Wire flash
Nebius raises GPU cloud prices 20% effective Oct 1
Editorial responsibility
- No named human review is recorded for this page.
- Source reporting is collected, normalized, translated or condensed automatically when needed.
- Automatically published source-backed update
Nebius announced a comprehensive price increase for its GPU cloud services averaging 20%, effective October 1, covering H100, H200, B200, and B300 chips. This marks the company's second hike since May 2025, when it raised on-demand prices 29% and spot prices 51%. Market analysts attribute the increases to sustained strong demand for AI computing power, with Nebius management revealing customer bookings extending into Q1 and Q2 of 2028 and demand visibility beyond 24 months. The price hike, coinciding with Nvidia's strong earnings guidance, boosted Nebius stock nearly 7% in after-hours trading and lifted other AI infrastructure stocks including CoreWeave, Marvell Technology, and Coherent. The supply-demand imbalance is shifting bargaining power toward data center operators, with contract terms becoming more favorable to suppliers. Nebius has secured major contracts including a $27 billion deal with Meta and a $2 billion investment from Nvidia, and its stock has gained approximately 150% year-to-date.
Source report
Nebius has announced a comprehensive price increase for its GPU cloud services, effective October 1, with an average hike of approximately 20%. The announcement, combined with Nvidia's strong quarterly earnings report, drove Nebius stock up nearly 7% in after-hours trading, reinforcing market expectations that the supply-demand balance for AI computing power remains tight.
Price Increase Details: Significant Hikes Across Multiple GPU Models
The price adjustment covers several chip models, including H100, H200, B200, and B300. According to pricing screenshots circulated on the X platform, per-GPU-hour rates have increased as follows:
- H100: $3.85 → $4.50 (+16.9%)
- H200: $4.50 → $5.40 (+20%)
- B200: $7.15 → $8.50 (+18.9%)
- B300: $7.85 → $9.50 (+21%)
This marks Nebius' second price increase since May of this year. In May, the company announced an average increase of 29% for on-demand capacity and 51% for spot capacity. For the B300, calculated from its pre-May price of approximately $6.10/hour, the cumulative increase now stands at roughly 56%.
In an email to customers in May, Nebius stated that the adjustments "reflect sustained strong demand for high-end GPU computing power, and even after the new pricing, Nebius continues to offer some of the most competitive GPU infrastructure prices in the market." As of press time, Nebius had not publicly responded to the latest price increase rumors, according to Stocktwits.
Market Reaction: Broad Sector Uplift
The pricing news, coinciding closely with Nvidia's impressive earnings guidance, reinforced market confidence in the continuation of the AI data center construction cycle. Key after-hours movements included:
- CoreWeave: +4%
- AAOI and Credo Technology: +3%
- Marvell Technology and Astera Labs: +2%
- Coherent: +2%
Supply-Demand Logic: Demand Visibility Exceeds 24 Months
Market analysts point to deeper supply-demand dynamics behind the price increases. According to X user @MelvinInvests, Nebius previously attempted manual price adjustments, but management indicated that demand was so strong that price adjustments alone could not balance supply and demand. The company subsequently conducted auction tests for scarce Blackwell computing power, where customers paid 15% to 20% above previous maximum list prices to secure access.
Key observations from the analyst:
- Customers have already booked computing power for Q1 and Q2 of 2028, with some orders involving tens of thousands of GPUs
- Current demand visibility has extended beyond 24 months, up from approximately 18 months previously
- The price increase covers both older Hopper architecture GPUs and newer Blackwell series chips, indicating broad-based demand across AI computing power types
"Nebius is benefiting from two forces simultaneously—continuously expanding computing capacity on one hand, and raising fees for existing capacity on the other," wrote @MelvinInvests. This suggests that, assuming high utilization rates, revenue per available GPU-hour will increase significantly, helping to improve profit margins and accelerate infrastructure investment recovery.
Fundamental Backing: Large Contracts and Accelerating Expansion
Nebius' recent fundamentals have supported market sentiment. According to Stocktwits, the company has accelerated expansion since securing a major contract with Microsoft last September, subsequently acquiring Staryps, Eigen AI, and Tavily. In March of this year, Nebius announced a $27 billion contract with Meta Platforms and received a $2 billion strategic investment from Nvidia.
Nvidia's quarterly earnings report, released Wednesday, further confirmed robust upstream demand, strengthening the market logic that AI data center construction by large tech companies will continue to drive demand in the computing power supply chain.
X user Jonah Lupton commented that Nebius' average price increase of ~20% implies significant bottlenecks in computing power supply within the AI ecosystem. He estimated that Nebius' current short-term customer contract prices may have already exceeded $60 million per megawatt, with future contract prices for Vera Rubin architecture products potentially breaking through $80 million per megawatt.
Stock Performance and Sentiment
Nebius stock has accumulated a gain of approximately 150% since the beginning of the year, making it one of the top-performing AI concept names. Data from Stocktwits shows:
- Discussions related to NBIS surged by more than 120% in the past 30 days
- Follower count increased by 7.3%
- Retail investor sentiment maintained a "bullish" rating as of Thursday morning
- Message volume at a "high" level
Shift in Bargaining Power: Data Center Operator Contract Terms Begin to Reverse
The supply-demand imbalance reflected by Nebius' price increases is reshaping contract negotiations across the data center industry.
Previously, major cloud providers held dominant positions in negotiations, often imposing stringent requirements: server racks needed near-100% uptime, with strict standards for temperature and humidity. One data center executive reported seeing contract clauses stating that if a rack went down due to power outage, overheating, or switch failure, the cloud provider could cancel six months of rent. If Service Level Agreement (SLA) breaches accumulated, the cloud provider could directly terminate the lease.
The executive noted that negotiating SLAs involves a trade-off between "optimal price" and "contract durability"—stricter terms command higher prices but carry greater risk. "Securing an SLA with lighter penalties is worth it even if the price is slightly lower."
However, as operators' bargaining power rises, these extreme clauses are being softened. Signs of a shift have also appeared in payment terms. A credit executive cited a case where a client rented only a small portion of a large data center, but the contract stipulated that if the client failed to pay on time, they would be liable for the full rent of the entire facility for a period. The data center owner reportedly admitted: "Look, we know this is outrageous... but we can just do it."
With cloud providers like Microsoft eager to get Nvidia server racks operational as soon as possible, the bargaining leverage of data center operators like CoreWeave is rising, and the initiative in contract negotiations is increasingly shifting to the computing power supply side.
Special Statement: The above content (including pictures and videos if any) is uploaded and posted by a user of NetEase Hao, which is a social media platform and only provides information storage services.
Source
网易财经Neutral / independent
Part of this Story
Nebius raises GPU cloud prices up to 21% as AI demand outstrips supply