NVIDIA B200 GPU Rental Price Drops to $4.22 Amid Stock Slump

According to Ornn, a platform providing real-time GPU compute price data for major hardware, the hourly rental price of B200 surged to $6.11 on May 30, marking the highest level in the past three months.
However, since that date, the platform’s data shows that the hourly rental price of B200 has continued to decline, falling to $4.22 by the end of last week.
image description


June 23 – NVIDIA’s stock price has been weak recently, and traders on the prediction platform Kalshi are keenly noting that the pricing power of the company’s AI chips is also quietly declining.



Despite the AI chip giant still recording approximately 12% gains year-to-date in 2026, its stock price has been fluctuating downward over the past month, with a cumulative decline of about 3%. In comparison, the VanEck Semiconductor ETF (SMH) has surged 84% so far this year and has soared 15% in the past month alone.



As the AI infrastructure buildout deepens, Wall Street’s fervent attention has quietly shifted toward memory chips and underlying architecture. This shift in sentiment has directly triggered a surge in Micron Technology and SanDisk, with both giants soaring nearly 60% in the past month alone.



During this period, NVIDIA appears to have been relegated to the sidelines.



Notably, as NVIDIA’s stock price remains depressed, the rental price of its B200 chip has also fallen. The NVIDIA B200 is a critical hub supporting the operation of current hyperscale data centers.



According to Ornn, which provides real-time GPU compute price data for major hardware, the hourly rental price of B200 climbed to $6.11 on May 30, setting a three-month high.



However, since that day, the platform’s data indicates that the hourly rental price of B200 has been on a downward trend, dropping to $4.22 by the end of last week.



Meanwhile, Kalshi traders are currently pessimistic about whether the rental price of NVIDIA AI chips can break back above May’s high. The contract is based on the rental price of NVIDIA B200 in the second quarter, and will settle if the price on the Ornn platform exceeds a specific threshold by June 30.



Most companies rent GPUs through cloud service providers or neoclouds, an emerging ecosystem. However, as demand for AI infrastructure grows, the cost of renting GPUs may fluctuate.
“Many people don’t know how much computing power they will need next year, and many current compute suppliers are also unclear about how many GPUs they should order and what capacity is required,” said Seoyoung Kim, a finance professor at Santa Clara University. “And manufacturers like Nvidia also don’t know how much to produce.”



In fact, Rich Privorotsky, head of Goldman Sachs’ One-Delta trading desk, also noted last week that the myth of AI “compute scarcity” appears to be losing its luster. Privorotsky pointed out that the most important indicator remains the rental price of compute resources: If scarcity persists, prices should remain firm and justify sustained capital expenditure. If supply increases and rental prices continue to decline, this will directly challenge the narrative of “compute resource shortage.”



Such pressure will first manifest in the hardware sector. Privorotsky noted that the beneficiaries are those companies that sell complete systems and monetize usage, rather than merely those selling “picks and shovels.” The greater risk lies upstream in the hardware and infrastructure stack, where expectations remain built on sustained scarcity.



Of course, some market participants remain relatively optimistic. Earlier this month, SpaceX and Google just reached a cloud services agreement. Under the agreement, Google agreed to pay SpaceX $920 million per month from October 2026 to June 2029. The computing power provided by SpaceX includes 110,000 Nvidia GPUs, CPUs, and memory.



Following the deal, RBC Capital Markets expressed optimism about Nvidia’s performance in the second half of 2026 and 2027, stating that the chip giant is “in the most advantageous position among its peers.” “Regardless of the specific reasons, these GPU rental agreements should, at least in the short term, alleviate ongoing concerns about Nvidia’s market share being eroded by ASICs,” the analysts wrote.



Leave a Comment

Your email address will not be published. Required fields are marked *