This event has ended. Showing historical data.
August 2026: Monthly average compute price of NVIDIA's H200
$0.00
1
11
August 2026: Monthly average compute price of NVIDIA's H200
$0.00
1
11
AI Analysis
Trader mode: Actionable analysis for identifying opportunities and edge
About This Event
August 2026 If the average value of NVIDIA H200 compute per hour is above X in August 2026, then the market resolves to Yes. The market resolves based on the average value of NVIDIA H200 compute per hour in August 2026, calculated as the arithmetic mean of hourly values reported by Ornn. Revisions to the underlying made after expiration will not be accounted for. The "USD" iteration of the index will be used unless explicitly stated otherwise. Values are rounded to two decimal places unless oth
Current Market Outlook
Kalshi traders are pricing a 98% probability that NVIDIA's A100 compute price averages above $0.50 per hour in August 2026. That near-certainty reflects a market consensus that GPU rental prices will remain well above this modest threshold, but it also reveals how little uncertainty traders see in the underlying data source.
The $0.50 threshold matters because it sits far below current market rates. As of late 2024, A100 rentals on major cloud platforms typically range between $1.50 and $2.50 per hour depending on configuration and provider. Even with aggressive price declines, reaching $0.50 would require a 65% to 75% drop over 21 months. That would represent an unprecedented collapse in GPU pricing.
Key Factors Driving the Odds
The dominant force keeping odds at 98% is the persistent supply-demand imbalance in AI compute. Cloud providers including AWS, Azure, and Google Cloud continue to report GPU shortages, with A100s still heavily utilized despite the newer H100 and H200 generations. Enterprise customers signing multi-year reserved capacity contracts lock in pricing that anchors the market above $1.00.
Historical depreciation patterns also support the high probability. Previous GPU generations like the V100 maintained rental prices above $0.50 for more than four years after launch, even as newer hardware entered the market. The A100 launched in 2020, and its rental prices have remained stubbornly elevated through 2024. A drop to $0.50 by mid-2026 would require an acceleration in depreciation that has no precedent in recent GPU pricing history.
The Ornn index methodology adds another layer of stability. The index averages hourly compute prices across multiple providers, smoothing out short-term volatility. For the monthly average to fall below $0.50, virtually every major provider would need to slash prices simultaneously, which competition alone rarely forces.
What Could Change These Odds
The most credible bear case involves a surge in alternative compute supply. If AMD's MI300 series and Intel's Gaudi 3 capture meaningful market share by 2026, providers might discount A100s aggressively to clear inventory. A global economic downturn could also suppress AI infrastructure spending, though GPU rental prices historically prove sticky even during recessions.
The index's reliance on Ornn's reported values creates a specific risk. If Ornn changes its provider coverage or methodology before August 2026, the reported average could diverge from actual market conditions. Traders assuming the index accurately reflects reality could be surprised by methodology-driven shifts.
The 2% probability of "No" is essentially pricing in a black swan scenario. Given that the A100 remains the workhorse of AI training and inference workloads, and that cloud providers have shown little appetite for price wars, the market's confidence appears justified. Anyone considering a contrarian position would need a specific thesis about why GPU prices would collapse faster than they ever have before.
AI-generated analysis based on market data. Not financial advice.
Overview
NVIDIA's A100 GPU, introduced in 2020, is a data center accelerator that has become a benchmark for AI compute. The A100 powers large language model training and inference, and its market price per hour is a key indicator of AI infrastructure costs. The prediction market question asks whether the average hourly price of A100 compute in August 2026 will exceed a threshold (X), based on data from Ornn, a provider of compute price indices. This market is relevant to businesses, researchers, and investors who track AI cost trends and the balance of supply and demand for GPU resources. Ornn, a company that aggregates compute prices from cloud providers and other sources, publishes indices for various GPUs, including the A100. The 'USD' iteration of their index is used, and the arithmetic mean of hourly values for August 2026 determines the outcome. Revisions to Ornn's data after expiration are ignored, so the market relies on the initial published values. This creates a clear, if somewhat narrow, resolution criterion. The price of A100 compute has seen significant volatility. In 2023, demand surged due to the generative AI boom, leading to shortages and high prices. By 2025, supply had increased, and newer GPUs like the H100 and H200 began to displace the A100 in some workloads, putting downward pressure on prices. However, the A100 remains widely used for inference and cost-sensitive training, so its price is not expected to collapse entirely. The August 2026 price will reflect ongoing supply dynamics, the pace of AI adoption, and the depreciation curve of the A100 as newer architectures become standard. Interest in this market stems from the broader significance of AI compute costs. For startups and enterprises, the price of GPU compute directly affects the economics of AI products. For investors, it signals the health of the AI hardware market and the competitive position of NVIDIA. The market also serves as a practical example of using prediction markets to forecast commodity-like prices in the tech sector.
Historical Context
The A100 GPU was announced in May 2020, with availability starting in late 2020. Initially, it was used for scientific computing and early AI models. The price per hour on cloud platforms was relatively stable, around $1-2 per hour for on-demand instances, but the landscape changed dramatically in late 2022 with the release of ChatGPT, which triggered a surge in demand for AI compute. By 2023, A100s were in short supply, with lead times stretching to months, and prices on secondary markets soared. Cloud providers rationed capacity, and some startups reported paying premiums for access. In response, NVIDIA ramped up production and introduced the H100, which offered better performance but at a higher price. As H100 supply increased in 2024, the A100 became the 'budget' option for many workloads. Prices began to moderate, and by 2025, the average hourly price for A100 compute was around $1.50, according to indices like Ornn's. The introduction of the H200 and Blackwell architecture further pressured A100 prices, as users migrated to newer GPUs for training, while A100s remained viable for inference. This historical trajectory suggests that by August 2026, the A100 will be a mature product, with prices likely to continue declining but not to zero, as it remains useful for certain applications.
Why It Matters
The price of A100 compute is a proxy for the cost of AI infrastructure. For startups and researchers, a lower price means more affordable experimentation and deployment, potentially accelerating innovation. Conversely, higher prices can stifle smaller players, concentrating AI development in well-funded organizations. The outcome of this market will indicate whether AI compute is becoming a commodity or remaining a scarce, high-value resource. For NVIDIA, the A100 price affects revenue from data center sales and the overall profitability of its GPU business. If A100 prices hold up, it suggests sustained demand for older GPUs, which could influence NVIDIA's product roadmap and pricing of newer models. For cloud providers, the price affects their margins and competitive positioning. The market also has broader economic implications, as AI adoption in industries like healthcare, finance, and transportation depends on compute costs. A significant price drop could lower barriers to AI integration, while a stable price might indicate a maturing market.
Current Status
As of early 2025, the average price of A100 compute per hour, as tracked by Ornn, is approximately $1.20-$1.80. The price has been gradually declining from a peak of over $2.00 in 2023, as supply has caught up with demand and newer GPUs have absorbed some of the load. However, prices have not fallen as quickly as some expected, due to persistent demand for AI inference and the high cost of newer GPUs, which keeps A100s attractive for cost-sensitive workloads. The market for August 2026 is likely to see continued gradual decline, but the pace is uncertain. Key factors include the adoption of NVIDIA's Blackwell architecture, which could further reduce A100 demand, and the potential for a slowdown in AI investment. Additionally, the resolution threshold (X) is not specified in the market description, but it will be set at a level that creates a meaningful bet. Given current trends, a threshold around $1.00-$1.50 might be plausible, but traders will need to assess the trajectory.
Frequently Asked Questions
What is the A100 GPU?
The A100 is a data center GPU from NVIDIA, based on the Ampere architecture, released in 2020. It is designed for AI training and inference, scientific computing, and other high-performance workloads. It comes in 40 GB and 80 GB memory variants.
How is the average compute price calculated?
The average price is the arithmetic mean of hourly values for A100 compute per hour, as reported by Ornn, for the entire month of August 2026. The 'USD' iteration of the index is used, and values are rounded to two decimal places.
What is Ornn?
Ornn is a company that publishes price indices for cloud compute resources, including GPUs. They aggregate pricing data from various cloud providers and marketplaces to produce a daily or hourly average price. Their indices are used by analysts and for prediction markets.
Educational content is AI-generated and sourced from Wikipedia. It should not be considered financial advice.
