TOTAL VOLUME:
$134.1b
24H VOL:
$141,541,542
24H TRANSACTIONS:
2,388,728,490
OPEN INTEREST:
$1,440,096,988
406,065
Markets across
30,522
events
MATCHED EVENTS:
2,692
PLATFORM COVERAGE:
5
Polymarket:
39%
VS.
Kalshi:
61%
Closed: Jul 31, 5:44 PM EST
Kalshi
This event tracks the pricing of GPU compute resources for NVIDIA's RTX 5090 graphics card during July 2026. The market measures the average hourly rental or usage cost of this compute resource, reflecting real-world pricing dynamics in the GPU compute market.
Resolution is determined by the arithmetic mean of hourly compute prices for the NVIDIA RTX 5090 throughout July 2026, as reported by Ornn and expressed in USD, rounded to two decimal places. The event establishes multiple price thresholds ($0.50, $0.75, $1.00, $1.25, $1.50, $1.75, and $2.00 per hour), each resolving independently to Yes if the monthly average exceeds that threshold. Data revisions made after the expiration of July 2026 are not incorporated into the final resolution.
Prediction market odds often diverge meaningfully from traditional analyst forecasts because they aggregate real-money incentives rather than relying on expert opinion alone. Traders betting on this market face direct financial consequences for accuracy, which can reveal gaps between published analyst estimates and market-derived expectations. Analyst reports on GPU pricing typically focus on historical trends and vendor roadmaps, while prediction markets incorporate forward-looking signals from participants with skin in the game. When this market's odds diverge significantly from consensus analyst views, it may signal either underappreciated catalysts or overconfidence in published guidance. Comparing the two reveals where market participants see risk or opportunity that traditional research has missed.
On Kalshi, this market is priced through continuous order-book matching, where buyers and sellers set the odds by placing limit and market orders. On Kalshi, prices reflect that venue's order book, liquidity, and how traders price the outcome right now. The price reflects the marginal probability that the RTX 5090's monthly average compute price will fall within specified outcome ranges during July 2026. As new information surfaces—such as NVIDIA earnings calls, GPU supply announcements, or competitive pricing moves—traders adjust their bids and asks, shifting the market price in real time. Kalshi's transparent order book allows participants to see depth and liquidity at each price level, helping traders execute efficiently and understand where consensus lies.
This market resolves around Aug 1, 2026, after the July 2026 period concludes and the RTX 5090's monthly average compute price can be verified from credible public sources. The outcome is determined by comparing NVIDIA's official or widely reported pricing data against the predefined outcome brackets offered in the market. Resolution hinges on establishing an accurate, defensible figure for the average compute price during that calendar month, which typically comes from NVIDIA's published pricing, earnings disclosures, or independent hardware pricing databases. Once the data is confirmed and the outcome bracket is determined, Kalshi settles all positions accordingly.
Several catalysts could shift this market significantly before resolution. NVIDIA's quarterly earnings announcements and forward guidance on GPU pricing strategy will likely trigger sharp moves, as will any major shifts in AI chip demand or competitive pressure from AMD or other manufacturers. Supply chain disruptions or manufacturing cost changes could influence NVIDIA's pricing decisions. Regulatory developments affecting AI compute access or export controls might also reshape expectations. Industry conferences where NVIDIA unveils new products or pricing tiers, as well as macroeconomic shifts affecting enterprise spending on GPUs, represent key watch points. Real-time tracking of competitor pricing and data-center adoption rates will help traders anticipate where NVIDIA may position the RTX 5090 in its product lineup.