Science and Technology
05/26/26
The market tracks whether the hourly compute price for an NVIDIA RTX 5090 GPU exceeds $1.46 by the end of 2026.
The price is driven by the supply constraints of AI data centers and cloud providers competing for GDDR7 memory-enabled hardware.
RTX 5090 hardware prices have already surged to $3,000–$4,000 retail, with forecasts of $5,000 by late 2026 due to persistent GDDR7 shortages.
If the compute price exceeds $1.46/hour, it confirms that hardware scarcity and AI demand have forced rental costs to more than 1.5x the current median of $0.87/hour.
Note: The resolution relies on available data for the time period; if no data is recorded by the expiration date, the market may resolve to "No" regardless of the theoretical price.
Industry sources predict RTX 5090 prices will hit $5,000 by late 2026 as AI demand consumes 20% of global DRAM capacity, creating a structural memory crisis.
byteiota.comCurrent hourly compute prices for the RTX 5090 show a market median of $0.87/GPU/hour across 162 configurations, with some providers starting at $0.33/GPU/hour.
gpus.ioRTX 5090 retail prices currently range from $3,000 to $4,000, having doubled from the $1,999 MSRP due to GDDR7 memory shortages.
byteiota.comPotential announcement of Micron's new HBM facility timeline, though relief for memory shortages is not expected until 2028.
byteiota.comEnd-of-year inventory checks and holiday demand spikes that historically exacerbate GPU price volatility in the fourth quarter.
byteiota.comAI-generated briefing. AI can make mistakes. This is not financial advice.
GDDR7 memory shortages are structural rather than temporary, with AI consuming 20% of global DRAM capacity and no supply relief until 2028, forcing compute prices upward.
byteiota.comIndustry analysts predict RTX 5090 hardware prices will reach $5,000 by late 2026, which would necessitate a significant increase in hourly rental rates to maintain profit margins for cloud providers.
byteiota.comtech4gamers.comCurrent hardware prices have already settled at a new floor roughly 50% above launch MSRP, and structural AI demand has permanently raised memory costs, suggesting compute prices will follow the hardware price trajectory.
tech-insider.orgAI-generated briefing. AI can make mistakes. This is not financial advice.
The current median hourly compute price is $0.87/GPU/hour, which is significantly below the $1.46 threshold, indicating a large gap must be closed in just five months.
gpus.ioWhile hardware prices are high, the wide variance in provider pricing (starting at $0.33/GPU/hour) suggests competitive pressure may keep average rental rates below the $1.46 target despite hardware scarcity.
gpus.ioSome rumors suggest GPU price hikes could be less severe than previously feared, and without a confirmed sudden drop in supply, compute providers may not raise rates to the extreme levels required to hit $1.46/hour.
techradar.comAI-generated briefing. AI can make mistakes. This is not financial advice.