The recent slump in Nvidia's stock price has been accompanied by a decline in the compute per hour price of its flagship B200 graphics processing unit (GPU). According to Ornn, which tracks live GPU compute prices, the B200's price peaked at $6.11 on May 30 but has since dropped to $4.22 as of June 21. This downward trend has led Kalshi traders to bet that Nvidia's chip prices will continue to fall. The company's reliance on GPU sales for data centers has made it vulnerable to fluctuations in demand for AI infrastructure. As companies like Google secure large-scale GPU rental agreements, such as its recent $920 million monthly deal with SpaceX, the dynamics of the market are shifting. While this deal is expected to drive growth for Nvidia in the second half of 2026 and 2027, the broader trend of declining GPU prices raises questions about the company's pricing power. With the cost of renting GPUs fluctuating as demand for AI infrastructure grows, manufacturers like Nvidia face uncertainty in planning production. As the AI buildout continues, the interplay between demand, supply, and pricing will be critical in determining the company's performance. The VanEck Semiconductor ETF's 84% gain this year, driven in part by the rise of memory chips and infrastructure, has outpaced Nvidia's 12% growth, highlighting the complexities of the evolving AI landscape.

Source: CNBC