This blog was authored by: Mike Gladders, Portfolio Lead | 20 May 2025

In our blog series focusing on HPE VM Essentials (VME), I’ve so far established in parts 1 and 2 that HPE VME is a viable alternative to Broadcom.  The key takeaway is that the features of VME hit a sweet spot for customers who only need vSphere Enterprise features but are being pushed toward paying for the full VCF suite.

If there is feature parity with Broadcom (some elements are still on the roadmap, as covered in part 2), then the next questions is: what about pricing?

The first important thing to cover is that HPE VME is back to good old socket-based licensing!

Socket-based Licensing!

One of the biggest impacts of the recent Broadcom licensing upheaval was the change to a per-core licensing model, with the introduction of a 16-core minimum per physical processor.  This caused a huge impact on those businesses with several smaller 4-core, 8-core or 10-core hosts.  Where they had been paying for a single socket for those hosts, they were now facing paying for a minimum of 16-core per socket, even though they weren’t using 16-cores of computing power.

At the other end of the spectrum, it also affected those organisations who had implemented a strategy around purchasing fewer physical hosts but with high core-per-socket densities—64 cores per socket for Intel or 128 for AMD (AMD now do a 192-core per socket CPU, but that wasn’t available pre-Broadcom changes).  On the old per-socket model, these organisations would have been paying for 1, 2, or maybe 4 sockets per host. Under Broadcom’s new model, they face subscription costs based on 64 or 128 cores per socket—a massive, unplanned increase.

Thank you, HPE, for seeing this considerable business impact and understanding that customers need the flexibility to be able to decide on core density that best fits their use cases, without the constraints of additional financial uplifts.

What difference does this make to pricing up a solution, then?

Let’s model a few different scenarios.  We won’t provide actual pricing here as there are a variety of different discounts customers might get through their HPE Channel Partners or Broadcom Partners, but we’ll take some average discounts from real world examples and explain the difference in terms of X.

So, let’s take a small one and just do the minimum billable configuration for a Broadcom host of 1 socket with 16 cores.  This is because even if you use fewer than 16 cores, Broadcom will still charge you for a minimum of 16.  There is no minimum core count for HPE VME, but we’ll use this as a starting point.

Just a quick note here – there was some information published through tech journalists suggesting Broadcom will now be implementing a minimum order of 72-cores. So, a 16-core order as per the example above wouldn’t actually be the minimum. But for the basis of a comparison, we’ll continue to use it.  Just bear in mind that this might not be a valid Broadcom order on its own.

Update: Since this 72-core minimum order was reported on 10^th April, nothing official has materialised, and the consensus is that this has now been scrapped, if it was ever planned.

At this 16-core configuration the pricing comparison is as follows:

So, we can see that VME, even at list price against discounted Broadcom pricing, is coming out significantly cheaper for a 1 socket,16 core deployment.  With both VME and Broadcom, you could purchase cheaper physical servers with lower core counts, but this will be the minimum that you pay for either hypervisor licensing.

Let’s move up to a 72-core minimum Broadcom purchase then:  3 hosts with 1 socket and 24 cores per socket, totalling 72-cores and meeting Broadcom’s 16-core minimum per socket and potential 72-core minimum per order.

So, it is pretty clear that the Broadcom pricing model just does not work well for smaller deployments.  SMBs are going to really struggle to justify Broadcom pricing.

Does it get better for larger deployments?

Let’s see…  We’ll ramp it up to 12 hosts, 2 sockets per host and 32 cores per socket.  So we’re getting fairly large now at 24 sockets and 1,536 cores.

Ah! So the Broadcom pricing doesn’t scale to larger sized environments either.  The disparity is getting even greater, even though the Broadcom minimums aren’t coming in to play now.  This is where the VME socket-based licensing becomes really powerful.  As the core count goes higher under the Broadcom , the cost continues to rise, but under the VME model, higher cores per socket make no difference to the pricing.

Let’s go a bit extreme now just to really drive home the point (sorry Broadcom).  Let’s say a customer has a use case for some HPE DL580 Gen11s with new 5th Generation AMD EPYC processors with 160 cores per socket and 2 sockets per server.  They have deployed 20 for a high-performance virtualisation environment – 6400 cores.

Ouch! Just for clarity: 50x is the difference between £20,000 and £1,000,000!

So, the closest Broadcom can come to VME pricing is the original 16-core comparison that we did.  This is the best-case scenario for Broadcom pricing.  I’ll drop the table again below.

A consistent pattern in the pricing comparison is that as core counts per socket increase beyond 16, Broadcom pricing rises – while VME remains flat. And even when core counts drop below 16, Broadcom’s 16-core minimum means the price doesn’t fall. VME still charges only a flat rate per socket.

That is pretty conclusive then, right?

So, there we go.  Obviously, cost is just one influencing factor when researching and designing a new virtualisation platform, but for most organisations it holds significant weight in the overall decision-making process.  From that perspective, HPE VME becomes an extremely compelling proposition.  At a time when CFOs are seeing rising costs for existing Broadcom estates, proposals showing significant savings are sure to stand out.

The balance against cost is obviously the features and capabilities that we discussed in part 2 of this blog series.  We felt that HPE VME needs to bring the 3^rd party integrations for Data Protection and Disaster Recovery to really be production ready, but those features are being actively worked on and are expected to be released this year.  Now is the time to start those initial discussions and proof of concepts so you can be in a confident position when your Broadcom support or subscriptions are next due for renewal.

Then there are other things to consider like:

• Existing skills and knowledge within teams and any additional training required
• The effectiveness of HPE support for VME, especially when compared against VMware support
• Existing relationships with vendors and partners
• Hybrid multi-cloud strategies
• Existing or planned Storage Area Networks and Storage Arrays
• Application modernisation journeys
• Sustainability targets

This is where Telefónica Tech UK&I excels. Come and talk to us about your cloud strategy, your virtualisation roadmap, and HPE VME. We’re always happy to share our insight and help customers navigate the next step in their journey.