Running IBM and Red Hat as separate negotiations left a median 20 percent on the table
The acquisition put two licensing models on one purchase order and left the buyer side running two negotiations. The renewal motion is one conversation, and treating it as two is worth a fifth of the outcome.
Prepared by Redress Compliance · August 17, 2026 · IBM and Red Hat advisory.
Executive summary
Separate negotiations left a median 20 percent on the table. Combined renewals consistently outperformed, through consolidated benchmarking, dual model negotiation, entitlement scoping, and co terminus packaging.
Red Hat subscription counts drifted as estates virtualised, because they were sized on socket pairs and virtual cores that moved underneath them. That drift inflated the renewal by 15 to 35 percent before any negotiation.
Two licensing models now sit on one purchase order. IBM prices in PVU and VPC against entitlement pools. Red Hat prices subscriptions per socket pair or per virtual core. Neither model reads the other.
Co terminus packaging is the structural move. Aligning the two renewal dates converts two annual conversations with separate leverage into one with combined leverage.
Two models, one purchase order
The acquisition consolidated the commercial relationship without consolidating the licensing logic. Both models persist, and they meter completely differently.
| Dimension | IBM side | Red Hat side |
|---|---|---|
| Metric | PVU and VPC against entitlement pools | Subscriptions per socket pair or virtual core |
| Sub capacity | Conditional on ILMT | Governed by the subscription model chosen |
| Drift behaviour | Pool consumption against retired products | Counts sized on hardware that virtualised underneath them |
| Renewal date | Set by the ELA or Passport Advantage cycle | Set independently unless co termed |
Neither model reads the other, and that is the opening. A buyer running two separate negotiations presents each vendor side with an isolated view of its own position, which is exactly the view that suits the seller. A combined position lets you benchmark consolidated spend, trade across both models, scope Cloud Pak entitlement against Red Hat coverage, and package the renewals co terminously. Those four moves together produced a median 20 percent recovery.
One purchase order, and the buyer still runs two negotiations
The acquisition changed who you buy from without changing how the products are licensed. IBM continues to price in PVU and VPC against entitlement pools; Red Hat continues to price subscriptions per socket pair or per virtual core. Both now arrive on one purchase order, and most buyers continue to negotiate them as though they were separate relationships, because internally they usually still are: different teams, different renewal dates, different technical owners. Across the consolidated renewals we have run, that separation cost a median 20 percent.
The recovery does not come from a single trick. It comes from four things that only become available once the position is combined: consolidated benchmarking, which prices the total relationship rather than two fragments; dual model negotiation, which lets a concession on one metric be traded against the other; Cloud Pak entitlement scoping against Red Hat coverage, which stops the same capability being paid for twice; and co terminus renewal packaging, which converts two annual conversations with separate leverage into one with combined leverage. None of the four exists while the negotiations run apart.
Underneath the commercial question sits a measurement one specific to the Red Hat side. Subscription counts were sized on socket pairs and virtual cores at a point in time, and estates virtualised underneath them. The count did not move because nothing forces it to, so the renewal prices a hardware shape that no longer exists, inflating it by 15 to 35 percent before any negotiation opens. That is the same class of problem as IBM pool drift, arriving through a different mechanism, and it is correctable with the same discipline of measuring before renewing.
The practical sequence follows from the structure. Establish both positions independently and accurately first, because a combined negotiation built on two inaccurate counts simply combines the errors. Then consolidate: one benchmark, one team, one calendar. The co term is worth pursuing even at a small cost, because it is what makes every subsequent renewal a single conversation rather than a recurring pair. The Red Hat subscription mechanics sit in the Red Hat cost brief, the IBM posture in the vendor management playbook, and the library in the IBM practice.
- 520 vendor benchmarks, from IBM ELA to Red Hat and OpenShift
- Both models reconciled against real deployment before the renewal opens
- A defensible position paper generated in minutes, not weeks
The four moves that combine the position
- Consolidated benchmarking, pricing the total relationship rather than two fragments that each look modest on their own.
- Dual model negotiation, so a concession on PVU or VPC can be traded against Red Hat subscription terms and vice versa.
- Cloud Pak entitlement scoping against Red Hat coverage, which stops the same capability being paid for on both sides of the purchase order.
- Co terminus renewal packaging, converting two annual conversations with separate leverage into one with combined leverage.
- Resize Red Hat subscription counts against current virtualisation, since counts sized on an older hardware shape inflate the renewal 15 to 35 percent.
- Fix both positions before combining them, because a combined negotiation on two inaccurate counts just combines the errors.
What the combined renewals showed
Consolidated renewals run across financial services, manufacturing, and public sector produced a consistent pattern:
On the combined renewal, through consolidated benchmarking, dual model negotiation, entitlement scoping, and co terminus packaging.
Renewal inflation where Red Hat subscription counts were sized on socket pairs and virtual cores that changed as the estate virtualised.
Two licensing models sit on one purchase order and neither reads the other. A buyer running separate negotiations presents each side with an isolated view of its own position, which is the view that suits the seller.
The renewal motion is one conversation. Treating it as two is worth roughly a fifth of the outcome, and the four moves that recover it only become available once the position is combined.
Your first five moves
- Establish both positions independently and accurately, since combining two inaccurate counts produces a combined error rather than leverage.
- Resize Red Hat subscriptions against current virtualisation, because counts sized on an older hardware shape carry 15 to 35 percent of inflation.
- Scope Cloud Pak entitlement against Red Hat coverage so the same capability is not paid for on both sides of the order.
- Benchmark the total relationship, not the two fragments, and negotiate across both models in one conversation.
- Pursue the co term even at a small cost, since it makes every future renewal a single negotiation. The IBM practice runs the combined position with you.
Frequently asked questions
What did the acquisition actually change?
Who you buy from, not how the products are licensed. IBM still prices in PVU and VPC against entitlement pools and Red Hat still prices subscriptions per socket pair or virtual core. Both now arrive on one purchase order.
Why do separate negotiations cost so much?
They cost a median 20 percent, because four recovery moves only exist once the position is combined: consolidated benchmarking, dual model negotiation, entitlement scoping across both, and co terminus packaging. None is available while the negotiations run apart.
Why do Red Hat counts drift?
They were sized on socket pairs and virtual cores at a point in time, and estates virtualised underneath them. Nothing forces the count to move, so the renewal prices a hardware shape that no longer exists, inflating it 15 to 35 percent.
What is co terminus packaging worth?
It converts two annual conversations with separate leverage into one with combined leverage, permanently. That makes it worth pursuing even at a small cost, because the benefit recurs at every subsequent renewal rather than once.
Can concessions be traded across the two models?
Only in a combined negotiation. Dual model negotiation is precisely the ability to trade a PVU or VPC concession against Red Hat subscription terms, and it is unavailable when each side sees only its own isolated position.
What is entitlement scoping across both?
Checking Cloud Pak entitlement against Red Hat coverage so the same capability is not paid for on both sides of the purchase order. Overlap is easy to create and invisible from either side alone.
Should we fix the counts before combining?
Yes. A combined negotiation built on two inaccurate counts combines the errors rather than the leverage. Establish both positions independently and accurately, then consolidate.
Is the 20 percent typical or best case?
Median across consolidated renewals run in financial services, manufacturing, and public sector. It is the middle of the distribution rather than the top of it, which makes it a reasonable planning assumption.
Does one purchase order mean one negotiation?
It should, and usually it does not. Most buyers keep two internal teams, two renewal dates, and two technical owners, which reproduces the separation the vendor benefits from even after the commercial relationship has consolidated.
Which side drifts more?
Both drift, through different mechanisms. IBM pools drift as products retire while entitlement stays assigned; Red Hat counts drift as the hardware they were sized against virtualises. The discipline is the same: measure before renewing.
