NVIDIA's pricing power is being directly questioned by the market. Trading data from the prediction market Kalshi shows traders are broadly betting that the computing power rental price for NVIDIA's flagship GPU chip, the B200, will fail to return to recent highs before the end of the second quarter. This "alternative data" signal resonates with NVIDIA's relatively subdued stock performance lately, making the market more cautious about the AI chip giant's near-term outlook.
According to data from the GPU computing power real-time price tracking platform Ornn, the hourly computing power price for NVIDIA's B200 chip climbed to a three-month high of $6.11 on May 30th, before declining steadily. As of June 21st, it had fallen to $4.22, a cumulative drop of over 30%. Concurrently, NVIDIA's stock price has fallen approximately 3% over the past month, while the Philadelphia Semiconductor ETF (SMH) has gained 15% over the same period. Memory chip stocks like Micron Technology and Sandisk have each surged nearly 60% in the past month.
The relevant contracts on the Kalshi platform are set to settle on June 30th, based on the B200 computing power price displayed on the Ornn platform. The current distribution of trader positions indicates that most believe the B200 price will struggle to surpass its late-May peak level before that date. This pessimistic bet directly targets the core question of whether NVIDIA can maintain its strong pricing power during the AI infrastructure build-out cycle.
Stock Lags Peers, NVIDIA Experiences Relative Stagnation
Although NVIDIA is still up about 12% year-to-date, its recent performance has clearly lagged behind the semiconductor sector as a whole. The SMH ETF has accumulated a substantial 84% gain year-to-date and a 15% gain for the month, creating a stark contrast with NVIDIA's trajectory.
Market attention is shifting. Wall Street is currently focused on memory chips and the next phase of AI infrastructure construction, with Micron Technology and Sandisk becoming primary targets for capital inflows, each gaining nearly 60% over the past month. NVIDIA has become noticeably marginalized in this round of market action, with investors growing increasingly sensitive to its lack of short-term catalysts.
B200 Computing Power Price Decline Reflects Uncertainties on Both Supply and Demand
The decline in B200 computing power prices reflects deep-seated uncertainties on both the supply and demand sides of the AI infrastructure market.
"Many companies don't know how much computing power they will need in the coming year, computing power providers don't know how many GPUs to order or at what scale, and manufacturers like NVIDIA don't know how much to produce," Santa Clara University finance professor Seoyoung Kim previously told CNBC. This three-way information asymmetry makes volatility in GPU rental prices common and makes it difficult for the market to form stable expectations for NVIDIA's revenue stream.
Most companies rent GPU computing power through cloud service providers or emerging "neocloud" platforms, with rental prices fluctuating with changes in AI infrastructure demand. The B200 price falling over 30% from its peak suggests a relatively loose supply of computing power in the short term, or a slowdown in the pace of demand growth.
Google-SpaceX Mega-Deal Provides Support, RBC Bullish on Second-Half Prospects
Despite recent weak data, NVIDIA still has fundamental support. Earlier this month, Google and SpaceX signed an agreement for Google to pay $920 million per month from October 2026 to June 2029 to rent AI computing power, which will utilize approximately 110,000 NVIDIA GPUs and related supporting hardware.
Following the announcement of this deal, RBC Capital Markets expressed optimism about NVIDIA's performance in the second half of 2026 and 2027, stating that NVIDIA is "in the most favorable position among its peers." Analysts wrote in a report: "Regardless of the specific reasons, these GPU rental agreements should help alleviate persistent market concerns about NVIDIA's share being eroded by custom chips (ASICs), at least in the near term."
The Signaling Value of Alternative Data
The reason Kalshi's betting data is drawing attention is that it provides a market signal independent of traditional sell-side research. The price mechanism of prediction markets directly reflects the actual positions and probability judgments of trading participants, rather than analysts' subjective forecasts.
Currently, Kalshi traders' positioning is skewed towards the B200 computing power price failing to break through its May high before the end of the second quarter, which aligns with NVIDIA's recent stock price trajectory. For investors, this means the market is still searching for a new equilibrium point between the long-term optimism brought by major procurement agreements like Google's and the short-term pressure on computing power pricing. Whether NVIDIA's pricing power can be validated in the next quarter will be a key indicator for observing the trajectory of its fundamentals.
Comments