AI Demand Remains Strong! CoreWeave (CRWV.US) and Nebius (NBIS.US) Both Raise Prices for Computing Services

According to Golden Ten Data, AI cloud infrastructure provider CoreWeave (CRWV.US), backed by Nvidia, announced that since June 30, 2026, the company has continued to sign new compute capacity contracts at higher prices, including short-term deals reached in the third quarter. Reportedly, in the third quarter of 2026, CoreWeave signed contracts with clients for terms of approximately three to six months, priced at around $40 million per megawatt. This price is calculated based on the ratio between annualized revenue and the power required to operate the relevant clusters. CoreWeave also reiterated that the company secured over $25 billion in new client net commitments at the start of the third quarter, which are not included in the revenue backlog as of June 30. As of August 11, the company’s total contracted power capacity increased to approximately 4.2 gigawatts, compared to about 3.7 gigawatts on June 30.

It is worth noting that the news of CoreWeave signing new compute capacity contracts at higher prices in the third quarter echoes reports that its cloud peer Nebius (NBIS.US) has raised its compute service prices, further confirming the market’s perception of sustained strong demand for AI compute power.

However, as of press time, CoreWeave was down 0.33% in pre-market trading on Thursday, underperforming Nebius and IREN (IREN.US), which were up more than 9% and 5% in pre-market trading respectively. This may be related to the company's plan to sell up to 35 million Class A common shares.

According to a notice sent by Nebius to its clients, Nebius will increase several on-demand compute resource prices starting October 1, with the latest Nvidia GPU seeing the largest hikes. Under the new pricing, the H100 will rise from $3.85/GPU hour to $4.50/GPU hour, an increase of about 17%; H200 will rise from $4.50 to $5.40, an increase of 20%; B200 will rise from $7.15 to $8.50, an increase of nearly 19%; Nvidia B300 has the largest absolute increase, rising from $7.85 to $9.50/GPU hour, up about 21%.

This price adjustment covers both the older Hopper architecture GPUs and the newer Blackwell series, indicating that market demand is not limited to a single chip generation but is distributed broadly across various types of AI compute power. This is Nebius’s second round of price increases within a few months. In May this year, the company announced a 29% average increase for on-demand capacity and a 51% increase for preemptible capacity. For example, the price of B300 has risen by about 56% compared to approximately $6.10/hour before May.

It is reported that Nebius previously attempted to adjust prices manually, but management indicated that strong demand meant such adjustments were still insufficient to balance supply and demand. Subsequently, the company ran auction tests for scarce Blackwell compute capacity, with clients actually paying 15%-20% above the previous highest price to secure access. Futurum Group CEO Daniel Newman posted on X: “AI compute demand is still off the charts. Nebius’s 20% price increase is just another example of strong demand.”

In addition, the news that CoreWeave signed new compute capacity contracts at higher prices in the third quarter may also boost investor optimism regarding the company’s Q3 performance. The company’s second-quarter revenue and profit both exceeded Wall Street expectations, and its backlog surpassed the $100 billion mark—up 246% year-on-year to about $104 billion, continuing to climb from $99.4 billion at the end of the previous quarter. Based on robust demand momentum, CoreWeave has also raised its full-year outlook. As one of the few newly listed cloud computing providers, CoreWeave’s performance is regarded as a key indicator of overall AI compute demand, and its better-than-expected results further confirm the ongoing surge in AI infrastructure investment.