Weekly Pricing Pulse

About Pulse

The Weekly Pricing Pulse tracks what changed in published GPU cloud pricing across AWS, Azure, GCP, OCI, CoreWeave, Lambda, Vultr, Nebius, and Crusoe over the last 7 days. AIForge Works compares each new harvest with the prior week and highlights buyer-relevant movement: comparable on-demand price changes, New Silicon, restored / expanded visibility, Deprecated SKUs, and likely price/SKU rolls.

This is not a live quote, workload-specific recommendation, or personalized account analysis. For workload-specific recommendations, use Cloud Advisor. To inspect neighboring SKUs, regions, or commitment terms, open the Multi-Cloud Pricing Analyzer. To compare your own GPU spend against a comparable floor, use GPU Audit.

Editor's note

The last seven days produced no confirmed public on-demand GPU price move above 1% across the nine providers we index. Comparable public on-demand GPU catalog listings had a small increase of 18, from 13,796 to 13,814. The four board floors were unchanged again this week: H100 at $3.85 per GPU-hour, H200 at $4.29, B200 at $6.69, and A100 SXM4 at $2.30.

AWS supplied the week's clearest catalog-expansion signal. Its eight-GPU p6-b300.48xlarge appeared in Seoul for the first time in our reviewed history, represented by Linux and four licensed operating-system variants. That is one B300 instance family appearing in a new catalog market, not five separate launches. We now track publicly listed on-demand B300 prices from three providers: AWS, OCI, and Nebius. CoreWeave also shows B300 on its public pricing page, but its on-demand rate remains contact-sales. As always, catalog visibility does not prove orderability or available capacity.

The remaining activity was mostly catalog reshuffling: 22 previously seen listings returned to view, while nine were no longer visible. For buyers, the practical takeaway is straightforward: this week's changes were about catalog visibility, not a market-wide price reset.

Ornn OCPI context

OCPI is the Ornn Cloud Price Index, a transaction-based benchmark for observed GPU cloud pricing that AIForge Works uses as market context alongside public provider catalog floors. Week over week, OCPI rose for H100 from $2.64 to $2.87, B200 from $5.93 to $6.64, and A100 SXM4 from $1.02 to $1.07, while H200 declined from $4.75 to $4.42 per GPU-hour. This Pulse uses the August 24, 2026 harvest for public catalog floors and the latest reviewed Ornn x ICE OCPI reading available for comparison, through August 23, 2026 20:00 UTC. It compares transaction-based OCPI readings with AIForge Works indexed Public Catalog floors for H100 SXM, H200, B200, and A100 SXM4.

Week at a glance

No provider had comparable on-demand price movement over the 1% reporting threshold. AWS, Lambda, Nebius, and Vultr changed catalog visibility through New Silicon, restored / expanded visibility, and Deprecated SKU entries, shown below.

These movement counts apply only to comparable public-catalog on-demand prices. OCPI is a separate transaction-based market index and is shown independently.

• On-demand pricing was quiet this week. We did not find any public on-demand GPU price changes above the 1% reporting threshold; the visible movement is catalog visibility.

Provider roll-up

No buyer-visible GPU catalog additions, removals, or public on-demand price moves above 1% were found this week for Azure, GCP, OCI, CoreWeave, and Crusoe.

Underlying provider feeds also shifted: Azure (-14,768 catalog records), GCP (+28 catalog records), Nebius (+16 catalog records). After normalizing to comparable GPU offers, purchase terms, and verified hardware specs, those feed-level changes did not alter the public on-demand GPU offers reported in this Pulse.

Offers seen in earlier verified harvests returned to view this week: AWS (6), Lambda (7), Nebius (8), and Vultr (1). Because these provider and SKU combinations appeared in earlier verified harvests, we treat them as restored catalog visibility rather than brand-new silicon launches.

Catalog changes this week

Catalog change counts summarize week-over-week visibility changes in the compared harvests. They show provider + region + instance/SKU observations that became newly visible or no longer visible between harvests. We do not describe these as confirmed capacity launches or retirements unless provider-published sources say so.

New Silicon

5 New Silicon GPU catalog entries were detected across AWS (5). Entries are grouped by provider so high-volume catalog churn does not hide smaller-provider changes.

New Silicon means a provider + region + instance/SKU observation newly visible in this harvest that is not paired with a same-provider, same-geo, same-instance deprecated SKU. It still does not by itself prove net-new capacity or a brand-new accelerator family.

Restored / expanded visibility

22 previously seen public GPU catalog observations were visible again this week across AWS (6), Lambda (7), Nebius (8), Vultr (1).

Each provider and SKU had appeared in an earlier verified comparison; the region shown is where the offer was visible this week. We classify these as restored or expanded catalog visibility, not a new hardware launch, proof of added capacity, or confirmation that the offer is currently orderable.

Deprecated SKUs

9 Deprecated SKU GPU catalog entries were detected across Lambda (5), Vultr (4). Entries are grouped by provider so high-volume catalog churn does not hide smaller-provider changes.

Deprecated SKU means the exact provider catalog SKU was present in the prior harvest but absent from this one. It is no longer visible as that exact SKU in this week's public-catalog comparison, even when a same-instance replacement looks likely; this does not prove the SKU is unavailable to buy.

Top on-demand movers by provider

No comparable offers met the 1% movement threshold for this section.

Up to 5 price observations per provider, ranked by confidence (high-confidence first), then by percentage change.

Methodology

Evidence through. Pricing evidence in this issue runs through August 24, 2026. This evidence-through date is recorded separately from the publication date shown in the article header.

What we compared. This Pulse compares last week's harvest (August 17, 2026) against this week's (August 24, 2026). Both harvests run automatically against the same 9 cloud providers.

How we compare prices fairly. Different providers package GPUs differently: some sell single-GPU instances, others sell 8-GPU instances. To compare apples-to-apples, we divide every instance price by the number of GPUs on that instance. All movement percentages in this Pulse are based on that per-GPU hourly rate, not the raw instance price.

What counts as a real move. We only flag price changes greater than 1%. Anything smaller is treated as rounding noise and excluded. New Silicon means a provider + region + instance/SKU observation newly visible in this harvest that is not paired with a same-provider, same-geo, same-instance deprecated SKU and has not been seen in the lookback window. Restored / expanded visibility means the provider and instance/SKU appeared in an earlier verified harvest and is visible again, sometimes in additional regions. Deprecated SKU means the exact provider catalog SKU was present in the prior harvest but absent from this one.

Scope and limits. The generated Pulse diff certifies on-demand movement only. All prices are in US dollars. Catalog entries missing a clean price or verified GPU specs are dropped before comparison. This Pulse compiles publicly published prices and is not a live quote, availability guarantee, or workload-specific advice.

This week's coverage. The August 24, 2026 harvest completed across 9 providers. After normalizing 6,280,662 raw catalog price observations against verified hardware specs, this Pulse identified 13,814 comparable public on-demand GPU observations for week-over-week movement analysis.

Next Pulse: Week of 2026-08-31.