AITOPAI热报
ProductsSelected

Cohere Launches Confidential Computing Version of Model Vault

Gorden Sun@Gorden_Sun
Post· 2 sources
View the post
Why we picked it

The real shift is that cloud AI inference no longer forces a trade-off between convenience and data custody — regulated industries can now run finance, medical or source-code workloads on shared cloud GPUs with verifiable hardware isolation. Enterprises evaluating private LLM deployment should compare this with on-premise setups. The Azure partnership also signals Cohere's continued toB focus rather than consumer-facing competition.

Cohere has released a confidential computing version of Model Vault, letting enterprises call cloud AI without exposing sensitive data. The system is built with Microsoft Azure and NVIDIA on NVIDIA H100 GPUs equipped with confidential computing technology. Before sending data, local software verifies the remote server's hardware and software security, then transmits encrypted content. Data is decrypted and processed for inference inside a hardware-isolated enclave on the GPU, invisible to Cohere, Microsoft operators and the underlying OS, and results are encrypted on return.

Original · Gorden Sun

Cohere专注做toB业务，活的的太滋润了。

Cohere 推出了机密计算版 Model Vault，让企业在云端调用 AI 时不用担心敏感数据泄露。

很多企业想用大公司的云端 AI 服务，但又不敢把财务报表、医疗记录、核心源代码等机密资料传上去，怕服务商或黑客在后台偷看。

Cohere 联合微软 Azure 与英伟达，基于搭载机密计算技术的英伟达 H100 显卡搭建了防护环境。企业在发送数据前，本地软件会自动核验对方服务器的软硬件是否安全合规，确认无误后才把加密内容发过去。

这样一来，企业的数据从离开自己电脑那一刻起就是全程加密的。数据传到云端后，直接进入芯片划定的一块硬件隔离空间里解密并运行 AI 推理，连 Cohere 官方、微软运维人员以及底层操作系统都无法偷看内存里的明文内容。AI 生成的结果同样会被加密送回。企业既能享受开箱即用的云端算力，又能牢牢守住核心商业隐私。

#Cohere#Microsoft Azure#NVIDIA#confidential computing#enterprise AI

Keep reading

Recent stories with the same tag or section
IndustryBuildersSelected01:07
Cohere's Rachad Alao on open science for foundation models

The talk links open model releases to concrete security practices like independent auditing and dynamic safeguards, rather than treating openness and safety as opposing goals. Researchers and ML engineers can follow the named stack — PyTorch, TorchTitan and vLLM — to reproduce or audit open models themselves. Cohere's stance is notable as a US-listed enterprise lab that publicly backs open science. origin

#Cohere#PyTorch#TorchTitan#vLLM
Event
官方账号PyTorch@PyTorch2 sourcesSource
ModelsUS vs ChinaSelected04:26
Cohere Embed 5 Fast Beats Google EmbeddingGemma 2 on 163 Datasets

Vendors increasingly publish self-run head-to-head embeddings benchmarks, so the 95% win rate should be read as vendor-reported until independent replication appears. Teams selecting an embedding model for retrieval or RAG pipelines can weigh Embed 5 Fast against EmbeddingGemma 2 on their own data instead of relying on one-side numbers. EmbeddingGemma 2 being open-weight makes such local testing straightforward.

#Cohere#Embed 5#EmbeddingGemma 2#embeddings
官方账号Cohere@cohereSource