Cloud/AI Governance/AI Spend
AI Spend
FinOps for your AI operation: the estimated token cost of every MCP server, the savings your policies produced, the cost per request and the return on your governance, in dollars.
AI Spend is the FinOps screen of AI Governance: it turns token consumption into dollars and policy enforcement into savings you can point at. The console's own pitch: "Find the spend you did not know you had. Pinpoint the servers costing you the most, measure your policy savings, and know exactly where to cut next."
AI Spend
Find the spend you did not know you had. Pinpoint the servers costing you the most, measure your policy savings, and know exactly where to cut next.
Cost & Savings Trend
Hourly estimated spend and savings over timeCost Distribution
Spend proportion by MCP serverCost by MCP Server
Token consumption in USD per server| MCP Server | Tokens | Cost | Savings | Requests |
|---|---|---|---|---|
| GitHub | 842,000 | $2.53 | $0.53 | 834 |
| Slack | 512,000 | $1.54 | $0.27 | 612 |
| Jira | 448,000 | $1.34 | $0.35 | 538 |
| Notion | 376,000 | $1.13 | $0.19 | 471 |
| Linear | 265,000 | $0.7950 | $0.1140 | 322 |
| Stripe | 214,000 | $0.6420 | $0.2130 | 604 |
| HubSpot | 118,000 | $0.3540 | $0.1230 | 538 |
| Vercel | 65,000 | $0.1950 | $0.0690 | 311 |
The mockup runs on the console's Sample Data. On a paid plan the same screen prices your live traffic, per organization.
The four counters
- Estimated Cost in dollars, with the tokens consumed behind it. Amber, on purpose: this is the number to bring down.
- FinOps Savings in dollars, with the tokens saved behind it. This is money your policies put back: every oversized payload the FinOps layer truncated before it reached the upstream API is tokens you never paid for.
- Cost per Request, your true unit economics: the spend divided by the requests that produced it.
- FinOps ROI, the return of your governance in one percentage: savings over cost, with the DLP and FinOps enforcement event counts that produced it right under the bar.
Where the money goes
Cost Distribution splits the period's spend by MCP server, and the Cost by MCP Server panel ranks them with a split bar per server: amber is what you spent, emerald is what your policies gave back on that same server. The heaviest bar is not a verdict; it is a decision waiting to be made. If a Connector's cost outruns its value, this bar is where you see it first, and the FinOps Guard policies in the Settings group are where you act on it.
Cost & Savings Trend plots both over time, hourly for Today and 3 days, daily for longer periods. The amber line is your spend; the emerald line riding underneath is your policies quietly paying part of the bill. The gap between them is the ROI chart in motion.
Server Cost Detail
The table closes the screen with one row per Connector: tokens consumed, cost in amber, savings in emerald, and requests, so cost per call is a glance away. The full console table also carries the Saved token count and the DLP and FinOps enforcement events per server, tying every dollar back to a policy action. Rows are clickable and open that server's cost detail; the empty state is the console's own: "No cost data recorded in this period."
Why this screen exists
Most teams discover their AI costs on an invoice. Here the spend is attributed per server, per period, in the same place where the enforcement that controls it lives. Mission Control's Cost vs Savings chart is the summary of this screen; when a number there raises an eyebrow, the Details link lands you here to find the server, the size of the bill and the policy savings already working on it.