AI spend
What the AI services in your clouds cost, straight from each provider's billing API and mapped to the same catalog as every other lane. Billing needs no agent and no log: if it costs money, it is in use. Last pull 1 h ago.
Last 30 days
$12,356
+25% vs the 30 days before
Previous 30 days
$9,912
same services, one month earlier
Annual run rate
$150,327
last 30 days, times twelve
Providers
3
Google Cloud · AWS · Azure
Trend
Daily AI spend, 90 days
06-1609-13
| Month | AWS | Azure | Google Cloud | Total |
|---|---|---|---|---|
| 2026-06 | $1,496 | $821 | $1,456 | $3,772 |
| 2026-07 | $3,460 | $2,241 | $3,303 | $9,004 |
| 2026-08 | $3,976 | $3,172 | $4,309 | $11,458 |
| 2026-09 | $1,909 | $1,723 | $2,229 | $5,861 |
Partial months at both ends of the window are shown as pulled.
By provider
Last 30 days
- Google Cloud · +30%4692
- AWS · +13%4124
- Azure · +33%3540
Worth a look
- Azure OpenAI on Azure grew 54% month over month to 2,665 USD. Find the owner before the next invoice does.
- Gemini API (AI Studio) on Google Cloud grew 310% month over month to 965 USD. Find the owner before the next invoice does.
- 965 USD went to the Gemini API outside Vertex AI last month. Those projects are billed like any other, but they sit outside VPC Service Controls and CMEK.
By service
13 billed AI services
| Service | Provider | Billed as | Last 30 days | Change | Share | Accounts | Evidence |
|---|---|---|---|---|---|---|---|
| Vertex AI | Google Cloud | Vertex AI | $2,827 | +17% | 23% | 2 projects | Open finding → |
| Azure OpenAI | Azure | Azure OpenAI | $2,665 | +54% | 22% | 2 subscriptions | Open finding → |
| Amazon Bedrock | AWS | Amazon Bedrock | $2,182 | +29% | 18% | 123456789012 | Open finding → |
| Amazon SageMaker | AWS | Amazon SageMaker | $1,443 | -6% | 12% | 123456789012 | Open finding → |
| Gemini API (AI Studio) | Google Cloud | Generative Language API | $965 | +310% | 8% | frank-side-project | Open finding → |
| Vertex AI Search and Agent Builder | Google Cloud | Discovery Engine | $560 | -5% | 5% | contoso-knowledge | Open finding → |
| Azure AI Services | Azure | Cognitive Services | $484 | -7% | 4% | Contoso Production | Open finding → |
| Azure Machine Learning | Azure | Azure Machine Learning | $391 | -2% | 3% | Contoso Data | Open finding → |
| Amazon Q | AWS | Amazon Q | $259 | +3% | 2% | 123456789012 | Open finding → |
| Google Cloud AI APIs | Google Cloud | Document AI | $199 | -8% | 2% | contoso-finance | Open finding → |
| AWS AI Services | AWS | Amazon Comprehend | $146 | 0% | 1% | 123456789012 | Open finding → |
| Dialogflow | Google Cloud | Dialogflow | $141 | -7% | 1% | contoso-support | Open finding → |
| AWS AI Services | AWS | Amazon Textract | $93.99 | +169% | 1% | 123456789012 | Open finding → |
Sources
Where the numbers come from
- AWS (demo)AWSok408 rows · 1 h ago
- Contoso (via Microsoft 365)Azureok331 rows · 1 h ago
- Google Cloud (demo)Google Cloudok505 rows · 1 h ago
AWS reads Cost Explorer (ce:GetCostAndUsage) grouped by service. Azure reads Cost Management per subscription (Cost Management Reader). Google Cloud reads the standard billing export in BigQuery, set on the Google Cloud connection. Only services with billing names in the catalog are kept.