Last fall, a lot of North Texas organizations made the same call. Windows 10 support was ending, the hardware refresh looked expensive, budgets were already committed and Microsoft was selling a one year extension for about sixty dollars a device.
Buying the year was often the right decision. It bought time, spread the cost and let leadership plan a refresh instead of panic-buying laptops in the fourth quarter.
The catch is that Extended Security Updates were designed as a bridge, and a fair number of organizations have been treating it like a destination. On October 13 the bridge tolls go up, and they keep going up every year after that.
What actually happens on October 13, 2026
Year one of Extended Security Updates ends. For consumer devices, that’s the end of the program entirely. For commercial and education customers, coverage can continue but the price ladder gets steep on purpose:
- Year 1, October 2025 through October 2026, roughly $61 per device
- Year 2, October 2026 through October 2027, roughly $122 per device
- Year 3, October 2027 through October 2028, roughly $244 per device
Carry a single machine all the way through and you’ve spent something like $427 per device to keep an aging operating system patched. After October 2028 there is no year four. Confirm current pricing and your specific licensing path before you commit, because the details vary by agreement, but the shape of it is not subtle.
Microsoft is not hiding the message here. The doubling isn’t a pricing accident. It’s a countdown with a dollar sign attached.
The license fee is not the real problem
Sixty-one dollars a device is annoying. A hundred and twenty-two is worse. Neither is what actually keeps this decision stuck.
The real issue is that a meaningful share of the Windows 10 machines still running out there cannot be upgraded in place. Windows 11 wants TPM 2.0 and a processor generation that a 2018 workstation simply does not have. So this was never really a licensing conversation. It’s a hardware refresh conversation that’s been wearing a licensing costume for about a year.
That’s why it keeps getting deferred. Renewing ESU is a purchase order. Refreshing forty workstations is a project, and projects need planning, scheduling, user disruption and a budget line somebody has to defend.
Three honest paths, and none of them are free
There’s no clever option here, which is worth saying plainly because you’ll read a lot of content this fall implying otherwise.
Refresh the hardware. Highest up front cost, lowest ongoing risk, and you’re done. If the machines are five or six years old you were going to face this anyway, and ESU has essentially been financing the delay.
Buy year two for part of the fleet. Reasonable when you’ve got specific machines tied to an application that isn’t ready, a department in the middle of a project or a capital cycle that really does start in Q1. This works as a decision. It works poorly as a default.
Move the workload rather than the device. For some roles a Cloud PC or a hosted desktop changes the math, particularly where the physical machine is mostly a keyboard and a screen. Not right for everyone and worth modeling carefully before anyone gets excited.
What we’d steer you away from is the fourth path, which is doing nothing and quietly running unpatched machines into next year. Your cyber insurance carrier has opinions about unsupported operating systems, and they tend to express them at claim time.
Where the compliance piece bites
If you’re a municipality, a water district, a healthcare group or anyone handling regulated data, this stops being a budget preference.
Unsupported operating systems show up as findings. CJIS, HIPAA and the FTC Safeguards Rule all expect systems to be supported and patched, and an auditor is not going to be charmed by an explanation involving capital timing. For public entities the added pressure is that the conversation happens in open session with a budget everyone can read.
Around DFW, we work with a number of organizations in exactly this spot. The pattern is consistent. The technical fix is the easy part. The hard part is sequencing it against a budget cycle and a council or board calendar that was set months ago.
What to do in the next three weeks
You don’t need a full project plan by October 13. You need enough information to make a deliberate choice instead of an accidental one.
Start with a current count. How many Windows 10 devices are actually still in service, not how many you think. Then sort them into which can upgrade in place, which need replacement and which are tied to an application or a piece of equipment that complicates things.
That last group is where the surprises live. The workstation running the machine controller. The PC attached to a lab instrument whose vendor supports exactly one operating system version. Those are the ones that turn a tidy refresh into a six week negotiation, and they are much better discovered in September than in the middle of it.
Then decide, on purpose, what gets refreshed now, what buys another year and what needs a different approach entirely.
The deadline is doing you a favor
Nobody enjoys a Microsoft deadline. But this one is forcing a conversation most organizations needed to have regardless, because hardware aging out is not really a Windows problem. It’s a lifecycle problem that Windows happened to put a date on.
Organizations with a working lifecycle plan saw this coming in 2023 and budgeted for it. Everyone else got a surprise, and the ones who bought a year of ESU bought themselves time to build that plan. October 13 is the day we find out who used it.
If you’re not sure where your fleet stands, that’s a short conversation and a worthwhile one. We’ll help you get an accurate count, sort out what can upgrade and what can’t and build a refresh plan that fits your budget cycle instead of fighting it. Ten minutes, no pressure and no IT jerks.
Call The Fulcrum Group at 817-337-0300 or visit www.fulcrumgroup.net. We’re in Keller, and we’ve walked a lot of North Texas organizations through this exact decision over the past year.
