện để bật

AWS được cho là đang nhận hàng tỷ USD máy chủ Nvidia mỗi tháng, dù một số trung tâm dữ liệu chưa hoàn thiện hoặc chưa có đủ điện để vận hành số GPU này.

#AWS#Nvidia#GPU#数据中心

Keep reading

Recent stories with the same tag or section
ProductsBuildersOfficial00:40
Chinese contentBayer Crop Science 在 Amazon Bedrock AgentCore 上构建农业多智能体决策引擎

拜耳用 Amazon Bedrock AgentCore 搭了个多智能体引擎，育种人员直接用自然语言查数据，文章还拆了 5 个关键设计决策，做企业级 Agent 的可以抄作业。

#Amazon Bedrock#AgentCore#Bayer Crop Science#AWS
Event
官方一手AWS Machine Learning Blog@Subbu Dandamudi2 sourcesSource
ProductsBuildersSelected23:32
AWS open-sources TorchNeuron PyTorch backend for Trainium

Open-sourcing the Trainium backend lowers the barrier for teams to run PyTorch workloads on AWS's in-house silicon without vendor-specific rewrites. Teams benchmarking inference costs against NVIDIA GPUs can now inspect and extend the translation layer directly. The agent-driven validation approach across hundreds of thousands of models is also a reference point for automated compatibility testing.

#TorchNeuron#AWS#Trainium#PyTorch
官方账号Amazon Science@AmazonScienceSource
ModelsBuildersSelected04:34
AWS Priming Converts Transformers Into Efficient Hybrid Models

Priming makes hybrid architectures practical by skipping full pretraining, which could shift how teams convert existing checkpoints into cheaper-to-serve models. Engineers deploying 8B-32B scale models can use the released weights to cut inference memory and cost immediately. It competes with from-scratch hybrid efforts like those from other labs at a fraction of the compute spend.

#AWS#hybrid models#sparse attention#state space models
官方账号Amazon Science@AmazonScienceSource