Broadcom isn't raising VMware prices. It's pricing SMBs out.
An eight-core physical host running VMware in your office now licences as 72 cores. Same hardware, same workload, nine times the licence footprint. That's not a typo and it's not a tier change. It's the floor that came in with Broadcom's restructure of the VMware product line, and for a Victorian SMB renewing in the next twelve months it is the single most consequential number on the page.
The shorthand on what Broadcom has done to VMware is "Broadcom raised prices." That is accurate but it misses what is actually happening. Hock Tan's playbook on every prior acquisition — CA Technologies in 2018, Symantec's enterprise business in 2019, now VMware — is the same: simplify the product line, force everyone onto subscription, set the floors high enough that the bottom of the customer pyramid leaves on its own, and concentrate revenue in the Fortune 500. The 72-core minimum and the mandatory bundle are not pricing decisions. They are segmentation decisions, expressed as price.
What actually changed
The full picture, condensed. Perpetual licences for vSphere, vSAN, NSX and Aria stopped selling in February 2024. Existing perpetuals can still run, but once the support contract lapses there are no more patches, no updates, no security advisories. The replacement SKUs are bundles — VMware Cloud Foundation (VCF) at the top, vSphere Foundation (VVF) below it. You don't buy the components separately anymore. Want vSphere on its own? You buy VVF and accept the vSAN and Aria capacity you may never use.
vSphere Standard — the entry-level SKU most Australian SMBs actually ran — is effectively a dead end. It's capped at vSphere 8 Update 3 with no upgrade path to vSphere 9. Once vSphere 8 reaches end of support, Standard customers have no supported route forward without moving to VVF or VCF. For an SMB planning a three-to-five-year infrastructure lifecycle, buying a product with a visible expiry date is not a plan.
That leaves VVF as the realistic floor for anyone who needs a supported path. VVF lists at approximately USD $190 per core per year on a one-year term. At current exchange rates that lands around AUD $265–$275 per core before reseller margin — in practice, Australian SMBs are seeing quotes in the AUD $230–$290 range depending on term length and channel. For a two-socket server with 16 cores per socket, that is AUD $7,360–$9,280 per year in hypervisor licensing alone, for a single host. Scale that across a three-node cluster and the annual licensing bill sits north of $22,000 before you touch storage, networking, or backup.
Then the per-core floors. Each CPU is licensed for at least 16 cores even if it has fewer. The smallest VCF order is 72 cores. Through 2025 that 72-core floor applied across the board. Late in 2025 Broadcom partially reversed it for some existing renewals, but the minimum still bites on new orders, subscription transitions, and any tier change — which captures most renewals. Late renewals themselves attract a 20% penalty applied retroactively to the first-year subscription price, with no grace period. Miss the anniversary by a day and the maths gets uglier.
The SMB cost shock, in one row
The most-cited public example is a UK university whose annual support went from the equivalent of $80,000 to $1 million — a 1,250% increase, driven entirely by the VCF bundle floor. SMB cases in the 350-450% band are now routine, particularly for businesses running a small number of low-core hosts. The eight-core-server-licensed-as-72-cores maths is not an edge case. It is how the model is structurally designed.
Why this is segmentation, not pricing
Look at who Broadcom is keeping and who it is losing, and the shape of the customer cull is hard to miss. Large enterprises with hundreds of cores already deployed barely feel the per-core floors — the minimums sit well below their actual usage. Their increases run in the 50-150% band, which negotiating rooms can absorb. SMBs running two or three hosts with eight or sixteen cores apiece are seeing 4-9× licence inflation before any list-price change at all, purely from the floor. That is not how a vendor structures pricing if they want to keep that customer. It is how a vendor structures pricing when they want the customer to leave on their own and feel like it was their choice.
This is not new behaviour from Tan. He ran the same playbook at CA in 2018 and Symantec's enterprise business in 2019: focus on the top decile of customers, drop the support cost of everyone else, take the revenue hit on the way down, rebuild margin on the concentration. VMware is the largest version of that move, not a different one.
The 72-core minimum is not a pricing decision. It's a segmentation decision, expressed as price.
What this means for a Victorian SMB facing a renewal
Three things, none of them complicated, all of them time-sensitive.
First, treat the renewal date as a hard deadline, not a target. The 20% retroactive penalty plus the missing grace period turns a slipped anniversary into a real cost that doesn't negotiate down. Calendar it now. Internal procurement workflows that took two weeks under perpetual licensing will not survive contact with this model.
Second, do the migration maths properly. The interesting question is not "how much more is VMware in 2026?" — it is "what is the three-year cost of staying versus moving, including the engineering time to migrate?" The honest answer for most 30-150 seat SMBs we look at is that staying on VMware now costs more over three years than migrating to Hyper-V, Proxmox, Nutanix, or a managed private cloud. The migration has stopped being the expensive option. The renewal is.
Third, stop trying to negotiate. The deal terms are not the lever. Broadcom's segmentation is structural, not transactional. Reps who used to discount to 60% of list now have neither the authority nor the incentive to do so. The right conversation is about exit, not discount.
This is the same dynamic the Cloud 3.0 unbundling is forcing one layer up the stack. The hyperscaler bundle is being unbundled by cost pressure; the hypervisor bundle is being unbundled by licensing pressure. The discipline is the same in both cases — place each workload deliberately, against the platform it actually fits, and stop paying for capabilities that aren't load-bearing.
Communicat runs Australian-hosted private cloud and managed Hyper-V for SMBs across Victoria specifically because we expected this moment. If you are renewing in the next twelve months, the first move is not a quote comparison — it is a workload audit. How many hosts. How many cores actually in use. Which workloads need vSphere-grade features and which ones do not. Most VMware footprints we look at are 30-50% over-provisioned for capabilities that nobody is using. That over-provisioning was free under perpetual licensing. Under the bundle model it is the line item Broadcom is monetising. Knowing what you actually run is what the next platform decision is built on.
If you would like a second pair of eyes on a 2026 renewal, that is the work we are running for clients across Victoria right now — workload audit, three-year cost comparison against Hyper-V, Proxmox, and managed private cloud, and a migration plan only if it stacks up. Get in touch and we will run the numbers with you.
A renewal in 2026 is a strategy decision
A perpetual-era VMware renewal was a procurement task. A 2026 VMware renewal is a five-year strategy decision about which platform your business runs on. Treat it as the latter — give it the time and analysis it deserves — and the answer for most SMBs is now somewhere other than VMware. The 72-core minimum, read correctly, is Broadcom telling you that themselves.
Frequently asked questions
What is the VMware 72-core minimum?
The 72-core minimum is the smallest VMware Cloud Foundation order Broadcom will accept under the post-acquisition licensing model. Below that floor, the licence count is rounded up to 72 regardless of how many cores the customer's hardware actually has. An eight-core server therefore licences for nine times its physical core count. Broadcom partially relaxed the rule for some existing renewals late in 2025, but it still applies to new orders, subscription transitions, and most tier changes — which captures most SMB renewals.
Is VMware vSphere Standard still available?
vSphere Standard still exists as a subscription but it is a dead end. It is capped at vSphere 8 Update 3 with no upgrade path to vSphere 9. Once vSphere 8 reaches end of support, Standard customers have no supported route forward without moving to vSphere Foundation (VVF) or VMware Cloud Foundation (VCF). For Australian SMBs planning a three-to-five-year infrastructure lifecycle, it is not a viable long-term option.
How much does VMware vSphere Foundation cost in Australia?
VVF lists at approximately USD $190 per core per year on a one-year term. At current exchange rates that translates to around AUD $265–$275 per core before reseller margin. Australian SMBs are typically seeing quotes in the AUD $230–$290 range depending on term length and channel partner. Multi-year commitments can bring the rate closer to AUD $210–$215 per core. For a typical two-socket, 16-core-per-socket server, that is AUD $7,360–$9,280 per year in hypervisor licensing for a single host.
Are perpetual VMware licences still valid in 2026?
Yes, but with limits. Existing perpetual licences continue to operate, but no new perpetuals can be purchased — Broadcom ended that channel in February 2024. Once a customer's support and subscription (SnS) contract expires, the perpetual licence keeps running but receives no further patches, updates, or security advisories. Most SMBs on perpetual SnS in 2026 are deciding between a subscription transition and migration off VMware entirely.
What is the late renewal penalty for VMware in 2026?
Late renewals attract a 20% penalty applied retroactively to the first-year subscription price, with no grace period. The penalty starts the day after the anniversary date. Broadcom does not negotiate the penalty out for SMB and mid-market customers, so a slipped renewal becomes 20% more expensive even if the contract is signed shortly afterwards.
What are the alternatives to VMware for Australian SMBs?
The realistic alternatives sit in three buckets. First, a different hypervisor on the same hardware: Microsoft Hyper-V, Proxmox VE, Nutanix AHV, or XCP-ng. Second, a managed private cloud — outsourced infrastructure on Australian-hosted hardware, which suits SMBs that want out of hardware management. Third, selective rebuild on hyperscaler IaaS for workloads that genuinely fit cloud-native economics. The right answer depends on workload profile, in-house skills, and renewal date. A workload audit is the prerequisite to choosing.
Should we renew VMware in 2026 or migrate?
The decision turns on three-year total cost rather than next-year renewal cost. For most SMBs running fewer than 100 cores of vSphere, migration cost — engineering time, retraining, transition risk — is now lower than the cost of staying on VMware over three years. Businesses with more than 200 cores or heavy NSX and vSAN dependencies face a more nuanced call. Either way, the decision deserves a workload audit before it deserves a quote.
Written by
Managing Director
John Zammit is Managing Director at Communicat IT, a Melbourne MSP serving Victorian SMBs since 1987. He writes about cloud economics, infrastructure strategy, and the gap between sales narratives and operational reality.