NVIDIA A100 · Hourly price on Jul 24 (Week 30)

Week 30: Jul 20, 2026 - Jul 24, 2026

$21.9K Vol.Closes Jul 24, 2026

Above $0.50$314.2 Vol.99%

Above $0.75$610.3 Vol.96%

Above $1.00$1.2K Vol.70%

Above $1.25$2.4K Vol.15%

Above $1.50$13K Vol.2%

Above $2.00$3.2K Vol.2%

Above $1.75$1.2K Vol.1%

View on Kalshi

At a glance

What

The hourly rental cost for NVIDIA A100 compute in the cloud market is currently fluctuating between deep spot prices near $0.33/hr and standard on-demand rates above $1.00/hr.

Who

Specialized cloud providers (e.g., PrimeIntellect, Vast.ai, Spheron) and decentralized platforms are offering significantly lower rates than traditional hyperscalers like AWS and GCP.

Context

The A100 80GB GPU market has seen a sustained price decline in mid-2026, with floor prices dropping 17% in a single month amid increased availability.

Stakes

If the hourly price remains above $0.50, it indicates that standard on-demand or reliable spot pricing is still commanding a premium over the absolute floor, impacting AI training and inference budgets.

Note: The "hourly price" in the resolution rule likely refers to the spot or effective market rate on Jul 24 rather than the highest on-demand sticker price; if the spot market dominates usage on that date, the price could fall below $0.50 given the current $0.33/hr floor.

Jul 20

The NVIDIA A100 spot price has fallen to $0.33/hr, marking a 17% drop from the previous month and approaching its 6-month low of $0.33–$5.77/hr.

gpufinder.dev

Jul 20

The cheapest confirmed A100 80GB instance in stock is $1.29/hr on PrimeIntellect, while on-demand pricing across 12 providers ranges from $0.40/hr to $32.97/hr.

gpufinder.dev

Jul 20

Spheron lists the cheapest A100 80GB on-demand rate at $1.07/hr with a spot price of $0.60/hr, while Vast.ai marketplace pricing dips to $0.67/hr for high-reliability hosts.

spheron.network

Now

AI-generated briefing. AI can make mistakes. This is not financial advice.

Bull case

The median price across 92 configurations is $1.07/GPU/hour, which is well above the $0.50 threshold, suggesting that most standard market transactions remain high.
gpus.io
Reputable on-demand providers like Lambda, FluidStack, and CoreWeave list A100 80GB rates between $1.21/hr and $1.62/hr, establishing a reliable floor significantly above $0.50.
gridstackhub.ai
Hyperscalers and enterprise-focused providers (AWS, GCP, Azure) continue to charge normalized rates of $3–$5/hr per GPU, maintaining a high average market price for reliable compute.
gridstackhub.ai

AI-generated briefing. AI can make mistakes. This is not financial advice.

Bear case

The market floor for spot instances has dropped to $0.13–$0.33/hr, with the cheapest confirmed A100 near its 6-month low, threatening to pull the hourly average below $0.50 if spot demand surges.
gpufinder.dev getdeploying.com
Vast.ai and Spheron offer spot/marketplace pricing as low as $0.60–$0.67/hr, creating a competitive pressure that could drive the effective hourly price down toward the $0.50 level for budget-conscious users.
gpufinder.dev gridstackhub.ai

Industries

NVIDIA A100 · Hourly price on Jul 24 (Week 30)

Week 30: Jul 20, 2026 - Jul 24, 2026

$21.9K Vol.Closes Jul 24, 2026

Above $0.50$314.2 Vol.99%

Above $0.75$610.3 Vol.96%

Above $1.00$1.2K Vol.70%

Above $1.25$2.4K Vol.15%

Above $1.50$13K Vol.2%

Above $2.00$3.2K Vol.2%

Above $1.75$1.2K Vol.1%

View on Kalshi

At a glance

What

The hourly rental cost for NVIDIA A100 compute in the cloud market is currently fluctuating between deep spot prices near $0.33/hr and standard on-demand rates above $1.00/hr.

Who

Specialized cloud providers (e.g., PrimeIntellect, Vast.ai, Spheron) and decentralized platforms are offering significantly lower rates than traditional hyperscalers like AWS and GCP.

Context

The A100 80GB GPU market has seen a sustained price decline in mid-2026, with floor prices dropping 17% in a single month amid increased availability.

Stakes

Jul 20

The NVIDIA A100 spot price has fallen to $0.33/hr, marking a 17% drop from the previous month and approaching its 6-month low of $0.33–$5.77/hr.

gpufinder.dev

Jul 20

The cheapest confirmed A100 80GB instance in stock is $1.29/hr on PrimeIntellect, while on-demand pricing across 12 providers ranges from $0.40/hr to $32.97/hr.

gpufinder.dev

Jul 20

Spheron lists the cheapest A100 80GB on-demand rate at $1.07/hr with a spot price of $0.60/hr, while Vast.ai marketplace pricing dips to $0.67/hr for high-reliability hosts.

spheron.network

Now

AI-generated briefing. AI can make mistakes. This is not financial advice.

Bull case

The median price across 92 configurations is $1.07/GPU/hour, which is well above the $0.50 threshold, suggesting that most standard market transactions remain high.
gpus.io
Reputable on-demand providers like Lambda, FluidStack, and CoreWeave list A100 80GB rates between $1.21/hr and $1.62/hr, establishing a reliable floor significantly above $0.50.
gridstackhub.ai
Hyperscalers and enterprise-focused providers (AWS, GCP, Azure) continue to charge normalized rates of $3–$5/hr per GPU, maintaining a high average market price for reliable compute.
gridstackhub.ai

AI-generated briefing. AI can make mistakes. This is not financial advice.

Bear case

The market floor for spot instances has dropped to $0.13–$0.33/hr, with the cheapest confirmed A100 near its 6-month low, threatening to pull the hourly average below $0.50 if spot demand surges.
gpufinder.dev getdeploying.com
Vast.ai and Spheron offer spot/marketplace pricing as low as $0.60–$0.67/hr, creating a competitive pressure that could drive the effective hourly price down toward the $0.50 level for budget-conscious users.
gpufinder.dev gridstackhub.ai