Primavera Cloud vs P6 EPPM: Which Licensing Model Costs Less at Scale

We compare the newer Primavera Cloud subscription against on-prem P6 EPPM perpetual licensing plus support, and we quantify where each model wins. We also name the 2026 migration pressure Oracle is applying and tell you exactly how to price the break-even before you commit.

The two models you are actually choosing between

This is not cloud versus on-prem in the abstract. It is two different licensing economics, and the crossover point depends almost entirely on user count, deployment lifespan, and how disciplined your provisioning is. On-prem P6 EPPM uses a perpetual license (a large fixed cost paid once) plus annual support at 22 percent of net license, repriced upward every year. Primavera Cloud (Oracle Primavera Cloud, or OPC) is a per-user, per-month subscription with support bundled in and no separate maintenance line.

The published reference numbers frame the gap. P6 EPPM perpetual sits at roughly $2,750 per named user with $605 annual maintenance (per Redress Compliance 2026 data). P6 Professional perpetual is about $2,570 per named user with $550 per year in maintenance, which is the standard 22 percent ratio. On the subscription side, P6 EPPM Cloud Service is quoted at $125 per user per month, and full Primavera Cloud implementations typically start around $100 per user per month with minimum-seat rules. Progress Reporter Cloud, the lightweight timesheet tool, runs about $12 per user per month.

Before you trust any of these, understand the transparency problem. Oracle no longer publishes current public pricing for Primavera Cloud. The tiers that circulate publicly derive from the Construction and Engineering Global Price List dated November 10, 2016, and are reference only. Actual pricing varies by reseller, region, and contract size. Treat every figure below as a modeling anchor, not a quote. For the wider portfolio context, see our Oracle Construction and Engineering licensing overview.

Oracle stopped publishing Primavera Cloud pricing. If your business case relies on a 2016 price list, you are negotiating blind.

The break-even math, run honestly

The single most important number is how many years the perpetual license amortizes before subscription overtakes it. Take a mid-size P6 EPPM estate of 100 named users. Below is the naive comparison most vendors will show you, followed by the version that survives contact with reality.

The pattern is clear and repeats across every estate we have modeled. On-prem P6 EPPM is more expensive in the first two years because of the upfront license, then it pulls ahead permanently once the perpetual cost is sunk. Subscription pricing only wins when the deployment is short-lived, the user count is small and volatile, or your provisioning discipline is so poor that you are paying support on named seats you never use. At scale and over a five-year horizon, perpetual plus support is cheaper by a wide margin, and the margin grows with every support escalation you avoid.

Two adjustments make the comparison honest. First, add support escalation to the on-prem line: Oracle's median renewal uplift runs 6.0 percent a year (per Oracle Licensing Experts, March 2026), above the 4 percent most buyers assume, and where no cap was negotiated (roughly 44 percent of estates reviewed) Oracle applies its standard 8.0 percent. Second, apply escalation to the cloud line too, because Cloud subscriptions reset at renewal and Oracle can reprice annually. The subscription is not a fixed rate for the life of the contract.

Where the subscription hides its costs

Primavera Cloud looks clean because support is bundled and there is no separate maintenance invoice. But four structural features inflate the effective per-user rate.

• Minimum seat floor. You must buy a minimum of five user licenses per module. Small teams and pilot projects pay for capacity they do not use. A three-person scheduling cell still funds five seats per module.
• Module fragmentation. OPC splits into four license types: Portfolio and Capital Planning, Progress, Schedule, and Task Management. A power scheduler who also does portfolio work needs multiple licenses unless you buy the Enterprise model, which grants everyone the Portfolio Planning license automatically and therefore inflates spend across the whole population.
• License-model coupling. Under the Project model all users automatically receive Progress, Schedule, and Task Management licenses; under the Enterprise model they also get Portfolio Planning. These 'convenience' models remove the ability to right-size per person, which is where the real savings sit.
• Renewal repricing. Unlike a perpetual license, the subscription has no floor. When the term ends, Oracle can raise the per-user rate, and any negotiated protection you built into an on-prem contract does not carry over.

That last point is the one buyers consistently miss. In 2026 Oracle is eliminating pricing protections that on-prem customers relied on. Customers migrating from on-prem to Cloud often assume their negotiated caps and discounts transfer. They do not. Cloud subscription contracts reset pricing at renewal. If you migrate to escape a support escalation, you may simply be trading a capped 6 percent problem for an uncapped renewal problem.

Subscription pricing has no floor. The rate you sign is the best rate you will ever see, and renewal only moves one direction.

The 2026 migration pressure, decoded

Oracle's 2026 sales strategy is built around OCI growth, targeting 46 to 50 percent cloud revenue growth year over year (per Redress Compliance, February 2026). To hit that, sales teams are pressuring on-prem customers to migrate, using two levers: support-lifecycle positioning and inflated migration business cases. Oracle markets Primavera P6 Migration to the Cloud as 'your last upgrade ever,' with Oracle Soar handling the automated move. That framing is a sales device, not a technical necessity.

The coercion mechanics matter because they change your negotiating posture. Perpetual licenses do not expire, but without active support you cannot upgrade to newer versions, and Oracle reserves the right to request a migration and charge backdated support fees. If support lapses and you later need to reinstate it, or an audit finds you running software without current support, Oracle may charge back support plus a 150 percent penalty on the missed fees. That penalty is the real stick behind the 'last upgrade ever' carrot. Understand it before you let support lapse as a cost-saving move.

There is also a hidden database migration cost. Customers running P6 on SQL Server must migrate their environment to an Oracle database as part of the cloud upgrade. That is a project cost, a re-platforming risk, and a potential new licensing exposure that never appears in Oracle's migration business case. Factor it in as a hard number before you accept the cloud TCO Oracle presents.

The traps that wreck both models

Whichever model you land on, several licensing traps apply. These are where audit findings and cost overruns come from, and they are largely independent of the cloud versus on-prem decision.

• Provisioning inflation (on-prem). Every user defined in the P6 system counts toward your obligation, including inactive, test, and dormant admin accounts. An estate that provisioned 150 users but actively uses 80 still owes 150 named-user licenses. This is the most common self-inflicted overspend and the easiest to fix before renewal.
• Indirect access. If your ERP, analytics platform, or data warehouse pulls P6 data and surfaces it to employees, those employees may need P6 licensing even though they never touch the P6 interface. This is a top audit finding and applies to both models.
• Reporting-tool licensing. P6 licenses cover scheduling and running standard reports through Oracle Analytics Publisher. Creating new or custom reports requires a separate full-use Analytics Publisher license. Custom dashboards are a frequent unlicensed-use finding.
• Restricted-use repricing. When an audit finds a restricted component used outside its intended context, Oracle's position is the use was never licensed, and the remedy is full-use licensing at list for the entire deployment, often with back support. See our analysis of the restricted-use Oracle Database under P6 EPPM.
• External contractor counting. Contractors and joint-venture partners accessing your P6 environment usually require their own licenses. This scales badly on large capital projects. Our guide on counting external contractors in Primavera deployments covers the exposure.

The CSI repricing clause you must understand before you drop anything

On-prem support contracts include a repricing clause tied to the Customer Service Identifier (CSI). Dropping any line within a CSI triggers repricing of the remaining lines at original list price, not the discounted net you have been paying. This is the trap that catches buyers trying to trim support by dropping unused modules. You save on the dropped line and lose far more on the surviving lines when they snap back to list.

The practical implication: never terminate a partial support line without first modeling the CSI-wide repricing impact. In many estates the repricing math means dropping a $10,000 line costs you $40,000 in lost discount elsewhere. If you want to reduce support cost, the cleaner routes are re-baselining your license count downward (removing dormant seats before renewal) or evaluating third-party support. See our decision framework on moving Primavera and Unifier to third-party support.

What the buyer should actually do

Run the decision in this order. It keeps you from accepting Oracle's framing and forces the numbers to lead.

• Baseline your real usage first. Count actively used seats, not provisioned seats. If you are paying support on 150 named users and use 80, your on-prem support cost is inflated by roughly 47 percent before you compare anything. Clean this up before renewal, because it changes the break-even.
• Model five years, not three. Run on-prem (license sunk plus escalating support at 6 percent and 8 percent) against cloud (per-user subscription plus renewal repricing). At 100-plus users over five years, perpetual plus support wins in almost every case we have modeled.
• Price the hidden migration costs. Add SQL Server to Oracle Database migration, re-platforming, and lost pricing protections to any cloud business case Oracle presents. These routinely add 15 to 30 percent to the true first-year cloud cost in our experience.
• Negotiate a support cap in writing. If you stay on-prem, cap the annual uplift at renewal. Roughly 44 percent of estates never negotiated one and pay 8 percent. A capped 3 to 4 percent uplift compounds into six figures of avoided cost over a five-year term.
• Do not let support lapse as a tactic. The 150 percent reinstatement penalty and backdated support exposure make lapsing far more expensive than it looks. If you want to exit Oracle support, exit deliberately through third-party support, not by omission.
• Prepare for the audit that follows a migration decision. Oracle audits intensify when a customer resists migration. Understand what the auditor examines across your Primavera and Unifier estate before you push back. Our overview of Oracle audit scope across the Primavera estate maps the exposure.

The headline conclusion: at scale and over a normal deployment lifespan, on-prem P6 EPPM perpetual plus disciplined support management is the cheaper model, often materially so. Primavera Cloud wins only for small, short-lived, or volatile user populations, or where your on-prem provisioning is so undisciplined that you are effectively paying subscription-level waste already. The 2026 migration pressure is a revenue strategy, not a cost argument in your favor. Make Oracle prove the TCO with your numbers, cap what you can, and keep the perpetual license you have already paid for unless the modeled break-even genuinely favors the move.

Related advisory: the full topic guide.

Frequently asked questions

Is Primavera Cloud cheaper than P6 EPPM at scale?

Rarely, once you model five years. On-prem P6 EPPM carries a high upfront perpetual cost (about $2,750 per named user) but then only 22 percent annual support, which sinks below cumulative subscription cost after roughly two to three years. At 100-plus users over five years, perpetual plus support is typically cheaper than $125 per user per month, even after 6 to 8 percent support escalation.

Does my negotiated on-prem discount carry over to Primavera Cloud?

No. Oracle Cloud subscription contracts reset pricing at renewal and do not inherit on-prem pricing protections. In 2026 Oracle is explicitly eliminating protections on-prem customers relied on, so a migration can trade a capped support increase for an uncapped renewal repricing. Never assume your discount transfers.

What is the 150 percent support reinstatement penalty?

If Oracle support lapses and you later need to reinstate it, or an audit finds you running software without current support, Oracle may charge the backdated support fees plus a 150 percent penalty on the missed amount. This makes deliberately lapsing support to save money far more expensive than it appears, and it is a key lever behind Oracle's migration pressure.

Why does dropping one support line increase my remaining costs?

Oracle support contracts include a CSI repricing clause. Dropping any line within a Customer Service Identifier triggers repricing of the surviving lines at original list price rather than your discounted net. Buyers who trim unused modules often lose more in restored list pricing than they save on the dropped line, so model the CSI-wide impact first.

Do external contractors need Primavera licenses?

Usually yes. Contractors and joint-venture partners who access your P6 environment generally require their own named-user or subscription licenses, and this scales quickly on large capital projects. Every provisioned account counts toward your obligation regardless of activity, so contractor access is a frequent source of unbudgeted cost and audit findings.

What hidden costs appear in an Oracle cloud migration business case?

Two big ones are routinely omitted. Customers on SQL Server must migrate to an Oracle database as part of the cloud move, a real project and re-platforming cost. And the loss of negotiated pricing protections at renewal exposes you to uncapped repricing. Both should be added to Oracle's TCO before you compare it to staying on-prem.