The trade is ownership against currency: a perpetual licence plus annual Subscription and Support survives a lapse — you keep running, unsupported but licensed — while a term subscription bundles support in, costs less to enter, and ends outright at expiry. IBM is deciding part of this for you, withdrawing perpetual options product by product and selling the catalogue subscription-first; the buyer’s job is no longer choosing a side so much as pricing the exit positions, timing the transition, and refusing to let a forced migration arrive as a silent repricing.
Published 9 March 2026 · Last reviewed 9 March 2026
For decades the default IBM transaction was a perpetual licence with an annual Subscription and Support contract on top: pay once for the right to run, pay roughly a fifth again each year for upgrades, fixes and someone to call. The subscription licence inverts that — one recurring fee, commonly on 12-, 24- or 36-month terms, with S&S folded in for the duration and the right to run ending when the term does.
What makes this comparison live in 2026 is that IBM is steadily removing the first option. IBM i licensing on newer Power systems moved to subscription-only terms; perpetual licences for certain operating-system tiers were withdrawn effective January 2026; Cloud Paks, SaaS editions and much of the distributed catalogue are sold subscription-first. Entitlements you already hold are not revoked by any of this — but the option to buy more perpetual capacity is disappearing product by product, and estate planning has to treat that as a moving boundary, verified per product, not a constant.
The decision therefore splits in two. Where both models are still offered, it is a genuine economic choice. Where they are not, the work shifts to negotiating the terms of an inevitable transition — which is a different exercise with different leverage, and the one buyers are least prepared for.
This guide compares licensing programs as published by the vendor; it is general information, not legal or licensing advice for your situation, and it names no firms. Program availability changes product by product — verify current terms against your own Passport Advantage agreement and product announcements. The firm directory lists IBM-capable advisors with balanced pros and cons, listed, not ranked.
| DIMENSION | PERPETUAL + S&S | SUBSCRIPTION |
|---|---|---|
| Right to run | Indefinite; survives an S&S lapse on the version you hold | For the term only; ends at expiry, no residual right |
| Support & upgrades | Via annual S&S renewal, a separate line per product | Included for the term; no separate S&S ledger |
| Cost shape | Capital-heavy entry, then recurring S&S with compounding annual uplift | Lower entry, full repricing exposure at every renewal |
| Lapse consequence | Run-as-is, unsupported; reinstatement is penal but the licence holds | Immediate unlicensed deployment; no run-as-is position exists |
| Capacity flexibility | Ratchets up; shrinking the estate strands paid entitlement | Can track down at renewal boundaries as well as up |
| Availability in 2026 | Narrowing; withdrawn for some products and platforms | The default sales motion across the catalogue |
| Where leverage sits | With the holder between renewals; S&S is the vendor’s annuity to protect | With the vendor at each term end, rising with workload criticality |
The lapse row is the one that prices everything else. The perpetual model carries a genuine option — the right to stop paying and keep running — and options have value even when never exercised, because the other side must negotiate as if you might. The subscription model deletes that option, which is precisely why the transition tends to arrive wrapped in incentives. What the incentive buys back is your walk-away position; treat it as a sale, not a gift.
Under perpetual, the recurring negotiation is S&S. The annual uplift compounds quietly, and the renewal is where caps, multi-year locks and re-harvesting of unused entitlement get won or lost. A lapse decision belongs in this conversation too: deliberately parking a static workload on lapsed support is a legitimate position with known costs — reinstatement charges chief among them — and it reads very differently from a lapse that simply happened.
Under subscription, the renewal is the whole game. Everything reprices at once, and recent renewal cycles have made “modernisation” a familiar wrapper for double-digit unit increases. The countermeasures are structural: start a year out, hold consumption evidence per product, negotiate price protection into the current term rather than hoping at the next one, and never let the term simply expire into a quote. This is the standing work of a renewal and contract negotiation engagement.
The transition itself is a negotiation event — the biggest of the three. Trade-up credits for surrendered entitlements, the treatment of remaining perpetual rights, co-term structure, caps on the first renewal after migration: all of it is settable exactly once, at signature. Audit findings have a way of arriving near these moments and resolving into subscriptions — the dynamics the Cloud Pak conversion guide dissects apply to the ownership question in full, and when the finding comes first, IBM audit defense is the discipline that keeps the settlement honest.
Compliance work survives both models. Perpetual PVU estates carry the ILMT and sub-capacity machinery; subscription estates swap part of that for entitlement-expiry risk and License Service metering on containers. Neither model removes the need for a continuously maintained licence position — they relocate where it breaks.
Perpetual economics still win — where they are still sold — for long-lived, capacity-stable workloads that would pay S&S every year regardless: the core transaction systems, the middleware that outlives three reorganisations. The longer the horizon and the flatter the capacity curve, the harder the subscription has to work to match a licence that amortised years ago. They also fit organisations that genuinely value the lapsed-support fallback, whether as negotiating posture or as the planned end-state for a system being wound down slowly.
Subscription fits volatile and uncertain footprints — workloads that may shrink, move to SaaS or be decommissioned inside the term — where paying only for the years used beats owning capacity you cannot return. It suits estates consolidating into Cloud Paks, where the subscription is the only shape on offer, and buyers who prefer one inclusive fee over administering a ledger of S&S lines. The honest entry in this column: it also fits every product where IBM no longer sells the alternative, which is a fit by fiat rather than analysis.
The test that decides it: for each major product, take your own five-year history of S&S payments and capacity change, then model both structures forward with a realistic renewal uplift on the subscription side. If the workload’s lifespan is shorter than the perpetual breakeven, subscribe. If it is longer and a perpetual option exists, the question becomes what you would pay for the exit option — and usually you already own it.
Letting an S&S lapse happen instead of deciding it. The run-as-is position is valuable when chosen for a static workload and expensive when discovered during an upgrade that now requires reinstatement or repurchase. Inventory which perpetual lines are deliberately parked and which are merely unpaid.
Treating subscription expiry like an S&S lapse. The instincts transfer badly: there is no unsupported-but-licensed state on the other side. Renewal calendars, owner assignment and a hard rule against running past term are the cheap insurance.
Surrendering perpetual entitlements without pricing them. Trade-up offers net the old licences against the new subscription at the vendor’s valuation. Obtain the side-by-side: what the discount is with and without surrender, and what reverts if the subscription is not renewed. The delta is the price of your fallback — decide it knowingly.
Absorbing “modernisation” repricing as if it were inflation. Unit increases dressed as program changes are negotiable exactly once per cycle, before signature. Benchmark against your own history, and make the uplift itself a negotiated term with a cap, not an annual surprise.
Taking transition advice from the parties paid by the transition. Vendor sellers and margin-bearing resellers have economics in the subscription closing. The independence test names the conflict; the fee-models guide explains how to pay an advisor so their incentives point at your outcome instead.
Yes — the perpetual licence survives the lapse. What stops is everything around it: version upgrades, fixes, security patches and the right to open support tickets. Rejoining later typically means reinstatement charges or repurchase rather than simply resuming payments, so a lapse should be a priced decision about a static workload, not a budget accident. A subscription, by contrast, ends outright at term.
Product by product, yes. IBM i licensing on newer Power systems moved to subscription-only terms, perpetual licences for certain operating-system tiers were withdrawn effective January 2026, and much of the distributed catalogue is sold subscription-first. Perpetual entitlements you already own are unaffected by a withdrawal — what disappears is the option to buy more, which is the fact to verify per product before building a plan on it.
It depends on three variables you control and one you do not. Yours: how long the workload genuinely lives, how stable its capacity is, and whether you would have paid S&S every year anyway. IBM’s: the repricing at each subscription renewal. Stable long-lived workloads tend to favour perpetual economics where the option still exists; volatile or shrinking footprints favour subscription. Model both against your own renewal history, and treat published uplift patterns as indicative.
Whatever the order document says. Trade-up transactions commonly surrender perpetual entitlements in exchange for credit against the subscription; side-by-side deals leave them intact but idle. The surrendered fallback is an asset with real value — price it into the deal explicitly, not silently inside a discount percentage.
The mechanics differ more than the magnitude. Perpetual PVU estates are audited on deployment counts and the sub-capacity discipline covered in the full-capacity versus sub-capacity guide. Subscription estates add a sharper failure mode: entitlements that expire, converting a licensed deployment into an unlicensed one immediately. Both models reward a continuously maintained effective license position.
No. This is a directory, not a ranking. Firms are listed alphabetically with balanced pros and cons. Independence is shown as a pro and reseller, Big-Four or vendor-side-audit ties as a con, both stated as factual trade-offs for you to weigh.
Modelling both structures on your own renewal history, pricing the fallback you would surrender, and structuring the transition terms before they structure you — that is precisely the work of an IBM licensing advisory engagement. The IBM hub maps the vendor’s wider licensing world, and the directory lists every firm covering IBM — with balanced pros and cons, listed, not ranked. Choosing between them? Start with how to choose an IBM licensing partner.
The criteria, questions and red flags →
Every firm for this work →
The repricing event, prepared for →
The bundle question next door →
The counting basis underneath →
Every field guide on the site →
Tell us the vendor, the service you need and where things stand, and we will route your brief to firms that genuinely cover that combination. The directory and matching are free for buyers, no vendor ever sees your brief, and we add no markup.
Our weekly dispatch on vendor audit programs, regional developments and one buyer move. Subscribe to The Licensing Radar.