Blog
Still on Windows 10? October 2026 is the expensive deadline
Windows 10 reached end of support in October 2025, and the estates that bought Extended Security Updates bought themselves a year. That year ends on 13 October 2026, and the price of staying doubles. Here is the honest cost curve, and the evidence-based way off it.
Windows 10 left mainstream support on 14 October 2025. Plenty of enterprises met that date the pragmatic way: enrol in Microsoft's Extended Security Updates (ESU) programme, keep the patches flowing, and defer the migration. Fair enough; ESU exists for exactly that reason. But the programme was priced to be a bridge, not a home, and the tolls are about to go up.
The clock that matters now
ESU Year 1 coverage ends on 13 October 2026. From that point, keeping a Windows 10 device patched means buying Year 2. Microsoft's published commercial list price for Year 1 was 61 US dollars per device; Year 2 doubles it to 122 dollars, and Year 3 doubles it again, before the programme closes altogether in October 2028.
There is a sharper edge for anyone who skipped Year 1. ESU is cumulative: enrol a device late and you must also buy the years you missed. A device joining in Year 2 costs Year 1 plus Year 2, which is 183 dollars before it receives a single new patch. Multiply any of these figures across ten thousand devices and the "cheap deferral" becomes a seven-figure line item that buys nothing except more time on an operating system Microsoft has already retired.
What ESU actually buys
It is worth being precise, because ESU is narrower than many renewal conversations assume. ESU delivers security updates rated critical or important. That is all. There are no new features, no non-security fixes, no design changes, and no general Microsoft support for the platform. Devices must also be current on cumulative updates before they can enrol at all.
Meanwhile the world around the operating system keeps moving. Application vendors are steadily dropping Windows 10 from their support matrices, security frameworks and cyber insurers take an increasingly dim view of extended-support estates, and every month on ESU widens the gap your eventual migration has to jump.
"We bought ESU" is a plan for a year, not a strategy
ESU is a bridge with a doubling toll. The only good reason to be on it is that you know exactly when you are getting off.
ESU is a perfectly rational purchase when it funds a migration that is actually happening. It is an expensive habit when it substitutes for one. The doubling curve is Microsoft telling you, in the clearest language a vendor has, that the programme is meant to be left. An estate that renews into Year 2 without a dated exit plan is not managing risk; it is renting it, at double last year's rate, with another doubling behind it.
What an evidence-based exit looks like
The estates that struggle with Windows 11 migration are almost never struggling with Windows 11. They are struggling with what they do not know: which devices meet the hardware requirements, which applications people actually use, and which of those applications will misbehave. So the exit starts with evidence, not with imaging.
Assess the estate first. A short, focused assessment establishes device-by-device Windows 11 readiness, real application usage rather than the licence list, and where the genuine blockers sit. Our Meridian platform, fed by Meridian's capture tool, builds exactly this picture from live estate data, so decisions rest on what is, not on what the CMDB remembers.
Rationalise applications before you migrate them. Most estates carry far more software than they use, and every redundant application is testing, packaging and remediation effort you do not need to spend. Rationalisation is where migrations shrink: we have cut a 26,000-application estate to about 1,000 before a migration. Few estates are that extreme, but the ratio of installed to needed is rarely flattering anywhere.
Then migrate on evidence. With a rationalised application set and a readiness-scored device estate, the migration itself becomes scheduling rather than archaeology: ring-based waves, hardware refresh only where the evidence says it is needed, known blockers handled up front, and rollback paths agreed before anyone touches a device.
The path before Year 2 bites
From September 2026 the arithmetic is simple. A short assessment produces the evidence in weeks. Application rationalisation, run against that evidence, shrinks the migration to the estate you actually use. A governed migration then lands wave by wave, and every device that reaches Windows 11 before 13 October 2026, or shortly after, is a device that never needs a Year 2 licence at all. Even a migration that runs into 2027 is cheaper started now: each completed wave is ESU spend you stop carrying.
ESU was the right call for the year it covered. The right call for this year is to make sure you never need Year 3.
Still carrying Windows 10 into 2027?
A short assessment puts evidence behind every device and application decision, and a governed migration gets you off the ESU curve before it doubles again.