A landmark year: highlights of Nebius’ journey in 2024

While the holiday season is perhaps not all about numbers — and more about not thinking about them for a while — they still pop up here and there. Whether it’s the GPU capacity you’re planning for next year or how many billions of parameters you’ll use in your next model, numbers are driving us into 2025. Here are some figures we are reflecting on these days — while keeping a sense of fun about them.

Our recent investment round points the way to a bright future, and we’re delighted to welcome some prestigious investors on board as new shareholders.

This $700 million adds to our own savings and gives us more to invest in building out AI infrastructure worldwide, while never forgetting our roots.

Nebius’ first data center in Finland, our home base, was entirely self-built, with meticulous attention to the building structure, adapting it to our in-house server and rack designs. And we structured it not just for cost-efficiency but also in the very Finnish way — with an on-site sauna.

This may seem pretty hot — admittedly, most employees prefer something closer to 80-100 °C.

Speaking of heat, we’re putting the excess heat from our servers to good use by redirecting it to the nearby town of Mäntsälä.

The pipes that deliver heat are, of course, not the only part of the data center’s communications. To support multi-host training for our clients, we rely on InfiniBand. In fact, the total length of optic cables installed in our data centers spans the full distance from our HQ in Amsterdam to Rome.

Of course, we visited Rome in 2024 to promote Nebius – along with 26 other cities across the United States, Europe, APAC and beyond. You’ll find us at AI conferences globally, ready to support participants wherever they may be.

Support means not only providing GPU infrastructure but also taking care of people in a warm and thoughtful manner. If the number of your socks is not even, well… that, of course, is something that never happened to us.

Going to conferences all around the world doesn’t mean we don’t have key places of interest. Our billboards in the Bay Area caught plenty of attention — exactly what we wanted.

If we’re this bold in marketing, how strong are our technological foundations? One way to measure it is through a patent portfolio. These numbers show we’re not just building AI infrastructure like others who came before us. Instead, we’re advancing technologies to push the industry forward.

Nebius brings additional technological value to the field. How else would you describe a cloud computing platform rewritten from scratch in less than a year? We own every single part of the Nebius tech stack, and every line of code is all our own. We think this makes Nebius truly unique in the cloud world.

As well as a cloud we also launched Nebius AI Studio, which gives GenAI app builders with endpoints for dozens of the most popular open-source models, as well as some more specialized offerings. These aren’t limited to LLMs but include vision models too. AI Studio is already making waves in the inference market.

Expanding our portfolio with tech products like these requires close collaboration between teams, especially when growing across multiple countries. To stay connected, we rely on Zoom for cross-border teamwork.

After productive brainstorming sessions, both online and offline, on how to make life even easier for AI developers, we make every effort to reconnect with our families. A sustainable way to commute is by bike — a very Amsterdam thing to do.

That’s a wrap for our numbers in 2024! Now it’s time to think about how to achieve even more impressive figures in 2025. If conquering the AI world is also on your agenda for next year, you know where to find infrastructure. And socks.

Nebius team

See also

Q4: Nebius cloud updates

After completing a full rewrite of our cloud from the ground up in October, we began improving the foundation that we brought to the market. The features and tools we have introduced since then span our entire range of cloud services, from Compute Cloud with NVIDIA accelerated computing in a new region, to managed MLOps solutions.

Nebius opens pre-orders for NVIDIA Blackwell GPU-powered clusters

We are now accepting pre-orders for NVIDIA GB200 NVL72 and NVIDIA HGX B200 clusters to be deployed in our data centers in the United States and Finland from early 2025. Based on NVIDIA Blackwell, the architecture to power a new industrial revolution of generative AI, these new clusters deliver a massive leap forward over existing solutions.

Nebius opens its first availability zone in the United States

Scheduled to go live in Q1 2025, the Kansas City availability zone will house thousands of state-of-the-art NVIDIA GPUs, primarily H200 Tensor Core GPUs in the initial phase, with the energy-efficient NVIDIA Blackwell platform expected to arrive in 2025.

Sign in to save this post