Nvidia is currently navigating a challenging period in the stock market, significantly trailing the broader semiconductor sector throughout 2026. While the VanEck Semiconductor ETF (SMH) has experienced a robust surge of nearly 85% this year, Nvidia’s shares have managed only a 12% gain. This underperformance is compounded by recent downward pressure, with the stock shedding approximately 3% over the past month, even as the wider semiconductor index continues to climb.

Market focus appears to be shifting away from pure GPU dominance toward memory chips and broader infrastructure components essential for the next phase of artificial intelligence development. Companies such as Micron Technology and Sandisk have capitalized on this trend, posting gains of nearly 60% in the last month alone. This rotation suggests that investors are re-evaluating where the most immediate value lies within the AI supply chain.

Adding to the uncertainty, the cost of computing power associated with Nvidia’s flagship B200 GPU has seen a notable decline. After peaking at $6.11 per hour in late May, the compute price for the B200 has fallen to approximately $4.22 as of late June. Prediction markets on Kalshi reflect this cooling sentiment, with traders betting against the likelihood of compute prices returning to their previous highs. This volatility in rental costs highlights the ongoing difficulty in balancing supply and demand for high-end AI hardware.

Despite these short-term hurdles, large-scale infrastructure deals continue to provide a floor for demand. A significant agreement between Google and SpaceX, involving the rental of massive computing capacity utilizing over 110,000 Nvidia components, underscores the continued necessity of the company’s technology. Analysts remain optimistic that such long-term rental agreements will mitigate concerns regarding competition from application-specific integrated circuits, positioning Nvidia to maintain its relevance as the AI buildout matures.

Key Takeaways

• Nvidia's stock has significantly underperformed the broader semiconductor sector in 2026, gaining only 12% compared to the SMH ETF's 85% rise.
• The cost of renting computing power for Nvidia's flagship B200 GPU has dropped from a May peak of $6.11 per hour to $4.22 per hour.
• Investor interest is rotating toward memory chips and infrastructure, though major cloud rental deals suggest sustained long-term demand for Nvidia hardware.

Editor’s Analysis & Impact

The current stagnation in Nvidia’s stock price reflects a maturing phase in the AI investment cycle. Initially, the market rewarded pure-play GPU manufacturers, but as the industry scales, the focus is shifting toward the ‘picks and shovels’ of the AI ecosystem—specifically memory and infrastructure. The decline in B200 compute prices suggests that the initial scarcity of high-end processing power is easing, or that cloud providers are optimizing their utilization rates. While this may dampen short-term margins, it is a natural evolution of a hardware-intensive market. The future outlook remains positive, provided Nvidia can maintain its technological moat against custom silicon alternatives. The long-term viability of the company will likely depend on its ability to secure massive, multi-year capacity contracts with hyperscalers, which serve as a hedge against the volatility of spot-market GPU rental prices.

Frequently Asked Questions

Q: Why is Nvidia underperforming compared to the broader semiconductor market?
A: Investors are currently rotating capital toward memory chip manufacturers and infrastructure providers, which are seen as critical components for the next stage of AI development, causing Nvidia to lag behind the broader semiconductor index.

Q: What does the decline in B200 compute prices signify?
A: The drop in hourly rental prices for the B200 GPU indicates a potential shift in supply-demand dynamics for computing power, suggesting that the extreme scarcity seen earlier in the year may be stabilizing.