Testing on-demand @NVIDIA B300 on Verda (formerly DataCrunch) provisioned via dstack. Isn't it cool?

Is complaining about the AI bubble just another bubble?

Andrey Cheptsov Constantin Dominique Paul With OSS LLMs - Europe’s on it. Llama was basically an (FAIR & HF) EU product. Qwen/DeepSeek-level work can happen here. I realise I sound like an old fart & I apologise - as an angel investor happy to back the teams building this. Am eg invested in Continue & Rasa.

dstack 0.19.39 is out. 🚀 This release focuses on bug fixes and documentation improvements. * Docs can now be built locally, making contributions easier. * Added /llms.txt and /llms-full.txt to help LLMs generate accurate commands and configs. * Added AGENTS .md to help agents

Wow, amazing how everyone suddenly became a TPU expert overnight.

Early access: B300 (Blackwell Ultra) VMs are now available on Verda (formerly DataCrunch) (former DataCrunch) via dstack. * Available on both on-demand and spot * Up to 8× GPUs per VM * Pricing: $4.95/h on-demand, $1.24/h spot If you need Blackwell nodes for training, eval or inference, you

It's the first neocloud to offer on-demand and spot @NVIDIA B300, and it can be used via dstack!

A new case study is out: Toffee AI breaks down how they use dstack to manage AI inference across neoclouds like Runpod and vast.ai. One control plane, consistent deployments, much less infra overhead. Read it here: research.toffee.ai/blog/how-we-us…

New blog from Toffee AI on how they use dstack to run model inference across multiple GPU clouds like Runpod and vast.ai. We built dstack for GPU-native orchestration, and messy GPU availability and pricing are a big part of that. I love seeing AI startups use it to tame

Stop words update: “banger.”

JetBrains launches a new IDE. Wha do you think? air.dev

Full blog here 👇 aerlabs.tech/blogs/dstack-g… A big shoutout to Andrey Cheptsov the driving force behind dstack, for building something genuinely thoughtful and developer-first. 🙌

November at Nebius brought yet another multi-billion hyperscaler deal and further expansion of our data center footprint, new leading benchmark results and deeper platform ecosystem integrations. Read the digest: nebius.com/blog/posts/dig…

Congrats to Nebius on the $3B Meta deal — a major boost for the cloud AI ecosystem. Exciting to see capacity scaling this quickly for both labs and startups!

Read how Toffee AI simplified their multi-cloud GPU stack with dstack and now ship LLM and image-gen inference across GPU clouds, all while reducing GPU spend by 2-3x. Read the full case study: dstack.ai/blog/toffee/

Excited to see dstack featured in Verda (formerly DataCrunch)'s monthly digest! We're thrilled about the dstack-Verda integration and how it simplifies training and inference orchestration for AI builders.

Over your xmas vacation, you should learn how to play with GPUs... ssh admin.hotaisle.app