MuleSoft licensing, explained for enterprise buyers. Editions, vCores and where your runtimes run.

A core based MuleSoft contract is priced on the edition, the committed vCore pool and the runtime each vCore runs on. Size the pool to measured use, price Runtime Fabric against CloudHub, and cap the escalator before Salesforce widens the deal.

How is MuleSoft licensed in 2026?

MuleSoft is sold on two kinds of contract. Most enterprises renewing today hold a core based subscription, priced on the Anypoint Platform edition, the vCores you commit to, and the runtime mix. Newer contracts are usage based, metered on activity instead of capacity, and the current price list sells that usage as Mule Credits.

The vCore is the unit of capacity. The edition decides which features each vCore can run, and the runtime decides whether it lives in MuleSoft's cloud or on your own infrastructure. Since Salesforce bought MuleSoft in 2018, the renewal also carries cross sell pressure from the wider Salesforce portfolio.

MuleSoft's public pricing page starts packages at $2,000 a month, billed annually. This guide concentrates on the core based subscription, where most renewals still sit. Read it with our Salesforce renewal guide, the Salesforce knowledge hub and our Salesforce advisory practice.

Which model is your contract on?

Open Access Management in Anypoint Platform. A usage based organization shows flows, messages and throughput entitlements there. If you see production, sandbox and design vCores, you are on the core based model. Current price points are in our MuleSoft pricing guide.

What do the Anypoint Platform editions include?

The edition sets the price of every vCore in the contract. Gold covers core integration, Platinum adds API management and monitoring, and Titanium adds advanced governance and the deepest monitoring. Very large deals are sometimes cut as a custom bundle.

Read the order form as well as the tier name. Premium connectors are often itemized as separate lines, and a connector whose integration has been retired can come off at renewal.

Which Titanium features justify the price?

MuleSoft's monitoring documentation reserves these for Titanium, or for Integration Advanced on usage based plans. If your platform team uses none of them, price Platinum.

• Retention. Monitoring data kept for more than 30 days.
• Custom metrics and alerts. Custom metrics, advanced dashboards and alerts on custom dashboards.
• Flow and connector charts. Performance views per flow and per connector.
• Log tools. Log points, raw archived log downloads and log search across applications.
• Export and reports. The Telemetry Exporter for audit logs and traces, Reports, and the Monitoring Archive API.

How do you right size the edition?

1. Score actual feature use. List the Platinum and Titanium features the team used over the last 12 months: custom dashboards and alerts in Anypoint Monitoring, policies in API Manager, and anything the platform team relies on.
2. Find the lowest tier that fits. Drop to Gold where Platinum features sit unused, and to Platinum where no one relies on the Titanium list.
3. Cap the uplift between tiers. Write the price of a later upgrade into the multi year deal.

What is a MuleSoft vCore and how is it counted?

A vCore is one virtual CPU of Mule runtime capacity with a paired memory allocation, and it drives the recurring cost. Five rules decide how many you consume.

1. One vCore equals one vCPU. It pairs with 1.5 GB to 2.0 GB of heap memory. Current documentation gives a 1 vCore worker 2 GB of heap on both CloudHub generations, while older sizing sheets quote 1.5 GB.
2. vCores are sold in packs. Contract lines are commonly quoted in packs of 4, 8, 16 and 32 vCores.
3. vCores are allocated to applications. Each Mule application reserves its worker size at deploy time, busy or idle, and two workers draw twice.
4. Non production has its own pool. Sandbox environments draw from a separate vCore entitlement.
5. The pool does not burst. You cannot deploy beyond the contracted vCores. Workers smaller than 1 vCore can burst CPU for short periods, but only inside the allocation you already pay for, and MuleSoft does not guarantee it.

How are cores counted on your own servers?

Hybrid standalone Mule runtimes are licensed on the maximum allocated cores they consume. The counting rules catch teams out on virtual machines.

• Logical cores count. The count is what the Java Virtual Machine running Mule can see. Physical cores are ignored.
• Every instance counts separately. Each VM or container is metered on its own, even on a shared host.
• Fractions round up. A container limited to 1.5 cores counts as 2.
• No limit means everything counts. A Mule runtime on an unconstrained 16 core VM consumes 16 cores, so pin allocations before usage is measured.

How is Flex Gateway licensed?

Flex Gateway is licensed apart from the Mule runtime vCore pool. It is a lightweight gateway built to replace the heavier Mule runtime for API gateway work, and MuleSoft now markets it as Omni Gateway.

• It runs without a Mule runtime. The gateway deploys at the edge with a very low compute footprint and consumes no Mule runtime vCores.
• It is metered on call volume. On usage based plans, gateway requests are totaled per month. Managed API instances are counted separately at the monthly peak, split into production, preproduction and unclassified.
• Hybrid deployments combine both. Flex Gateway at the edge, Mule runtime carrying the integration logic in the core.

Older core based order forms define their own Flex Gateway count, so read that definition before you forecast a saving.

Which MuleSoft runtime costs least: CloudHub, Runtime Fabric or Standalone?

Standalone on your own servers carries the lowest per vCore cost in the quotes we review, and CloudHub the highest, because CloudHub includes MuleSoft's infrastructure and operations. Runtime Fabric on your own Kubernetes sits between them, with more operating work for your team.

Worked example: repricing a CloudHub heavy contract

Say you commit 48 production vCores on CloudHub, and assume $10,000 per vCore a year for round numbers (an illustration, not a MuleSoft list price). The usage report shows a peak of 36 vCores, four of them 1 vCore API proxy applications that Flex Gateway could replace.

The 16 Runtime Fabric vCores cost $96,000 to $120,000 instead of $160,000. The annual bill falls by $160,000 to $184,000 before gateway and Kubernetes running costs, and resizing the pool does most of that work.

What Salesforce cross sell should you expect at a MuleSoft renewal?

Expect the MuleSoft renewal to arrive with other Salesforce products attached. Three motions come up most.

1. Data Cloud bundle. Data Cloud added to the renewal at what is presented as a discount.
2. Agentforce inclusion. Agentforce credits added to the bundle as a strategic position.
3. CPQ tie in. CPQ proposed as a downstream consumer of the MuleSoft integration.

Why we would keep MuleSoft out of the wider Salesforce bundle

The usual advice is to fold MuleSoft into the broader Salesforce agreement, on the theory that a larger deal earns a larger discount. We disagree. A blended quote hides the per vCore rate, so you cannot benchmark it or price the unused vCores.

It also ties MuleSoft to renewal dates and uplifts set for products your roadmap may not need. Get MuleSoft quoted as separate lines, settle the vCore count and the escalator, and only then decide whether co terming with the CRM contract is worth it.

What will the account team say, and how should you answer?

• "Customers your size all run Titanium." Ask which Titanium features your team used last year, and have Platinum priced alongside.
• "You need headroom for next year's projects." Buy growth when projects are funded, with a price hold on extra vCores at today's rate.
• "Runtime Fabric saves nothing once you count Kubernetes." If your clusters and platform team already exist, that cost is sunk. Ask for both runtimes quoted per vCore side by side and add your own running cost.
• "The usage based model will cost you less." Ask for a conversion quote built from 12 months of your own flows, messages and throughput.

Where can you cut cost at a MuleSoft renewal?

There are nine places, each tied to one line on the invoice. Work through them before the next contract refresh, well ahead of the first quote.

• Edition right sizing. Drop to the lowest tier that fits actual feature use.
• Runtime mix shift. Shift workloads from CloudHub to Runtime Fabric where the Kubernetes platform exists.
• Flex Gateway adoption. Replace heavy Mule runtime gateways with Flex Gateway.
• vCore right sizing. Measure actual vCore consumption and shrink the contracted pool.
• Non production pool. Price the sandbox vCore pool at a deeper discount than production.
• Renewal escalator cap. Put a written ceiling on the multi year escalator.
• Bundle pushback. Decline cross sell bundles that do not fit the integration roadmap.
• Multi year discount. Negotiate a multi year term at a fixed discount.
• Independent audit support. Bring in an independent adviser before any MuleSoft true up.

MuleSoft licenses cleanly when the runtime mix matches your operating model, the vCore pool matches actual consumption, and the renewal escalator stays inside a written cap.

Which contract terms should you ask for?

Ask for terms that keep the vCore count and the unit price flexible over the whole term.

• Escalator cap. A written cap of 3 to 5 percent a year against the prior year's net price, so the discount carries forward.
• Reduction right. Fewer vCores allowed at each anniversary, so a retired project stops costing money.
• Growth price hold. Extra vCores at the current unit rate for the full term.
• Runtime portability. vCores usable across CloudHub, CloudHub 2.0, Runtime Fabric and hybrid standalone, so a runtime shift does not mean buying twice.
• Bundle severability. Separate terms and renewal dates for Data Cloud, Agentforce or CPQ, so dropping one does not reprice MuleSoft.

How do you check your own MuleSoft usage before a renewal?

Start with Usage Reports in Anypoint Platform, then match them against entitlements and worker sizes. MuleSoft sees the same data, so bring your own reading of it.

• Usage Reports. vCore consumption and peak across CloudHub and CloudHub 2.0. It opens only at the root organization and needs the Usage Viewer permission. For Runtime Fabric and hybrid standalone, check the core limits you set against the contract as well.
• Access Management. Production, sandbox and design entitlements, split by business group.
• Runtime Manager. Worker size and count per application. Export one row per application.
• Anypoint Monitoring. CPU and memory per worker, which shows the applications that would run on a smaller size.

What have we seen in recent MuleSoft renewals?

In the MuleSoft renewals we have advised on, the runtime mix produced the largest savings, because so many customers still run the CloudHub heavy setup they bought early. Escalators ran high where no one pushed back, and caps of 3 to 5 percent were negotiable on multi year deals.

Edition and pool sizing need only usage data and a clear ask. Runtime and gateway changes take engineering time, so start them months before the renewal date.

How does Redress help with MuleSoft renewals?

We run MuleSoft renewals through the Vendor Shield subscription, the Renewal Program, the Benchmark Program and the Software Spend Assessment. Each engagement is led by a former Salesforce commercial executive who now works for the customer.

See our benchmarking work and MuleSoft renewal service, read about us, find our locations, or contact us.

What to do next

1. Score feature use against your tier. Check which Anypoint Monitoring and API Manager features the team actually used this year.
2. Inventory the runtime mix. Map every Mule application to CloudHub, Runtime Fabric or Standalone.
3. Price the Runtime Fabric option. Compare per vCore cost across runtimes, including Kubernetes running costs.
4. Identify Flex Gateway candidates. Edge gateways with little transformation logic.
5. Right size the vCore pool. Cut it to measured consumption plus a small buffer.
6. Negotiate the escalator cap. Get it in writing on multi year terms, with a reduction right.
7. Test cross sell against the roadmap. Decline bundles your integration plans do not call for.
8. Line up independent support early. Engage an adviser before the next renewal or true up.

Frequently asked questions