The AI Costs page provides cost visibility for AI services across your cloud providers. It surfaces token-level costs broken down by service, model, and token type, giving you a unified view of AI spending across multi-cloud environments and multiple LLM models.
Access AI Costs
- Go to one.newrelic.com > All capabilities > Cloud Cost Intelligence.
- Click AI Costs in the left navigation.
Cloud provider filters
To filter AI cost data by cloud provider, select an option from the provider tabs at the top of the page:
- All: All cloud providers
- AWS: Amazon Web Services only
- Azure: Microsoft Azure only
- GCP: Google Cloud Platform only
AI top spends
The AI Top Spends section surfaces your highest AI cost areas across four dimensions. Each widget shows the total for the selected period, the comparison value from the previous period, the percentage change, and a breakdown of the top three items.
Services
Displays total AI costs grouped by cloud AI service (for example, AmazonBedrock, Azure AI Services, AmazonSageMaker). Use this to identify which services are driving the most spend across your cloud providers.
Models
Displays total AI costs grouped by LLM model (for example, Claude Opus 4.6, Claude Sonnet 4.6). Use this to compare cost across model versions and providers.
Token Cost
Displays total AI spend broken down by token type:
- Cache Write Token: Cost for writing tokens to cache
- Input Token: Cost for input tokens processed by the model
- Cache Read Token: Cost for reading tokens from cache
Use this to understand which token operations contribute most to your AI costs.
Token Usage
Displays total token consumption (in billions) broken down by token type. Use this alongside Token Cost to identify high-volume, low-cost patterns versus low-volume, high-cost patterns that may indicate optimization opportunities.
Sugerencia
Compare the Token Cost and Token Usage widgets together to understand the cost-per-token efficiency across different token types and models.
Cost Trends
The Cost Trends section provides time-based analysis of your AI costs with interactive charts and controls.
Chart controls
- View options: Switch between Hourly, Daily, Weekly, or Monthly views
- Group by: Select dimensions to group cost data (up to 3 selections)
- View by: Choose a visualization format (availability depends on your group-by selections)
Understanding the data
- Total cost: Aggregate AI spending for the selected period with percentage change versus the previous period
- Average cost (Daily): Daily average AI spending with percentage change versus the previous period
- Stacked bar chart: Visualizes AI cost distribution over time, color-coded by cloud provider
- Output token: Cost for output tokens generated by the model in response to prompts
Cost data table
Below the chart, a table displays AI cost data broken down by cloud provider:
Column | Description |
|---|---|
Cloud provider | The cloud provider (for example, AWS Marketplace, Azure, AWS, Google Cloud) |
Total AI Cost | Aggregate AI spending for the selected period (sortable) |
% of Spend | Percentage of total AI spending attributed to this provider |
Compared Cost | AI cost for the equivalent previous period |
Trend | Percentage change versus the previous period, color-coded (green for decrease, red for increase) |
Visualization options
With a single group-by selection:
- Bar chart: Displays AI cost data as vertical bars
- Line chart: Shows AI cost trends over time
- Tree chart: Hierarchical visualization of AI cost data
- Data table: Tabular view of AI cost data
With multiple group-by selections:
- Heat map: Color-coded visualization across dimensions
- Data table: Tabular view of AI cost data
Filters and customization
Date range selector
Use the date range selector at the top right to:
- Choose a predefined time period
- Set custom start and end dates
- View costs in your preferred timezone (UTC displayed by default)
Engineering costs filter
If your organization uses cost allocation tags, filter by the Engineering costs tag to narrow AI cost analysis to specific teams, projects, or cost centers tagged in your cloud environment.