October 11, 2027. That is the date vSphere 8 hits End of General Support — five years after its October 2022 GA, per Broadcom’s standard lifecycle policy. It sounds comfortably far away until you do the arithmetic the other direction: Gartner has pegged comprehensive VMware migration programs at 18 to 48 months, and even a disciplined in-place platform upgrade across a large estate eats two to four quarters. Run that math backward from October 2027 and the pay-or-leave decision is not due next year. It is due now, in 2026 — and most of the IT teams I talk to have not internalized that.
This brief lays out the three options — upgrade to VCF 9, buy extended support, or execute an exit — with the dates, the cost signals, and a quarter-by-quarter timeline you can hold your team to.
What changed
Three facts define the situation. First, vSphere 8’s End of General Support date is October 11, 2027 — after that, no patches, no security fixes, no support tickets under standard terms. Second, vSphere 8 was the last VMware release you could hold on a perpetual license. Everything forward is subscription, on Broadcom’s terms, and those terms moved again in 2025 when the per-CPU licensing minimum jumped from 16 to 72 cores. Third, the destination Broadcom wants you at is VMware Cloud Foundation 9, which went GA in June 2025.
Here is the part teams underestimate: VCF 9 is not a version bump. Practitioners who have run the upgrade describe it as an infrastructure redesign — it restructures how storage, networking, and operations tooling fit together, folding vSAN, NSX, and the Aria-era management stack into a single opinionated platform. Upgrade paths are constrained, too: getting to the current 9.x releases requires specific source versions, so estates on older VCF or standalone vSphere builds face a multi-hop sequence, not one maintenance window. If you budgeted this as “another vSphere upgrade,” you budgeted it wrong.
Why it matters
Because every option takes longer than the calendar suggests. An exit to another hypervisor at enterprise scale runs 18 to 48 months when you count discovery, procurement, pilot, migration waves, and the operational retraining nobody budgets for. A VCF 9 adoption at scale is realistically 6 to 12 months of design and execution. And the money has its own clock: FY2027 budgets get locked in during 2026 planning cycles. If the line item is not in the budget you are building right now, you have already chosen — you have chosen to pay whatever the renewal costs, from a position of zero leverage. Broadcom’s negotiators know exactly where October 2027 sits on your calendar; the honest framing is in our companion brief on negotiating a Broadcom VMware renewal.
The takeaway a VP can repeat in a meeting: the deadline is October 2027, but the decision deadline is Q4 2026. Miss the second one and the first one gets expensive.
Option 1: Upgrade to VCF 9 and stay with Broadcom
The case for staying is real. VMware’s platform remains the most feature-complete virtualization stack in the market, your operations team already knows it, and the ecosystem of backup, monitoring, and DR tooling around vSphere is unmatched. For estates deep in vSAN and NSX, VCF 9 genuinely consolidates what used to be four products’ worth of lifecycle management. If you are running thousands of VMs with hard availability SLAs and heavy VMware-specific automation, staying is often the lowest-risk path.
The case against is the cost structure and the loss of optionality. You are trading perpetual licenses you own for a subscription priced per core, with a 72-core minimum per CPU that punishes smaller hosts, and you are buying the full VCF bundle whether or not you want NSX and Aria. List pricing varies by commitment, but budget for a multiple of your old maintenance spend — the renewal shocks documented since 2024 have run from 2x to well past 5x. The full picture is in our brief on VMware licensing changes in 2026. Who it fits: large, VMware-native estates where migration risk outweighs subscription cost, and organizations that can negotiate multi-year terms before their leverage evaporates.
Option 2: Buy extended support and rent time
Broadcom has historically offered paid extended support past End of General Support, typically sold per year at a premium over standard support — as of mid-2026, expect it to be offered, priced to discourage, and capped at roughly two years of runway. Understand what it is: security patches and break-fix, no new features, no certification of new hardware. It is a bridge, not a destination.
Used well, extended support is the tool that turns a panicked 2027 migration into an orderly 2028 one. Used badly, it is an annual tax on indecision. The test is simple — if you have a signed migration plan with dated waves, buying a year of extended support is rational insurance. If you are buying it because the decision meeting keeps slipping, you are paying premium rates to avoid a conversation. Who it fits: organizations mid-exit whose migration waves genuinely cannot finish by October 2027, and regulated shops with change-freeze calendars that make a 2027 cutover reckless.
Option 3: Execute an exit
The two exit destinations that come up in nearly every evaluation are Nutanix and Proxmox, and they sit at opposite ends of the market.
Nutanix AHV: the enterprise landing zone
Nutanix offers the closest thing to feature parity with vSphere — HA, live migration, integrated storage, DR orchestration, and a mature management plane in Prism. Its Move tool automates VMware-to-AHV migrations with replication-based cutover, which materially shortens migration waves, and mainstream backup vendors support AHV natively. The honest caveat: Nutanix is premium-priced. You are trading one enterprise vendor for another, and if your exit is purely about cost you may be disappointed at quote time — the savings versus VCF are real but usually not dramatic. Where Nutanix wins is operational simplicity for hyperconverged estates and a vendor relationship that currently wants your business. Who it fits: enterprises above roughly 200 VMs that need support SLAs, want a supported migration path, and are exiting for predictability as much as price.
Proxmox VE: the open-source counterweight
Proxmox VE, built on KVM and LXC, has moved from homelab favorite to a credible enterprise contender, with a native VMware import wizard since version 8.1 and support subscriptions that cost a fraction of any commercial hypervisor. Clustering, HA, live migration, and integrated backup are all in the box. The honest caveats: management tooling at very large scale trails vCenter, the third-party ecosystem is thinner (though Veeam now supports it), and support is subscription-based European-business-hours engineering, not a global TAC. Who it fits: Linux-capable teams, estates under a few hundred hosts, edge and branch deployments, and organizations for whom the licensing line item is the whole point. Our full ranking is in VMware alternatives for 2026.
| VCF 9 (Broadcom) | Extended support | Exit (Nutanix / Proxmox) | |
|---|---|---|---|
| Decision deadline | Q4 2026 (renewal leverage) | Mid-2027 at the latest | Q3–Q4 2026 |
| Execution time | 6–12 months | Weeks (contractual) | 18–48 months full program |
| Cost signal | Multiple of legacy maintenance | Premium over standard, per year | Nutanix: moderate savings; Proxmox: large savings |
| Best fit | Large VMware-native estates | Mid-exit or change-frozen orgs | Cost-driven or leverage-seeking estates |
| Main risk | Lock-in, future price rises | Paying to defer, not decide | Migration effort, ops retraining |
The backward-planned timeline
Plan backward from October 11, 2027, and the quarters assign themselves:
- Q3 2026 (now): Inventory the estate — hosts, cores, VMware-specific dependencies (vSAN, NSX, SRM, vendor appliances). Get budget placeholders into FY27 planning for both the stay and leave scenarios.
- Q4 2026: Decision quarter. Run pilot workloads on at least one alternative, get real quotes from Broadcom and one challenger, and put a signed direction in front of leadership. This is also your maximum-leverage renewal window.
- Q1–Q2 2027: Execute the first migration waves or the VCF 9 design-and-upgrade sequence. Staying put should mean upgrade runbooks tested in a non-production pod by Q2.
- Q3 2027: Bulk migration waves or production VCF cutover. Anything still undecided here should trigger an extended-support purchase — deliberately, not by default.
- Q4 2027: vSphere 8 is out of general support. Whatever remains on it is now technical debt with a meter running.
What to do about it
Three moves, in order. One: treat Q4 2026 as the real deadline and put the decision meeting on the calendar this month — with the estate inventory as pre-read. Two: price all three options for your actual core counts, because the 72-core minimum and bundle composition change the math in ways a spreadsheet from 2023 will not predict. Three: whichever way you lean, run a pilot on an alternative anyway — a working Nutanix or Proxmox proof of concept is worth double-digit percentages in a Broadcom negotiation, even if you ultimately stay.
The short version: October 2027 is not the deadline. It is the consequence. The deadline is this fiscal year.
Frequently asked questions
When does vSphere 8 support end?
vSphere 8 reaches End of General Support on October 11, 2027 — five years after its October 11, 2022 GA under the standard lifecycle policy. Technical Guidance typically extends about two years beyond that, but without patches or new bug fixes.
Can I keep running vSphere 8 after October 2027?
Technically yes — the software keeps running. But you lose security patches and support entitlement, which most security frameworks and auditors treat as a compliance finding. For internet-adjacent or regulated workloads, running an unsupported hypervisor is a risk most CISOs will not sign off on.
Do I have to move to VCF 9, or can I buy vSphere on its own?
Broadcom’s packaging as of mid-2026 centers on VCF for enterprise estates, with vSphere Foundation as the slimmer tier. Standalone perpetual vSphere is gone — vSphere 8 was the last perpetually licensed release, so any forward path with Broadcom is a subscription.
How long does a VMware exit actually take?
Gartner’s range for comprehensive migration programs is 18 to 48 months. Small, simple estates with tools like Nutanix Move or the Proxmox import wizard can finish in one to two quarters; large estates with NSX, SRM, and deep automation sit at the long end.
Is extended support available for vSphere 8?
Broadcom has historically sold extended support past end-of-support dates, priced per year at a premium. Terms are negotiated, not published — treat it as available but expensive, and as bridge financing for a migration already in motion, not a strategy.
Enterprise Techie publishes vendor-honest analysis like this daily — get the brief by email, free.