Underneath every hardware lifecycle decision is a money decision, and it comes down to three options: lease the gear, buy it outright, or buy it once and run it far past its depreciation into the ground. Each has a genuinely different risk profile. Each has a sales pitch attached that emphasizes its strengths and goes quiet on its costs. Sorting the real tradeoffs from the financing pitch is worth doing before you sign anything.

Buy outright

You pay the full capital cost up front and own the asset. The pitch is simple: no interest, no ongoing payment, the gear is yours.

  • The real advantage: once it's paid for, it's free compute. A bought server that's fully depreciated and still healthy costs you nothing but power and the occasional part. This is where the "run it into the ground" option gets its enormous value — you've already paid, so every extra year is pure return.
  • The real cost: capital up front, and you own the disposal and the obsolescence risk. If the workload outgrows the box or the technology moves, you're holding an asset you have to retire and replace with more capital. You carry the full lifecycle risk because you own the full asset.
  • Best when: you have the capital, the workload is stable, and you intend to run the gear for its full useful life and then some. Buying to run for three years and discard is the worst case for buying — you took all the capital risk for a short horizon.

Lease

You pay a regular amount over a term and hand the gear back (or renew) at the end. The pitch leans on cash flow and "always current hardware."

  • The real advantage: no large capital outlay, predictable operating expense, and someone else owns the obsolescence and disposal risk. At term end, you refresh to current gear without a disposal problem. For organizations that value cash-flow smoothness and want to stay on a refresh treadmill, this is genuine.
  • The real cost: you pay more in total — the lease includes the financing cost and the lessor's margin on the residual-value bet. And you own nothing at the end; the payments stop buying you anything the day the term ends unless you renew or buy out (often at an unattractive price). The "always current hardware" benefit only materializes if you actually refresh at term — if you'd have happily run the gear five more years, the lease forced you off an asset that still had value.
  • Best when: you genuinely want the refresh treadmill, value predictable opex over total cost, and don't want to own disposal/obsolescence. It's a bad fit for gear you'd otherwise happily run into the ground — you're paying a premium for a flexibility you won't use.

Run it into the ground

Not a separate purchase method — it's what you do after buying: run the owned, fully-depreciated gear far past its book life, replacing parts as they fail, until the economics or the risk finally say stop.

  • The real advantage: the lowest total cost of ownership available, by a wide margin, for the right workloads. You already paid the capital; every additional year of service is return on a sunk cost. Parts are cheap on the secondary market for older platforms. This is how disciplined shops get extraordinary value out of hardware budgets.
  • The real cost: rising failure risk (the bathtub tail), growing security exposure as the platform ages past support, and the operational burden of keeping older gear fed with parts and patches. Run into the ground without instrumentation and you're gambling; run it with SMART data, failure monitoring, and a risk assessment and you're extracting maximum value from an asset you own.
  • Best when: the workload is stable and survivable-if-it-fails (secondaries, clustered roles, non-critical services), the gear is healthy by the data, parts are available, and you have the discipline to monitor wear and track the security exposure. It's the right answer far more often than the five-year-refresh crowd admits — and the wrong answer for anything where a surprise failure or an unpatchable exposure would genuinely hurt.

What the financing calculator leaves out

Vendor and lessor calculators are built to make their preferred option look best. The things they tend to omit:

  • The value of the run-into-the-ground years. Lease calculators assume you refresh at term; they never price the three to five extra years you'd have gotten by owning. That omission is the whole game.
  • The total-cost-of-leasing premium. Monthly numbers hide that the sum plus financing exceeds the purchase price, and that you own nothing at the end.
  • Your actual failure risk. Both "buy and refresh on schedule" and "lease" quietly assume hardware must be replaced on a calendar. Your real failure data often says otherwise, which shifts the math toward owning and running longer.
  • The disposal and obsolescence risk you're really carrying. Buying means you own it; that's a real cost the buy pitch underweights and the lease pitch overweights.

The honest frame

There's no universal answer, but there is a clear structure. If the workload is stable and survivable and you have the operational discipline, buy and run it into the ground — it's the lowest total cost, full stop, and the most undervalued option in the room. If you genuinely want the refresh treadmill and value predictable opex over total cost, lease — just don't do it for gear you'd rather own and keep. And buy-to-refresh-on-schedule is really the weakest of the three: you take the full capital risk of ownership but throw away the run-into-the-ground value that was the main reason to own.

The lifecycle decision and the money decision are the same decision wearing two hats. Decide them together, with your real failure data in hand, and ignore whichever calculator was built to sell you the answer that pays its author.