Amsterdam-based Nebius Group, an AI infrastructure firm, introduced on Monday that it has entered into definitive agreements for a $700M (roughly €667M) personal placement financing from a choose group of institutional and accredited buyers.
It consists of participation from Accel, NVIDIA, and sure accounts managed by Orbis Investments.
The Dutch firm will use the funds to additional construct out its full-stack AI infrastructure – together with large-scale GPU clusters, cloud platforms, and instruments and providers for builders.
Arkady Volozh, founder and CEO of Nebius Group, says, “The inspiration of our enterprise is our experience in constructing superior know-how infrastructure. We’ve demonstrated the size of our ambitions, initiating an AI infrastructure build-out throughout two continents. This strategic financing offers us further firepower to do it sooner and on a bigger scale. I’m grateful to our buyers for the belief they’ve positioned in us – our group is able to ship.”
Issuing Class A shares
Nebius Group will situation 33,333,334 Class A shares for $21.00 every in a non-public placement. This value is about 3% greater than the typical value of the shares since buying and selling began once more on Nasdaq. The deal will shut as soon as normal situations are met, says the corporate within the press launch.
Moreover, the Board has determined that the Firm now not wants to purchase again its Class A shares after seeing sturdy buying and selling and liquidity since buying and selling resumed on Nasdaq on October 21, 2024.
In August 2024, shareholders accepted a plan to repurchase as much as 81M Class A shares at a most value of $10.50 every. This value was primarily based on obtainable money after promoting the Firm’s Russian enterprise.
John Boynton, Chairman of the Board of Nebius, says, “The authorisation to doubtlessly repurchase shares was initially meant to supply legacy shareholders who needed to exit our new enterprise a possibility to take action, particularly in gentle of the extended suspension of buying and selling on Nasdaq. Based mostly on the sturdy degree of investor engagement and technical dynamics that we’ve got noticed following the resumption of buying and selling on Nasdaq, we consider that these shareholders who could have needed to exit have had a possibility to take action at a value greater than the utmost repurchase value authorised by shareholders,”
“The Board has decided that the easiest way to maximise worth for the Firm’s shareholders is to take a position our capital into our core AI infrastructure enterprise, the place the Firm believes there’s a substantial market alternative,” provides Boynton.
Because of the mix of strategic financing and the choice to not deploy any capital towards repurchasing Class A shares, the Firm now expects to ship an annualised run-rate income (“ARR”)1 by year-end 2025 of $750M to $1B.
Nebius Group: Constructing AI infrastructure
Based by Arkady Volozh, Nebius is a know-how firm that gives the infrastructure wanted for the worldwide AI business. This consists of GPU clusters, cloud platforms, and instruments for builders.
The corporate’s core enterprise is an AI-centric cloud platform constructed for intensive AI workloads.
With proprietary cloud software program structure and {hardware} designed in-house (together with servers, racks, and knowledge centre design), Nebius offers AI builders the computing, storage, managed providers, and instruments they should construct, tune, and run their fashions.
The AI-native Nebius GPU cloud is designed to handle the total ML lifecycle – from knowledge processing and coaching to fine-tuning and inference – multi functional place.
The brand new Nebius AI Studio inference service expands the corporate’s choices for app builders. It gives entry to superior open-source fashions in a user-friendly surroundings at aggressive costs per token.
The Dutch firm, a most well-liked cloud service supplier within the NVIDIA Associate Community, presents high-end infrastructure optimised for AI coaching and inference.
The corporate’s core AI infrastructure division is supported by a group of roughly 400 expert engineers, delivering a real hyperscale cloud expertise tailor-made particularly for AI builders.
