FY26 Cloud Cost Review — Optimization Opportunities by Org
This is the document at the center of Redundant's opening. Rob submitted it on Monday as "Cloud Cost Status, FY26 Month One." By Wednesday it had a new title, a distribution list of eleven vice presidents, and a communications-guidance attachment from the Office of Chief Counsel. The word low-utility was engineered to be precise enough for Finance and dull enough for Legal. It traveled anyway.
Internal Memorandum
Cloud Cost Status, FY26 Month One
FY26 Health Logistics Cloud Cost Review — Optimization Opportunities by Org
[Document retitled by Office of the CTO prior to distribution. Original filename retained in revision history.]
| From | R. Coleman, Director, Cloud FinOps |
|---|---|
| To | CTO Staff (distribution: 11 VPs, Chief Counsel) |
| Date | FY26 · Month One [precise date redacted] |
| Re | Consolidated cloud run rate; low-utility findings by org |
1. Summary
Consolidated cloud spend is running at $480M annualized, up 30% year over year against business growth of 9%. Absent intervention, current commitments project a further 55% increase next fiscal year. This trajectory has been reviewed with Finance and is not considered survivable under the current plan.
2. Methodology — the “low-utility” flag
Spend is flagged low-utility where the underlying system shows no measurable request-path activity — internal or external — over the trailing 90 days, as measured by the owning team’s own observability stack, at thresholds set by the owning team’s senior staff.
Language note (approved by Finance, Nov.; revised Jan. following developer-environment feedback): low-utility does not mean waste, dead, or useless. It means: by every technical signal we can measure, this system is not doing meaningful work. It does not mean we know what it is for. It does not mean cut it.
3. Findings by org (extract)
| Org / Platform | Monthly run rate | Flagged low-utility |
|---|---|---|
| Helios (platform refresh) | $4.1M | $2.0M / mo — development environments, no promotion to production in 6 mo. |
| Mercato (“The Future of Always-On Sales”) | $1.4M | ≈ $1.0M / yr — priority-service instance classes with no live allocation window in trailing 90 days |
| Atlas (trace & custody) | ||
[Duplicate-platform analysis — Mercato/Atlas overlap, modeled over six months — moved to Appendix B. Appendix B not present in recovered copy.]
4. Items requiring caution
Several flagged categories are low-utility by signal but load-bearing by function: exception review queues, cold-chain monitors, chain-of-custody approval desks. These systems register near-zero activity precisely because they exist for the exception, not the routine. FinOps recommends operational review before any action is taken on these line items. The true waste and the safety work currently share a spreadsheet; the flag cannot tell them apart. People can.