PE Portfolio Multi-Tenant Oracle Consolidation: $14.2M of Run-Rate Spend Removed

The Brief: One Portfolio, Nine Oracle Estates

The fund's operating partner had inherited a portfolio where every Oracle relationship had been left to the portfolio company CIO. Nine portfolio companies. Six industries. Seven distinct Oracle account teams. Three live Oracle ULAs, four Java SE Universal Subscriptions on different metrics, and two unresolved LMS audit exposures. Cumulative run-rate Oracle spend across the portfolio was $34M a year. Oracle's account teams were treating each portfolio company as a separate sales target, which meant every renewal was being negotiated in isolation — and Oracle was winning at headline pricing in seven of nine. The portfolio CFO brought us in to design and run a multi-tenant Oracle consolidation that would survive Oracle's predictable counter-moves.

Oracle's playbook against portfolio consolidation is well understood by anyone who has run it from the sales side. They will argue that aggregation is a "change of use" that triggers re-licensing. They will argue that any shared-services architecture amounts to multi-tenant deployment requiring an Oracle Cloud-only commercial structure. They will threaten audits across every portfolio company simultaneously. The buyer-side defence depends on getting the contractual architecture right before Oracle's account teams realise consolidation is in motion. We did the consolidation under a confidentiality wrapper that prevented coordinated Oracle counter-moves until the contract changes were already signed.

Forensic Position Assessment Across Nine Companies

The first eight weeks were a forensic, evidence-based inventory of every Oracle Order Form, every OMA, every Schedule, every Java SE Universal Subscription, and every active support stream across the portfolio. We catalogued the Processor Metric, NUP, Employee Metric and Universal Credits commitments at each portfolio company. We mapped contractual restrictions on entity transfer, change-of-control, and shared-services use. We benchmarked the actual deployment against the licensed entitlement at each company. The audit exposure was concentrated in three companies — predictably, the three that had been carved out of larger groups and had inherited Oracle contracts without ever benchmarking the entitlement. Total compliance gap across the portfolio: $4.8M of potential back-licence claim. We dealt with that before Oracle did.

Our Approach: Five Workstreams in Parallel

1. Master Customer Entity at the Fund Level

   Oracle's standard OMA treats every signing entity as a separate customer. We negotiated a master customer relationship at the fund/holding level with documented portfolio company affiliate definitions, then re-papered each portfolio company's Oracle agreements to inherit from the master. This protected the portfolio from Oracle's "change-of-control" playbook on add-on acquisitions and gave us a single contractual surface to negotiate from at every renewal.

2. Java SE Universal Subscription Right-Sizing

   Three portfolio companies were paying for Java SE Universal Subscription against an Employee Metric that included contractors, board members and acquired-but-not-integrated employees. One was paying twice — once on the legacy Java SE subscription and once on the Employee Metric, after Oracle's rep had failed to terminate the original contract. We rebuilt the count, terminated the duplicate, and pushed back on the Employee Metric definition with documented evidence. Two companies were migrated to OpenJDK on a controlled, evidence-based exit; one stayed on Oracle Java SE at the right-sized count.

3. ULA Co-Termination and Selective Exit

   The three live ULAs were on staggered three-year terms with different exit windows. We co-terminated two, certified them out at exit with forensic deployment counts, and re-papered the third as a smaller, more defensible Term licence position. Oracle's account team argued for portfolio-wide PULA expansion. We declined, with evidence: PULA pricing assumed perpetual growth that the portfolio's actual usage trajectory did not support.

4. Support Re-Pricing and Third-Party Support

   Two portfolio companies running stable, frozen Oracle Database versions were moved to third-party support, removing 100% of Oracle's 22% annual support stream on those workloads. A third was kept on Oracle support but had its support footprint right-sized after we identified $1.1M of support fees being paid on licences that had been retired in a prior data centre exit but never removed from the Oracle support contract.

5. Audit Exposure Closed Before It Was Surfaced

   The $4.8M compliance gap was closed quietly — through evidence-based right-sizing, deployment changes, and selective re-licensing — before Oracle's LMS team had reason to issue a notification. The result was zero formal audit, zero back-licence claim, and a defensible Oracle position across all nine portfolio companies.

The Results

By month 11, the portfolio's Oracle spend had moved from $34M a year to $19.8M, with a single contractual surface, an evidence-based affiliate definition that protected future add-on acquisitions, and zero formal audit exposure. Two portfolio companies had been exited to third-party support; two had migrated Java to OpenJDK; three had certified out of ULAs. The remaining Oracle workloads were running on right-sized, defensible licence positions with documented evidence behind every count.

Key Takeaways for Private Equity Sponsors