If you pay for Anypoint and the vCore line on the order form no longer matches what Runtime Manager is actually using, this page is for you. The job is to right-size workers, not to guess a list price Salesforce does not publish.
We are an Official Salesforce Consulting Partner. MuleSoft work sits with that practice. This is not the homepage offer (one supervised agent on a support, sales, or document workflow). If that is the pain, book the assessment and say so. If the pain is a vCore bill, book the same 30 minutes and bring the artefacts below.
vCores did not vanish in 2024
New customers since March 2024 buy Integration Starter or Integration Advanced, metered on Mule Flows and Mule Messages, not on vCores. Existing vCore contracts were not end-of-lifed. Most people who still open this URL are on those contracts. Read your order form before you argue with a blog that assumes the new SKU.
Salesforce does not publish a public vCore list price. Third-party 2026 ranges disagree with each other, from the high teens into the mid tens of thousands of US dollars per runtime vCore per year. Treat that as order-of-magnitude only. Do not budget from it. The unit that matters is the production runtime vCore on your order form.
Non-production capacity is usually cheaper than production on that form. CloudHub hosting is typically inside the runtime line rather than a surprise on another page. Unused workers still bill. That last sentence is where the money is.
What “unused” actually looks like
Idle cost is almost never a missing discount. It is workers you provisioned and then did not take back.
Oversized non-production: a full replica of production for a QA environment that runs a nightly test. Apps that never take traffic and still have a worker. Applications sized at 1.0 vCore that would run on 0.2 with the same SLA. High availability left on because it was the default in a template, on an app that is not in the HA set. Production sized for a peak week in November and left there all year.
None of those needs a new licence conversation. They need utilisation, from Anypoint Monitoring or Runtime Manager, read against the contract.
Cutting vCores without Salesforce on the call, and without utilisation, is how you shrink a worker that is actually hot and then buy the emergency add-on. We will not do that.
What to bring to 30 minutes
The order form, including the vCore (or flow and message) lines and the dates. A week of utilisation for the workers you care about, from Anypoint Monitoring or Runtime Manager. The list of applications you believe are production, and the ones you are unsure about.
We read utilisation against the contract and tell you which workers to shrink, merge, or leave. Leave is a valid answer. A quiet worker that is the HA pair for a payment API is not waste.
If the contract has already moved to flows and messages, the same 30 minutes is still useful: you are then arguing about message volume and unused apps, not vCores. Bring the form anyway.
Start with 30 minutes. Say it is MuleSoft. Related platform work lives on Salesforce and MuleSoft.