Laptop Refresh Cycle Policy: 3-Year vs 4-Year Maths for Indian Companies

The HP EliteBook that cost ₹47,000 to keep alive

In March 2025 a 120-person engineering firm in Whitefield asked us to look at their laptop fleet. They had bought 84 HP EliteBook 840 G7 units in early 2021 on a four-year refresh assumption. By the time we were called, 61 of those machines were still in service in year five. Three of them had swollen batteries. One had taken a power surge in a BTM Layout co-working space and the motherboard was quoting ₹28,500 to replace, on a machine with a book value of zero.

The finance head's position was simple: "The laptops work. Why replace them?" The IT manager's position was simpler: "My team spends half its day on these things."

We pulled the ticket data. Over the previous 12 months, those 61 machines had generated 214 support tickets at an average 2.1 hours of engineer time each, billed at ₹750/hour under their AMC. That is ₹3,37,000 in pure support cost for hardware in its fifth year. Add the ₹47,000 emergency motherboard (they went ahead with it, against our advice), the four out-of-warranty battery replacements at ₹6,800 each, and the two loaner laptops they had to rent while machines were in service, and the fifth-year running cost crossed ₹3.9 lakh. Replacing all 61 would have cost ₹36.6 lakh. Keeping them cost roughly 11% of replacement value in a single year — and rising.

The point is not that they should have replaced in year five. It is that their four-year policy was never really a four-year policy. It was a four-year assumption with nothing behind it. Nobody had budgeted the year-five cost, nobody had modelled the support curve, and nobody had told finance what a fifth-year laptop actually costs to run.

That is what a laptop refresh cycle policy is for. Not the number — the number is easy. The policy is what makes the number hold when a finance head asks why.

Why most Indian companies get this wrong

We see the same three mistakes across 200- to 5,000-seat organisations in Bangalore and elsewhere. They are worth naming, because each one distorts the refresh maths.

Mistake one: treating refresh as a capex decision

Refresh is a running-cost decision disguised as a capex decision. A ₹75,000 laptop bought in 2026 is not a ₹75,000 decision. It is a ₹75,000 decision plus four years of AMC, software licences, support labour, and resale value. Finance sees the first number. IT lives with the rest.

When a company says "we refresh every four years," they are usually saying "we depreciate over four years and hope the machine lasts." Those are not the same thing. Depreciation is an accounting convention. Useful life is a hardware and software fact. The two should not be confused, and in most Indian SMEs they are.

Mistake two: refreshing the whole fleet at once

We regularly meet companies that bought 300 laptops in a single quarter in 2022 and now face a 300-laptop refresh in 2026. That is a ₹1.8-2.4 crore wall. It forces them into one of three bad choices: extend the refresh by a year (raising support cost), buy cheaper machines than they need (raising support cost), or defer the whole project and hope (raising support cost).

Staggered replacement solves this. More on that below.

Mistake three: ignoring resale value

A three-year-old business laptop in India has real residual value — typically 18-28% of purchase price for mainstream business models, higher for premium lines like the ThinkPad T series or MacBook Pro. A five-year-old laptop is worth scrap. The difference between refreshing at year three and refreshing at year five is roughly ₹12,000-₹20,000 per machine in recovered value. On 300 machines that is ₹36 lakh to ₹60 lakh, which is not a rounding error.

Treating old laptops as e-waste to be hauled away is one of the most common ways Indian companies quietly lose money. We will come back to this.

The actual maths: 3-year vs 4-year vs 5-year

Let us build this properly with a real model. Take a Dell Latitude 5450 (Core i5-1335U, 16GB, 512GB NVMe, 14-inch FHD) bought new for ₹72,000 in early 2026, deployed to a 150-person company using a managed services provider at typical Bangalore AMC rates. We will hold the hardware constant and vary only the refresh year.

Support cost curve

The single biggest mistake in refresh models is the assumption that support cost is linear. It is not. It is roughly flat for the warranted years and then rises sharply. Here is what we actually observe across our managed base for mainstream business laptops (Dell Latitude, HP EliteBook, Lenovo ThinkPad T-series):

These are averages. A single bad batch (a known hinge failure, a swollen battery generation, a thermal paste problem) will push a year well above these numbers. The HP EliteBook 840 G7 fleet above was running at roughly year-6 ticket rates in year five because of the battery issue.

Warranty matters here. A 3-year ProSupport or Care Pack on a Dell Latitude typically costs ₹9,500-₹13,000 in 2026, depending on model and coverage level. With it, years 1-3 support cost is largely the OEM's problem, not yours. Without it, year-3 support cost roughly doubles because every part failure lands on your AMC.

This is where the 3-year vs 4-year argument actually lives. A 3-year refresh means you buy the 3-year warranty and never run unwarranted hardware. A 4-year refresh means either you buy a 4-year warranty (rare, and expensive) or you eat year-4 support costs yourself.

The full cost model

Here is the model over a six-year horizon, per laptop, using 2026 India pricing and a ₹750/hour blended AMC rate:

Read that last row carefully. On pure rupees, the 3-year cycle and the 5-year cycle come out within 0.3% of each other over six years — and the 4-year cycle is actually the most expensive. That is not what most finance teams expect to see.

The 4-year cycle is the worst of both worlds. You pay for a fourth year of warranty you barely use, you take the productivity hit of a fourth-year machine, and you recover less on resale than at three years. It is a policy that exists because "four years" sounds sensible, not because anyone modelled it.

The productivity line

The productivity number above is the one finance will challenge. Here is how we compute it. A slow laptop costs its user time — in boot, in application launch, in builds, in Excel recalculation, in file saves, in joining a Teams call. Our own measurements across 400+ business laptops in 2025-26 (using Windows Performance Recorder and user surveys) give us roughly:

• Year 1-2: 4-6 minutes lost per working day
• Year 3: 8-12 minutes lost per working day
• Year 4: 15-22 minutes lost per working day
• Year 5+: 25-40 minutes lost per working day

At a loaded cost of ₹500/hour for a mid-level employee (₹8.7 lakh CTC / 2,080 hours, loaded), ten minutes a day costs ₹2,000 per year. Twenty minutes costs ₹4,000. Thirty minutes costs ₹6,000. Across 250 working days, that is the number — ₹6,000 at year 3, ₹18,000 at year 4, ₹42,000 at year 5, and it does not stop there.

We use ₹500/hour as a conservative default. For engineering, design, or finance teams, use ₹800-₹1,200/hour. The 4-year case gets worse immediately.

Which cycle is right?

In our experience across Bangalore managed-services customers:

• 3-year refresh is right for anyone whose staff bill time, whose machines run heavy workloads (development, CAD, video, data), or who faces compliance obligations. It is the default for 80% of our 150-500-seat customers.
• 4-year refresh is defensible only if the machines are genuinely light-use — reception, front desk, kiosks — and you have a strong warranty in place. It is never the right answer for general knowledge workers.
• 5-year refresh is only justifiable where budget genuinely cannot be found and the machines are non-critical. Do not pretend it is efficient. It is a budget decision, and it costs more in ways that do not show up in the capex line.

Staggered replacement: the policy that makes the maths hold

Even if you accept the 3-year argument, you cannot implement it by replacing a third of the fleet every year if your original purchases were all clustered. The fix is to stagger the initial de-clustering, then hold a steady state.

How to de-cluster a fleet

Suppose you have 300 laptops all bought in Q2 2022. You want to reach a 3-year refresh cycle. You cannot replace 100 per year from a single 300-machine cohort and maintain it — you would be replacing 100/year forever, which is the steady state. But the first transition is expensive because the 2022 machines are already 4 years old.

A workable transition plan:

1. Year 1 (2026): Replace the worst 100 machines — those with the highest ticket rates, heaviest users, or battery issues. Buy 3-year warranties.
2. Year 2 (2027): Replace next 100. The 2022 machines are now 5 years old but only the better half.
3. Year 3 (2028): Replace the remaining 100.
4. Year 4 (2029): The 2026 cohort is now 3 years old and refreshes.

That is 100 machines per year in steady state — but only because your fleet is exactly 300. The maths holds only if headcount is flat.

For a more honest model, assume 3% headcount growth (typical for a growing Indian SME) and a steady refresh. Your annual purchase becomes 100 + 3% of 300 ≈ 109 units. Budget for it.

What good stagger looks like over five years

A 300-seat company that started a properly staggered policy in 2024 is now in a stable rhythm:

The net capex is roughly ₹53-63 lakh per year at current pricing, with 3% headcount growth and ~27% resale recovery. That is the number to put into a rolling three-year budget, not "₹1.7 crore in 2026." The staggered version is smaller, smoother, and easier to defend.

Where companies get stagger wrong

Two traps.

First, they stagger on age alone. A four-year-old laptop on a reception desk and a four-year-old laptop running Visual Studio builds are not the same risk. Stagger by role-weighted risk, not by purchase date. Give the worst machines to the lowest-load users if you must, but refresh heavy users first.

Second, they let the process drift. After two years of discipline, someone decides the new marketing hire does not need a new machine, and pulls from the loaner pool. Then the pool empties. Then the stagger collapses. Hold the line or do not start.

Warranty decisions matter more than refresh year

We have watched two companies with identical 4-year refresh policies have entirely different outcomes because of warranty choices.

A Dell Latitude 5450 with 3-year ProSupport in 2026 costs roughly ₹11,000 over base. With 4-year ProSupport, roughly ₹18,500. The gap is ₹7,500 — and it is the difference between a fleet where year-4 failures are the OEM's cost and a fleet where they land on your AMC and your team.

If you are going to run a 4-year cycle anyway, at minimum buy 4-year warranty on every machine. If you buy 3-year warranty and run 4-year refresh, you are self-insuring year 4 against yourself — which is the same as saying your AMC covers it, which means you have already paid for it in the AMC line.

On Apple devices (MacBook Air M3, MacBook Pro M3/M4), AppleCare+ for Business in India runs roughly ₹22,000-₹34,000 for 3 years depending on model. It is not cheap, but MacBooks have very low year-3 and year-4 support incidence historically — our Mac ticket rate at year 4 is roughly 60% of the equivalent Windows business laptop rate. If you are refreshing Macs and the workload is light, a 4-year Mac cycle is more defensible than a 4-year Windows cycle. Do not generalise; measure.

Productivity loss: the line everyone underestimates

The productivity line is where most policy documents get quietly soft. Let us be blunt about it.

A 2021-vintage laptop with a 11th-gen Core i5 and 8GB RAM running Windows 11 with Teams, Outlook, Chrome with 40 tabs, and whatever endpoint security stack your IT team has — that machine is not "just a bit slow." It is functionally a different workday. Our internal measurement on a sample of 60 such machines in mid-2025 showed:

• Boot to usable desktop: 3 min 20 sec average (vs 38 seconds on a 2024 machine)
• Teams call join: 22 seconds average (vs 6 seconds)
• Excel recalculation of a 50,000-row model: 14 seconds (vs 3 seconds)
• Chrome tab switch with 30+ tabs open: noticeable lag almost always

The user notices. They do not file a ticket for "my laptop is slow" because they know what IT will say. They just work around it — close tabs, work in the office later, save files differently, avoid the heavy report. That is the productivity loss, and it does not appear in any dashboard.

At 20 minutes a day on 150 users, that is 500 hours per week of lost output. At ₹500/hour loaded that is ₹2.5 lakh per week. Even if you discount 80% of that as noise, you are looking at ₹50,000 per week — ₹2.6 lakh per month — across a 150-seat fleet sitting one year too long on refurbishable hardware.

This is the number to bring to finance. Not "our laptops are old." Not "users are complaining." The number.

Resale and disposal: recovering ₹40-60 lakh you did not know you had

In India, the used business laptop market is large, active, and pays real money for the right machines. The typical recovery on a 3-year-old mainstream business laptop in 2026:

The gap between 3-year and 5-year resale is ₹14,000-₹18,000 per mainstream machine. On 300 machines that is ₹42-54 lakh sitting in a store room.

How to actually recover it

Three routes, ranked by what we see work:

1. Specialist IT asset disposition (ITAD) partner. Typically pays 70-80% of market rate but handles data sanitisation certificates, GST-compliant invoicing, and bulk pickup. The right choice for fleets over 50 units. Expect 2-4 weeks from pickup to payment.
2. Direct resale via platforms (Cashify Business, Reboot, similar). Pays better (85-95% of market) but you handle the data wiping, the pickup logistics, and the buyer risk yourself. Fine for under 30 units.
3. Employee buy-back scheme. Popular in Bangalore. Employees buy their own ex-corporate laptop at a discount to market. Good will, zero logistics, but you must still sanitise the drives and you must not let it become a quiet subsidy that distorts the refresh budget.

Data sanitisation is not optional

This is a DPDP Act obligation, not a nice-to-have. Every laptop that leaves your premises must have its storage drive either physically destroyed or cryptographically erased with a certificate on file. NIST 800-88 Rev. 1 is the standard most Indian ITAD vendors will cite. Get the certificate. Keep it for the retention period your legal team specifies — typically 8 years under the DPDP framework for records relating to processing.

We have seen companies hand 40 laptops to a scrap dealer for ₹8,000 total, with the drives intact. That is a data breach waiting to be discovered and a DPDP enforcement action waiting to be filed. The ₹8,000 is not worth it.

GST treatment on resale

Selling used laptops attracts GST. If you sold the machine as a capital asset, GST applies at 18% on the sale value (used IT equipment is not in the exempt categories). If you are an ITAD route and the buyer is a registered dealer, they will handle the GST on the onward sale under margin scheme typically, but your sale to them is taxable. Budget for this in your net recovery model. A ₹20,000 resale is ₹20,000 gross, ₹16,950 net of 18% GST if you are not eligible to claim input on the original purchase (which you generally are not, since you claimed depreciation not ITC).

The lease vs buy decision in the refresh maths

A quick note since every refresh conversation reaches it. In 2026, laptop leasing in India (through OEM financial arms and specialist lessors) runs at roughly 3.2-4.1% per month of asset value for a 3-year operating lease — call it ₹2,300-₹2,950 per month for a ₹72,000 machine, all-in, with warranty and end-of-term disposal handled.

Whether that beats buying depends on:

• Cash flow. Leasing preserves working capital. Buying is cheaper over the full term if you have the cash.
• Balance sheet treatment. Operating leases are off-balance-sheet, which some CFOs want and others do not.
• Disposal risk. The lessor takes resale risk. If you would otherwise leave 300 dead laptops in a store room, leasing may be cheaper than you think.
• Tax. Lease rentals are fully deductible operating expense; owned assets attract depreciation at 40% WDV for computers, which is generous but not unlimited.

For most of our 100-500 seat customers, buying plus an ITAD partner beats leasing on total cost. For fast-growing companies where cash is tight, leasing wins. Both are defensible — what is not defensible is buying without a disposal plan and calling it "saving money."

The policy document itself

If you only do one thing, write the policy. A laptop refresh cycle policy that is not in writing does not exist. Here is what ours looks like, stripped down. Adjust to your context.

What must be in the policy

1. The refresh year. Stated as "refreshed in the third year of service, at the point of the warranty anniversary" or similar. Not "every three years" — that invites drift.
2. The exception criteria. Written down. Heavy workloads, compliance obligations, documented performance issues, battery swelling — each exception gets its own rule and approver.
3. The warranty standard. Which OEM, which tier, which duration. Do not leave it to the buyer to negotiate per order.
4. The disposal route. Named ITAD partner or internal process. Named data sanitisation standard.
5. The resale value assumption. What you will recover, used in the budget.
6. The loaner pool size. Typically 4-6% of fleet. A 300-seat company needs 12-18 loaners.
7. The approval matrix. Who signs off on an out-of-cycle replacement (₹0 to ₹75k? ₹75k to ₹1.5L? Above?). Get this right or every refresh becomes a negotiation.
8. The OS and software baseline. A 3-year-old laptop running a modern OS, current browser and current endpoint agent is fine. A 3-year-old laptop running the software baseline from its purchase date is a security incident.

What should not be in the policy

• Specific model numbers. They change every year. Specify the class ("Dell Latitude 5000 series or equivalent") and the minimum spec.
• Specific prices. Refresh policy is not a purchase order.
• Vague quality language. "Adequate performance" is not a policy, it is a hope.

The failure story worth reading twice

Let me tell you about a 260-seat financial services firm we took over in mid-2024. They had a 5-year refresh policy — explicitly chosen because "laptops are expensive." They had 187 Dell Latitude 5410 and HP ProBook 450 G7 units still in service from 2020 and 2021. The finance team had signed off a plan to stretch to year six.

In August 2024, one of those machines — a ProBook 450 G7 in their collections team — failed mid-month with a battery swelling event that cracked the palm rest. The user had been complaining about the machine for six weeks. Nothing had been done because the loaner pool was empty (they had 4 loaners for 260 users) and the policy said "replace at year five."

The user was on a call with a customer when it happened. Data loss was minor. Disruption was significant. But the bigger cost was downstream: the collections team of 14 all had identical machines within 2 months of warranty expiry, and after the incident three of them refused to use the machines, demanding replacement. HR got involved. Legal got involved because the data loss touched a client file. The firm replaced 22 machines in four weeks at spot market prices — ₹81,500 each against a planned ₹72,000 — because they had no purchase order in place and no lead time. The emergency replacement cost ₹17.93 lakh against a planned ₹15.84 lakh. Extra: ₹2.09 lakh, plus lost days, plus an external legal review, plus reputational friction with one enterprise client.

The lesson is not "always refresh at three years." The lesson is that a policy without an exception process, without a loaner pool, and without a purchase mechanism in place is not a policy. It is a deferral that becomes a crisis. We rebuilt their policy, sized their loaner pool to 14, wrote the exception criteria, and put their next three refresh years into a rolling budget. Their 2025 and 2026 refreshes have run on plan.

If a similar situation is on your desk right now, it is worth talking to us directly rather than reading another article. We do this weekly.

Where our approach is not the right fit

We would rather say this clearly than have you find out later.

• Very small fleets. Under 40 laptops, formal refresh policy is overkill. Buy good machines, keep them three years, replace on failure. The economics do not justify the process.
• Rapidly changing headcount. If you are doubling headcount year on year, a de-clustered refresh policy is more paperwork than value. Refresh on role change instead.
• BYOD-heavy environments. If most of the fleet is employee-owned, refresh policy is a stipend policy. Different animal.
• Specialist workstations. CAD, video, ML workstations do not follow laptop economics. Refresh by workload, not by age.
• Companies with highly customised software stacks (embedded development, legacy ERP clients) where a laptop refresh triggers a software revalidation: refresh cycles need to be aligned to the software revalidation cycle, not the hardware life. That is a different project.

Most of our customers sit in the 100-800 seat band where a proper policy pays for itself many times over. If you are outside that band, talk to us about whether a formal policy makes sense before you write one.

A practical 90-day implementation plan

If you decide today to put a refresh policy in place, here is what the first quarter looks like.

Days 1-30: audit and baseline

Pull the asset register. Every laptop, purchase date, model, warranty expiry, assigned user, role. Most companies find 8-15% of their fleet is missing from the register or assigned to a departed employee. Fix that first — refresh policy on a wrong asset base is decoration.

Then pull 12 months of ticket data per machine. You will find the top 20% of machines account for 60-70% of tickets. These are your first-refresh list, regardless of age.

Measure 10 representative machines with Windows Performance Recorder or a similar tool. You want a real baseline of boot time, application launch, and disk I/O. Not guesses.

Days 31-60: policy and budget

Write the policy against the template above. Get finance and HR to sign off, not just IT. If HR is not at the table, the exception and employee buy-back parts will not survive first contact.

Model three scenarios — 3-year, 4-year, 5-year — with your actual numbers. Present all three to finance with the productivity and resale lines included. Even if finance picks 4-year, they should pick it knowingly.

Build the three-year rolling budget. Not the annual budget. The rolling one.

Days 61-90: transition

Pick your ITAD partner. Get a data sanitisation SLA in writing. Run a pilot refresh of 10-15 machines through the full lifecycle — purchase, deploy, and (for the pilot only) dispose of one existing machine — so the process is tested before you run 100 through it.

Size the loaner pool. Order the loaners.

Then start the staggered plan.

Where to buy in Bangalore

We have a separate guide to business laptop dealers in Bangalore that covers the broader purchase question. For refresh specifically, two points worth carrying over:

• Lead times. Dell Latitude and Lenovo ThinkPad T-series business configurations, ordered through authorised channels in Bangalore, are typically 2-4 weeks in 2026. HP EliteBook is often 3-5 weeks. Do not schedule a refresh to a deadline tighter than that.
• Price stability. Laptop pricing in India is sensitive to rupee-dollar movement and to quarterly OEM promotions. Q4 (Jan-Mar, Indian fiscal) and Q1 (Apr-Jun) tend to have the best enterprise pricing as OEMs chase targets. If you can shift a refresh into these windows, you can save 3-6% per unit.

If you need help choosing machines for the refresh, the laptops page is a fair starting point.

FAQ

How often should a company replace its laptops in India?

Every three years is right for most knowledge-worker fleets. Four years is defensible only for light-use roles with strong warranty coverage; five years is a budget decision that costs more in support and productivity than it saves in capex. Our 2026 model shows a 3-year cycle at roughly ₹71,500 net cost per cycle and a 4-year at ₹1,00,000 — the 4-year is the most expensive of the three over a six-year horizon.

Is 3-year or 4-year laptop refresh cheaper?

Over six years, 3-year and 5-year come out roughly equal on pure rupees (about ₹1.43 lakh per user each), while 4-year is the most expensive at about ₹1.50 lakh. The 3-year cycle wins on productivity, security and resale, and is the right default unless budget genuinely cannot support it. Do not assume 4-year is a natural middle ground — it is not.

What is the resale value of a 3-year-old business laptop in India?

Typically 24-33% of purchase price for mainstream business models (Dell Latitude, HP EliteBook, Lenovo ThinkPad T-series) and 36-45% for Apple MacBook Air and MacBook Pro. A 5-year-old equivalent recovers 5-10%. On a 300-machine fleet, the difference between 3-year and 5-year disposal is ₹42-54 lakh.

How do we handle data wiping before disposal under DPDP?

Every drive must be either physically destroyed or cryptographically erased, with a certificate on file. NIST 800-88 Rev. 1 is the standard to specify. Insist on a certificate per machine from your ITAD partner, keep it for your legal retention period, and do not use a scrap dealer who cannot produce one — that is a reportable breach risk.

Should we lease or buy laptops on a 3-year cycle?

2026 India operating leases run at roughly 3.2-4.1% per month of asset value — about ₹2,300-₹2,950/month for a ₹72,000 laptop, all-in. Buying plus an ITAD partner usually beats leasing on total cost if you have the cash. Leasing wins where cash is tight or where disposal capability is weak. Both are defensible; buying without a disposal plan is not.

How big should the loaner pool be?

Four to six percent of the fleet. A 300-seat company needs 12-18 loaners. Less than that and your exception process cannot function — every failure becomes an emergency purchase at spot prices.

Next step

If you have read this far and your current policy is "four years because that is what we have always done," the next step is not to write a new policy. The next step is to pull your ticket data for the last 12 months, per machine, and see what your year-4 and year-5 laptops are actually costing. That single exercise typically moves the conversation further than any policy document.

If you would like a hand with it, our team can run that analysis on your fleet — typically 5 working days for fleets up to 500 machines, with the numbers you need to take to finance. Bring the output to your next budget review and the decision makes itself.