As of June 1, 2026, based on Gate TradFi market data, NBIS’s intraday gain exceeds 15%, reaching a high of $272. The price has set a new all-time high again. Behind this strong performance, is it driven by short-term sentiment, or is there long-term, verifiable value support?
Market analysts believe that the growth in AI industry compute demand far outpaces the expansion speed on the supply side. AI compute shortages will persist for at least 3 to 5 years. Hyperscale cloud providers—large tech companies like Microsoft, Amazon, and Google that build their own data centers—are constrained by physical infrastructure construction cycles, making it impossible to build a sufficiently large GPU cluster in the short term. This creates an irreplaceable market space for specialized AI cloud providers like Nebius.
Positioned as a “new cloud” provider in AI infrastructure, Nebius offers enterprise-grade high-performance compute resources based on NVIDIA GPUs by building its own data centers. The core of its business model is: it does not get involved in comprehensive businesses like storage and databases in traditional cloud services, but instead focuses on the vertical of GPU-intensive computing, providing dedicated compute power for AI model training and inference. As AI applications move from technical exploration to industrial deployment, demand for compute in inference scenarios is growing at an exponential pace. According to industry reports, the number of tokens that inference models need to process is 20 times earlier levels, while compute requirements expand to more than 150 times. The AI cloud infrastructure track where Nebius operates is at the forefront of a structural demand surge.
Nebius Revenue Surges Sixfold
Nebius’s financial performance in Q1 2026 provides the most direct data backing for its market narrative. Based on the company’s disclosed financial report data, Nebius Q1 revenue reached $399 million, up 684% year over year from $55.3 million in the same period last year, exceeding market expectations of $371.4 million. The core business segment achieved net profit of $621 million, reversing the loss position from the prior-year period. Adjusted EBITDA also rose from a loss of $53.7 million to a profit of $130 million. The rapid revenue growth is not a one-off event—it is built on a foundation of long-term collaboration with leading technology companies.
In September 2025, Nebius signed a five-year GPU infrastructure supply agreement with Microsoft worth $17.4 billion. In the same year, the company also secured a $27.0 billion, five-year AI infrastructure deal with Meta. Combined, these long-term contracts totaling more than $44.0 billion provide Nebius with revenue visibility far above the industry average level. In March 2026, NVIDIA made a strategic investment of $2.0 billion in Nebius to obtain about 8.3% equity. The two sides plan to jointly deploy more than 5 gigawatts of AI compute infrastructure by 2030. This capital cooperation not only validates Nebius’s technology roadmap, but also means it holds a priority position within NVIDIA’s supply chain system. From a financial perspective, Nebius’s revenue growth rate, customer mix, and the degree of long-term contract coverage together form the core fundamental support behind this round of price increases.
How Token Factory Reshapes the Business Logic of AI Compute
In November 2025, Nebius officially launched the Token Factory inference platform. The platform allows AI enterprises and developers to deploy and optimize open-source or custom models on Nebius’s infrastructure, and to be billed in units of tokens. Token Factory is not a traditional compute rental tool. It integrates capabilities such as high-performance inference, post-training of models, and fine-grained access control, supporting mainstream open-source models including NVIDIA Nemotron, Llama, and DeepSeek. Its technical core is to reduce inference cost and latency by up to 70% through methods like model quantization and distillation, while ensuring 99.9% service-level agreement availability.
In May 2026, Nebius acquired AI inference optimization company Eigen AI for about $643 million, further strengthening Token Factory’s technology stack. Eigen AI’s core team previously led the development of large-model inference optimization technologies such as AWQ 4-bit quantization and SpAtten sparse attention. These technologies can significantly reduce GPU memory usage and improve GPU utilization. The strategic significance of Token Factory is that it upgrades the commoditized “compute rental” business into a platformized service of “compute + software.” NVIDIA founder Jensen Huang defined token as “revenue” rather than “cost” in his June 1, 2026 keynote speech at GTC Taipei. This directly echoes Token Factory’s business logic. When each token generated by a company can be quantified and priced, expanding an AI factory shifts from capital expenditure to a natural choice for capacity expansion.
How Multiple Signals from the Capital Markets Drive Price Up
Behind this round of NBIS price increases, multiple signals from the capital markets create a compounding effect. On June 1, 2026, at the Computex conference held in Taipei, NVIDIA CEO Jensen Huang publicly praised Nebius. In the slide deck he showed, a heart-shaped symbol connects the NVIDIA and Nebius logos. Huang specifically mentioned the quality of Nebius’s customer base, including Cursor, Revolut, Shopify, and other well-known companies.
On the same day, the hedge fund Situational Awareness LP, managed by former OpenAI researcher Leopold Aschenbrenner, filed a 13G with the U.S. Securities and Exchange Commission, disclosing that it holds about 5.6% passive equity in Nebius, totaling about 12.41 million shares. The fund had previously held positions in IREN and CoreWeave, and its purchase of Nebius allowed it to complete a布局 (portfolio setup) across three top targets in the new-cloud track.
Optimism from institutional players is also gaining momentum. Arete Research raised Nebius’s target stock price from $291 to $380, representing the highest target price issued by all Wall Street analysts. Citizens JMP raised its target price to $270, while Bank of America raised it to $205. On the short-pressure side, Nebius’s short position ratio is about 21.1%. Based on recent average daily trading volume of 19.6 million shares, the estimated number of days required to cover is about 2.3 days. This squeeze effect, to some extent, amplifies the upward elasticity of the stock price.
Nebius’s Differentiation in Competitive Landscape on the Same Track
Nebius’s rise is not an isolated phenomenon. On the NeoCloud track it operates in, CoreWeave and IREN are also widely watched targets. The three companies differ significantly in their business models: CoreWeave entered the market first, activating the highest amount of compute capacity, and mainly focuses on AI model training, deeply tied to Microsoft and OpenAI. IREN emerged from Bitcoin miners, and its core advantage lies in locking in large quantities of relatively cheap electricity resources. Nebius, meanwhile, emphasizes building a software stack more, placing its focus on the inference market rather than the training market. Growth in Nebius’s non-megacap customer contracts is rapid, and it also owns high-valuation subsidiaries such as Avride and ClickHouse, giving it a relatively healthy financial structure. On June 1, 2026, Nebius announced that it would raise quotations for new customers for its H100 and B200 class GPUs by 30% to 70%. This pricing move directly responds to prior bearish narratives that GPUs would depreciate quickly and compress profit margins. Pricing power is an important metric for measuring the depth of a compute service provider’s moat. In a market where supply and demand are highly imbalanced, being able to pass upstream hardware cost increases directly through to downstream customers is direct proof of market speech power.
Does Expansion in Capital Expenditure Mean Overextension?
In this round, Nebius significantly raised its 2026 capital expenditure guidance to a range of $20.0 billion to $25.0 billion, expanding far beyond prior expectations. In Q1 alone, spending on equipment and intangible assets reached $2.47 billion, up 355% year over year. This large amount of capital will mainly be used to build a new gigawatt-class AI factory in Pennsylvania, USA, where about 1.2 gigawatts of power resources have already been secured.
To support this aggressive expansion, in March 2026 Nebius issued convertible senior notes via private placement to raise about $4.3 billion, including 1.250% notes maturing in 2031 and 2.625% notes maturing in 2033. Large-scale debt financing combined with high capital expenditure forms the financial pressure Nebius faces at present. If the notes are converted into equity in the future, there will be risks of dilution to existing shareholders’ equity. However, looking at industry cycles, AI infrastructure investment is in a historic expansion phase. Expansion decisions by upstream semiconductor manufacturers such as TSMC and Micron are typically based on forecasts of long-term demand cycles for the industry, rather than short-term prosperity. This is a cycle signal that cannot be ignored for the entire track.
Frequently Asked Questions FAQ
What is Nebius’s core business?
Nebi us is a “new cloud” service provider focused on AI infrastructure. It delivers high-performance compute resources based on NVIDIA GPUs to enterprises through self-built data centers, covering both AI model training and inference scenarios. The company also holds multiple technology assets under its umbrella, including Avride, ClickHouse, and Toloka, covering areas such as autonomous driving, data analytics, and data labeling.
What is Token Factory, and why is it important?
Token Factory is an AI inference platform launched by Nebius in November 2025. It allows enterprises to deploy and optimize open-source or custom large models in units of tokens. The platform integrates capabilities such as high-performance inference, model optimization, and access control, making it a key strategic layout for Nebius’s shift from “compute rental” to a platformized service of “compute + software.”
What are the main drivers behind NBIS’s price increase in this round?
Main drivers include: Q1 revenue up 684% year over year; public endorsement from NVIDIA CEO Jensen Huang at the Computex conference; the fund managed by a former OpenAI researcher building a 5.6% equity position; multiple institutions such as Arete Research raising target prices; Nebius announcing a GPU price increase of 30% to 70% validating pricing power; and the squeeze effect created by short positions.
What is Nebius’s financial foundation?
As of Q1 2026, Nebius revenue was $399 million, up 684% year over year, and adjusted EBITDA turned from loss to profit at $130 million. The company has two long-term contracts with Microsoft ($17.4 billion) and Meta ($27.0 billion), as well as NVIDIA’s strategic investment of $2.0 billion. But the company is currently in a high-capex expansion period, with 2026 capital expenditure guidance as high as $20.0 billion to $25.0 billion.
What are Nebius’s main risks?
Main risks include: potential equity dilution from large-scale debt financing; ongoing cash flow pressure from high capital expenditure; risk of delays in power infrastructure construction—around half of data center projects in the U.S. that are planned are delayed due to grid bottlenecks; and intense competition among peers.
