Savings calculations
Baseline methodology, calculation formulas, and pricing logic behind every number in the Savings Report.
This page is the technical reference for the Realized savings report. It covers the baseline methodology (how Cast AI establishes a pre-optimization reference point) and the exact formulas that turn that baseline into projected cost and savings figures, including hibernation savings.
Baseline overview
To say "Cast AI saved you money," Cast AI needs a reference point: what would this cluster have cost, and how would it have looked, without Cast AI? That reference point is the baseline, and it's the foundation every number in the report is measured against.
The baseline isn't a single dollar figure. It's a small model of your cluster's behavior before Cast AI's optimizations took effect, made up of four numbers plus a time window:
- CPU overprovisioning factor: how much more CPU was provisioned than workloads actually requested.
- RAM overprovisioning factor: the same, for memory.
- CPU unit cost: the effective price you paid per CPU core, per hour.
- RAM unit cost: the effective price you paid per GiB of RAM, per hour.
These four numbers are measured over a baseline period, a historical window representing your cluster before Cast AI optimization. Once captured, the model stays fixed and is reapplied going forward to reconstruct what your projected (no-Cast-AI) cost would be on any given day.
NoteWhy a model, and not just an old bill? Clusters grow and shrink over time, so comparing today's bill to a fixed dollar amount from months ago wouldn't be fair. Instead, the baseline captures ratios and rates (how much you tended to overprovision, and at what price) and reapplies those ratios to today's actual workload demand. That keeps the comparison fair as your cluster scales up or down.
What the numbers mean in practice
| Parameter | What it captures | Example |
|---|---|---|
| CPU overprovisioning factor | Average ratio of provisioned CPU to requested CPU during the baseline period | 1.8 means 80% more CPU was sitting on nodes than workloads ever asked for |
| RAM overprovisioning factor | Average ratio of provisioned RAM to requested RAM | 1.4 means 40% more memory was provisioned than requested |
| CPU unit cost | Average cost per CPU core, per hour, during the baseline period | $0.031 per core-hour |
| RAM unit cost | Average cost per GiB of RAM, per hour, during the baseline period | $0.004 per GiB-hour |
These four values are averaged across your entire baseline period and applied uniformly. They don't currently distinguish between on-demand, spot, and fallback capacity.
How your baseline is determined
Cast AI computes baselines automatically, once per day, for every eligible cluster. It tries three methods in order, from most to least specific to your cluster, and uses the first one that has enough data:
1. Your cluster's own history → most reliable
2. Other clusters in your org → fallback
3. Cast AI's broader fleet average → last resortA cluster only moves to the next method if the previous one didn't have enough data to compute a baseline. Once a cluster gets a baseline from an earlier method, it's never re-evaluated against a later one.
Your cluster's own history
This is the highest-quality baseline, because it's built entirely from your cluster's own pre-optimization behavior. Cast AI looks at the period between when your cluster first connected and when Cast AI took over active node management. That takeover point is when more than 20% of your nodes became Cast AI-managed. Cast AI calculates the four parameters from your cluster's actual metrics during that window.
Your cluster qualifies for this method only if there's at least 7 days of history between cluster creation and the point Cast AI took over. Without a week of pre-optimization data, the cluster falls through to method 2.
In the report and API, this baseline type is labeled Cluster history.
Other clusters in your organization
If your cluster doesn't have enough of its own history, Cast AI looks for other clusters in your organization that already have a Cluster history baseline, and averages their overprovisioning factors and unit costs. The baseline window becomes the combined span of those peer clusters' own baseline windows.
The assumption here is that clusters within the same organization tend to share infrastructure patterns and pricing, making a sibling cluster's measured behavior a reasonable stand-in for yours.
In the report and API, this baseline type is labeled Peer clusters.
Industry average
If your cluster has no usable history of its own, and no eligible peers in your organization, Cast AI falls back to averaging overprovisioning and cost across all clusters observed on the fleet. It uses the single day your cluster connected to Cast AI's active management. This is a genuine cross-fleet average, not limited to your organization.
In the report and API, this baseline type is labeled Industry average. Think of it as "no data more specific to this cluster was available for comparison."
NoteAn industry-average baseline is the least specific of the three, since it isn't based on your own workloads or infrastructure. If your Savings Report is using an industry-average baseline and the numbers look off, contact your Cast AI representative about a manual baseline review.
When a baseline can't be computed yet
A cluster needs to be at least 14 days old, and Cast AI needs to be actively managing it, before a baseline is computed at all. A cluster younger than that, or not yet under active management, shows no realized savings. Unless the workload autoscaler manages it, it stays absent from the report entirely rather than showing zero savings. This is intentional: it avoids showing noisy, unstable savings numbers for clusters that just connected.
Once your cluster clears both thresholds, Cast AI computes a baseline on its next daily run, and the cluster starts appearing in your report.
Manual baseline adjustments
Sometimes the modeled baseline doesn't reflect reality. For example, if you have out-of-band knowledge of your actual pre-Cast-AI costs, or the automatic calculation landed on a window that wasn't representative. Two adjustments are available through your Cast AI representative:
- Override the baseline: hand-set any of the four parameters directly. Fields you don't specify keep their current values. Once a baseline is overridden, Cast AI's automatic daily calculation will never touch that cluster again, so the override sticks until it's changed manually.
- Recalculate over a custom date range: re-run one of the three calculation methods above, but over a date range you choose instead of the one Cast AI selected automatically.
Both baseline types are labeled Overridden in the report once applied, so you can always tell when a baseline reflects a manual adjustment rather than the automatic calculation.
Baseline assignment summary
| Your cluster's situation | Baseline you'll get |
|---|---|
| At least 7 days of history before Cast AI took over node management | Cluster history |
| Not enough own history, but your org has another cluster with a Cluster history baseline | Peer clusters |
| No own history and no eligible peers in your org | Industry average |
| Younger than 14 days, or Cast AI hasn't yet taken over management | No baseline yet |
| Cast AI representative manually set the baseline | Overridden (see Manual baseline adjustments) |
Calculation formulas
Actual cost
Actual cost is the straightforward sum of real dollars spent on provisioned CPU and RAM over the period you're viewing, with no modeling involved.
actual cost = actual CPU cost + actual RAM costProjected cost
Projected cost is the modeled estimate of what you'd be paying without Cast AI. It's the number every savings figure is derived from.
projected CPU cost =
if (day is on/after the baseline period ends):
greatest(original requested core-hours, current requested core-hours)
× CPU overprovisioning factor
× greatest(current effective $/core-hour, adjusted baseline $/core-hour)
else:
actual CPU costRAM is calculated the same way, using the RAM overprovisioning factor and RAM unit cost. The RAM baseline unit cost is adjusted to your commitment tier identically.
How the formula works:
- Before your baseline period ends, projected cost equals actual cost. Cast AI hadn't optimized the cluster yet during the baseline window, so there's nothing to attribute savings to. Savings only start accruing once the baseline period closes. See Realized savings.
- Demand is the larger of original or current requested core-hours. If a workload's requests were rightsized down, projection uses the larger, pre-rightsizing original request. If a workload's requests grew above its original level, projection uses the larger current request. Either way, the model never assumes you'd be running leaner than you actually run today.
- Demand is scaled up by the overprovisioning factor, the ratio your cluster historically ran at without Cast AI's node-level optimization.
- The result is priced at whichever is higher: your current effective rate or your adjusted baseline unit cost. This ensures the counterfactual is never underpriced. If Cast AI lowered your unit cost (for example, through spot adoption), the projection reflects the higher pre-optimization rate.
- The baseline rate is adjusted to your current commitment tier before this comparison, so a change in your commitment tier doesn't distort the result (see Commitment tier changes).
Realized savings
realized savings = projected cost − actual costThis is zero for any cluster without a baseline, and zero for any day inside the baseline period, since projected cost equals actual cost during that window (see above).
Workload autoscaler savings
workload autoscaler savings =
(original requested core-hours − current requested core-hours)
× greatest(current effective $/core-hour, adjusted baseline $/core-hour)counted only on days when workload autoscaling (VPA) was actually active on the cluster.
NoteThis calculation runs independently of the baseline period. Unlike realized savings, workload autoscaler savings can accrue even on days that fall inside your baseline window, since rightsizing isn't tied to when Cast AI took over node management.
Hibernation savings
When a cluster hibernates, Cast AI scales its nodes to zero and the cluster incurs no compute cost. The report counts the cost the cluster would have incurred as savings:
hibernation savings = hibernated hours × as-if-running hourly rateHow the formula works:
- Hibernated hours are the hours in a hibernation window. A window opens when the cluster hibernates and closes when it resumes, is archived, or is deleted. A window still open at the end of the report range is counted up to the range end. Hours are attributed per calendar day, so partially hibernated days contribute only their hibernated share.
- The as-if-running rate is the modeled cost of running the cluster for one hour. It's an uptime-weighted average of the daily projected cost over the 30 days ending on the day the cluster hibernated. It uses only data from before hibernation started. Days after the baseline period ended contribute projected cost; days inside it contribute actual cost. The rate uses the same formula as Projected cost.
- Actual cost during hibernation is treated as zero. Hibernated hours contribute their full as-if-running cost as savings.
- On clusters with a baseline, hibernation savings accrue only after the baseline period ends. The gating matches realized savings. Clusters without a baseline reach the report only through workload autoscaler management. They are priced at their actual cost per hour before hibernation, with no baseline scaling or gating.
- Each window's rate is captured as the window happens. Cast AI recomputes it while the cluster remains hibernated, so baseline adjustments take effect on the accruing savings. After the window closes, the rate is fixed.
Reported cost fields
The report combines the CPU and RAM figures above into the fields you see in the UI and API:
| Field | Value |
|---|---|
| Actual cost | Actual CPU cost + actual RAM cost |
| Projected cost | Projected CPU cost + projected RAM cost |
| Autoscaler savings | Autoscaler CPU savings + autoscaler RAM savings |
| Workload autoscaler savings | Workload autoscaler CPU savings + workload autoscaler RAM savings |
| Hibernation savings | Hibernation CPU savings + hibernation RAM savings |
| Total savings | Currently equals autoscaler savings only, since the workload autoscaler impact is already included by decreasing the CPU and RAM demand |
NoteThe console adds hibernation savings into the Baseline spend and Realized savings cards, and into the realized savings values of the savings-over-time chart. The API keeps hibernation savings in a separate field, and its total savings value excludes them.
Listing prices vs. discounted prices
The report can show actual cost and current unit price using either your discounted (effective, negotiated) rates or listing (on-demand list) prices, ignoring any discounts. Toggling this changes what "actual cost" means, but not your baseline:
- Discounted (default): reflects what you actually pay: negotiated rates, real spot prices. To ensure these reflect your actual cloud spend, configure Price adjustments.
- Listing: uses on-demand list prices, useful for showing savings against undiscounted, rack-rate pricing.
Two things to know about this toggle:
- Listing prices aren't available before August 6, 2025. If you select a time range that starts before this date, the report automatically falls back to discounted prices regardless of your selection.
- Your baseline unit cost is always based on discounted pricing, since that's what was available when your baseline was calculated.
- In Discounted mode, that stored baseline rate is rescaled to your current commitment tier (see Commitment tier changes).
- If you switch to listing prices, actual and current costs rise to list-price levels while the baseline floor stays at your discounted historical rate. The "whichever is higher" pricing rule described in Demand and price selection still applies on top of that.
Commitment tier changes
Your baseline unit costs capture the discounted rates in effect during the baseline period, after whatever cloud commitments you already had. Your commitments are, for example, an enterprise agreement, committed-use discounts, Savings Plans, or your configured Price adjustments. If those commitments change afterwards, your commitment tier moves with them, and the stored baseline rate no longer reflects what that capacity would cost you today. Savings figures then misattribute the change: overstated when your discounts deepened, understated when they shrank.
To keep the comparison fair, the report rescales the baseline unit cost to your current commitment tier before using it in any savings calculation:
adjusted baseline unit cost = baseline unit cost × (current discount level ÷ baseline-window discount level)The discount level is the ratio of your effective (discounted) cost to the listing cost of the same capacity, measured per cluster and separately for CPU and RAM. For example, if your effective CPU price is $0.021 per core-hour and the listing price for that same capacity is $0.030, your CPU discount level for that window is 0.7. The current discount level is measured over the report's selected time range. The baseline-window discount level is measured over the cluster's baseline period.
In practice:
- Discounts deepened since the baseline: the baseline rate is adjusted down, so savings no longer credit a price reduction Cast AI didn't cause.
- Discounts shrank since the baseline: the baseline rate is adjusted up accordingly.
- Commitment unchanged: the ratio is 1 and the stored baseline rate is used as-is.
- Manually overridden baselines are never rescaled: the value set by your Cast AI representative is used exactly as provided (see Manual baseline adjustments).
Two boundaries to be aware of:
- The adjustment applies only in the default Discounted pricing mode. In Listing mode the current side is priced at list, and the stored baseline is used without rescaling.
- Discount levels are measured against listing-price data; see Listing prices vs. discounted prices for the data-availability boundary. For a baseline window that falls entirely before listing-price data exists, the baseline-window discount level is treated as 1 (undiscounted).
Historical data quality
A small number of data-quality safeguards apply to historical data, mainly affecting workloads' original (pre-rightsizing) requests:
- Some workload types had a data-collection gap that produced zero values for original requests before Cast AI resolved it in early 2026; the report substitutes current requests as a reasonable proxy for those older data points.
- Values that are implausibly large (well beyond what any real pod would request) are excluded and replaced with a sane fallback, to guard against occasional data-collection errors.
- For time ranges that start before October 2025, the report uses the earliest reliable snapshot of original requests it can find, rather than assuming zero.
These safeguards only affect historical accuracy for older data. They don't change how current and future savings are calculated.
Scenario reference
Per-day savings
| Day is... | Workload autoscaling active? | Realized savings that day | Workload autoscaler savings that day |
|---|---|---|---|
| Inside the baseline period | No | None (projected = actual) | None |
| Inside the baseline period | Yes | None (projected = actual) | Accrues |
| After the baseline period ends | No | Accrues | None |
| After the baseline period ends | Yes | Accrues | Accrues |
Hibernated days
On clusters with a baseline:
| Day is... | Hibernation savings that day |
|---|---|
| Hibernated, inside the baseline period | None |
| Hibernated, after the baseline period ends | Accrues at the as-if-running rate |
| Partly hibernated, after the baseline period ends | Accrues for the hibernated hours only |
A fully hibernated day contributes zero actual cost and zero nodes to the report's averages. No workload autoscaler savings accrue for it, because a hibernated cluster reports no workload metrics.
Demand and price selection
The projected-cost formula always picks the higher option for both demand and price. If Cast AI lowered your unit cost (for example, through spot adoption), the projection reflects the higher pre-optimization rate:
- Demand uses whichever is larger: original (pre-rightsizing) requests or current requests.
- Price uses whichever is higher: your current effective unit cost or your adjusted baseline unit cost.
Known limitations
- Clusters without a baseline are absent, not zeroed. A cluster that's too new, not yet under active management, or that produced no usable data in any of the three baseline methods doesn't appear in either report view. If the workload autoscaler manages the cluster, it appears even without a baseline. Such a cluster can show workload autoscaler and hibernation savings; its realized savings stay zero.
- Clusters created before node-level monitoring became available (February 2025) may have a less precise baseline window. For these clusters, the baseline period ends when Cast AI took over credentials rather than when it took over active node management, which can make the overprovisioning and cost factors slightly less representative.
- An industry-average baseline draws from Cast AI's entire fleet, not just your organization. This is intentional (it's meant to be a genuine cross-fleet reference point), but it means an industry-average baseline for your cluster is influenced by data from unrelated organizations.
- Savings only start after your baseline period ends. A report whose entire time range falls inside the baseline window will show projected cost equal to actual cost and zero realized savings (workload autoscaler savings may still appear, since that calculation isn't tied to the baseline window).
- Baseline factors don't distinguish between on-demand, spot, and fallback capacity. All four baseline parameters are single averaged values applied uniformly across lifecycles, so spot-versus-on-demand nuance isn't captured in the model.
Updated 12 days ago
