Mainframe · Calendar

The mainframe MIPS capacity squeeze nobody saw coming.

The seller team running the mainframe portfolio this quarter is sitting on capacity growth data they did not have at the last renewal cycle. The data is now the lead slide. Buyers without their own capacity view are walking into a conversation already shaped.

By The Negotiation Desk · Mainframe · Archetype: The Calendar

Published 25 Aug 2024

The Broadcom mainframe account teams are running a different motion in 2026 than they ran in 2023. The change is not in the contract. It is not in the unit of measure. It is in the data the seller is bringing to the opening meeting. The post 2024 capacity telemetry has produced a multi year view of buyer side MIPS consumption that did not exist at the prior renewal. The view shows that most enterprise mainframe estates have grown capacity quietly across the last three years, sometimes by 8 to 22 percent against the contracted band, and the seller is now using that growth as the opening of the renewal conversation. Buyers without a parallel view of their own capacity history are walking into the opening meeting at a disadvantage they did not have in 2023.

This article is a Calendar piece, written on live deals this quarter. It describes what the Broadcom mainframe team is actually doing in their opening meetings right now, what the data they are presenting actually shows, and what the buyer side preparation needs to look like to be in a different conversation. If the buyer's next mainframe renewal is between now and the third quarter of 2027, the work in this article is the work that needs to be done in the next 90 days.

What the seller is opening with

The opening pack now includes a three year capacity trend. The trend is built from MSU consumption data the seller pulls from the post 2024 mainframe telemetry, normalised against the in force contract band, and presented as a chart in the first ten slides of the renewal pack. The chart shows two lines. The contracted capacity, flat across the term. The consumed capacity, sloping upward. The gap is the squeeze. The conversation that follows is built around what to do about the gap.

The seller's proposed answer to the gap is a renewal at a higher band, with a forward growth allowance baked in, plus a true up on the prior period where the consumed capacity exceeded the contracted capacity. The true up is the part most buyers are not prepared to defend, because the contracted band language often includes peak measurement provisions that the seller can reach for if the buyer has not read them. The forward growth allowance is the part that compounds, because the higher band carries forward into every subsequent renewal.

What the data actually shows

The data is real. Most enterprise mainframe estates have grown capacity quietly across the last three years. The growth is driven by workload consolidation, new application deployment on mainframe footprints that had been deprioritised, and in some sectors a re shoring of workloads that had been moved to distributed platforms and then returned. The growth is not uniform. Some quarters are higher than others. Some workloads are seasonal. Some of the apparent growth in the seller's chart is noise that disappears when the data is read at the right granularity.

"The chart in their pack showed twenty two percent growth over three years. The same data read against rolling quarterly averages and stripped of two known seasonal peaks showed eleven percent. The remaining eleven was still a conversation. It was a different conversation."Mainframe Lead, The Desk

The buyer side preparation

The buyer's preparation has three parts. The first is the buyer's own capacity history, pulled from the buyer's mainframe operations team and normalised against the buyer's contracted band. The seller's chart should not be the first capacity chart the buyer sees on the buyer's own estate. The second is the contract language on the unit of measure, the peak measurement provisions, and any softening or capping provisions that may apply to growth above the contracted band. The third is an operational reality statement that documents the drivers of the consumption growth, the seasonal patterns, and the workloads that are intended to be migrated off the mainframe in the renewal period.

Each of those three preparations needs to be in place before the opening meeting. The seller's chart is the seller's argument. The buyer needs the equivalent on the buyer's side. Without it, the opening meeting closes with the seller's reading as the only reading on the table, and the renewal conversation that follows is anchored to that reading.

The four available restructure paths

Four restructure paths have produced outcomes on live mainframe deals in the last 18 months. The first is a band amendment that explicitly excludes seasonal peaks from the consumption measurement. The second is a forward commit that prices growth at a lower marginal rate than the contracted band, in exchange for a longer term. The third is a workload exit commitment that scopes a portion of the consumed capacity for migration off the mainframe within the renewal period, with the contracted band priced against the net of the exit. The fourth is a hybrid IPLA and MSU restructure that moves a subset of the licensed software to a different unit of measure, reducing the total exposure to the consumed MIPS metric.

Of the four, the band amendment is the cleanest where the consumption growth has a documented seasonal component. The workload exit commitment is the cleanest where the buyer has a credible migration plan. The forward commit and the hybrid restructure are the cleanest where neither of the first two paths is operationally available.

Average capacity growth seller is presenting+8% to +22% over 3 years

Growth that survives buyer side normalisation+4% to +14%

True up exposure on opening renewal pack$3M to $18M

Settled exposure where buyer brought a parallel view$0 to $4M

What we have seen this quarter

An Asia Pacific bank brought us a mainframe renewal in the first quarter of 2026. The seller's opening pack showed a 19 percent capacity growth over three years against a contracted band, with a proposed renewal at a higher band plus a $12M true up on the prior period. The buyer's own capacity history, normalised against rolling quarterly averages and stripped of two known seasonal peaks tied to regulatory reporting cycles, showed an 8 percent growth. The settled renewal closed at a band amendment that excluded the seasonal peaks, no back true up, and a forward commit that priced growth at a lower marginal rate. The total economic value of the outcome across the term was approximately $18M against the opening seller position.

The takeaway

The mainframe renewal motion in 2026 opens with a capacity trend chart the seller did not have in 2023. The chart is the seller's argument. Buyers without a parallel view of their own capacity history are walking into a conversation already shaped.
The data is real but the reading is negotiable. Seasonal peaks, workload patterns, and contract language on peak measurement provisions all soften the seller's chart. The buyer side normalisation often halves the growth number that survives to the closing position.
Four restructure paths produce outcomes. Band amendment, forward commit, workload exit commitment, and hybrid IPLA restructure. The choice depends on which path is operationally available, not which one is theoretically cleaner.