Our MLPerf® Training v6.1 results: 1st on @NVIDIA Vera Rubin at rack scale, in both pre-training and post-training. NVIDIA Vera Rubin NVL72 on Nebius trained up to 1.63x faster than NVIDIA Grace Blackwell Ultra NVL72 at matched scale. Plus updated Inference v6.1 results 👇
https://t.co/rK2J91TkHV
Nebius is a key partner of Slush 2026. 🇫🇮
Whether you’re training on thousands of GPUs or running inference at scale, come talk to the team about AI Cloud and Token Factory. Our startup team will be there all week at booth #7B23.
See you in Helsinki!
#Slush2026#Nebius #AICloud
I'm long $NBIS, and with all the FUD lately I doubled my position this past week.
So with $META now dropping the news that it's building a cloud business to sell its excess AI compute, this deep dive feels timely. In it I break down:
- Why hyperscalers are renting the overflow instead of building.
- The $46B backlog, and the two-customer risk sitting under it
- Why NBIS owns all three layers of the neocloud stack
- The 2030 case, and what has to go right for it
- Why I think it's becoming the AWS of inference
On this one, I got to leverage my network ⚡
The piece was reviewed by someone inside Nebius, who added their own insight on where the inference market is heading.
Straight from the source, not a pitch deck.
Thanks for your time @demian_ai, I really appreciate the input. ⚡
Go check it out for free on my substack:
https://t.co/NS03YNgLAA
Bullish Neoclouds.
Long $NBIS.
We’re back with more data center ASMR. This one's for anyone whose happy sound is the slide and click of a server seating into a rack. Fresh tracks every Friday through July.
https://t.co/THXDEmaLDm
Season 2 of the Nebius for Startups Podcast launches today. Our first guest, @taiuti, CEO of @reactorworld, discusses what happens when the training data you need simply doesn't exist, and how that problem led him to world models with host @Josh_Liss. 🎧 https://t.co/2dNHyS6OJr
NVIDIA Nemotron™ 3 Ultra is now live on Nebius Token Factory.
It’s built for long-running agents across coding, deep research and enterprise workflows.
Nemotron 3 Ultra delivers frontier reasoning with up to 5x faster inference and up to 30% lower cost for agentic workloads.
For builders, the next question is production: performance, reliability, economics and control.
Run it today: https://t.co/kHWvYquwoy
Berlin on May 28 for Applied AI Conference 2026. At $NBIS we’re building the AI cloud for teams shipping real AI products. Will also be catching up with a few customers on the sidelines. Ping me if you’ll be there.
Nebius founder and CEO Arkady Volozh joined @Accel's Spotlight On podcast to discuss our approach to building for the AI era, including full-stack platform strategy, our partnership with NVIDIA, contract with Meta, capacity planning at scale, and the expansion of our inference platform.
Listen to the full episode: https://t.co/ieQZJ1kUkk
@Atlassian I just usually write my (- name) and if it’s a to a friend who is colleague or to my manager then I write stuff like “Live Love Laugh “ etc . 😂
What is art? In this case, it’s decommissioned datacenter materials.
Our circular centers are one way we cut down on e-waste by repurposing electronic components at the end of their lifecycle.
ℹ️ Note: WhatsApp is currently experiencing outages in multiple countries; incident not related to country-level internet restrictions or filtering #WhatsAppDown