NEW YORK, January 8, 2026, 13:18 EST — Regular session
- Nebius shares rose about 4% in midday trade, after a choppy open
- The AI cloud firm is pitching early access to Nvidia’s next Rubin platform from H2 2026
- Traders are watching a Jan. 29 product webinar and the next earnings update expected mid-February
Nebius Group N.V. shares rose about 4% on Thursday, climbing back toward $100 after sliding the prior session, as investors kept trading around the AI cloud name’s next-generation GPU plans. The Nasdaq-listed stock was last up 3.8% at $100.03, after opening at $96.99 and swinging between $96.06 and $102.49; Nvidia fell about 2% while AI infrastructure peer CoreWeave edged higher.
The move matters because Nebius has become a volatile way to play the scramble for graphics processing units (GPUs) — chips used to train and run large AI models — at a time when investors are trying to separate booked demand from deliverable capacity. In November, the company said it had signed a roughly $3 billion, five-year AI infrastructure deal with Meta, following a $17.4 billion agreement with Microsoft in September, while reporting a sharp jump in revenue alongside heavier spending and a wider loss.
Nebius has been telling investors it can win by building “neocloud” capacity — specialized cloud built around GPUs — and then layering services on top, rather than trying to outspend the biggest hyperscalers. It has said it is expanding in the U.S. and Europe and aims to secure 2.5 gigawatts of power capacity by 2026 as it pushes into more traditional industries. Reuters
Earlier this week, Nebius said it would deploy Nvidia’s Rubin platform through its Nebius AI Cloud and its Token Factory inference service, with Vera Rubin NVL72 systems expected to be available starting in the second half of 2026. “We are proud to be one of the first on the market to offer Vera Rubin GPUs,” founder and CEO Arkady Volozh said in the release. Nebius
Nvidia unveiled Rubin at CES and said the platform is aimed at lowering the cost per token — a basic unit of AI output — for complex “mixture-of-experts” models, a popular design that routes work to specialist sub-models. Nvidia also listed Nebius among companies expected to adopt Rubin, alongside cloud and infrastructure players including Microsoft and CoreWeave. NVIDIA Investor Relations
For traders, the timing is the point: early commitments to next-gen systems can matter when supply is tight and delivery slots are scarce, even if the hardware does not ship for months. It also gives the market something concrete to trade beyond big-contract headlines.
But the setup cuts both ways. Nebius is still in build-out mode, and any slip in data-center timelines, power availability, or GPU delivery could hit revenue recognition and keep losses elevated, while competition from larger cloud providers and other GPU specialists remains intense.
Next up, investors are likely to watch a Nebius “Builder Hour” webinar on January 29 for signs of product uptake and capacity messaging, and then the company’s next quarterly update, which Nasdaq currently estimates around February 18. Nebius+1
