For most of the history of enterprise networking, the CapEx model wasn't really a model — it was just how things worked. You needed network equipment. You bought it. You depreciated it over five to seven years. You ran it until it failed or until a refresh cycle forced replacement. The capital expense was a fact of doing business, like buying desks and chairs.

Over the last decade, a quiet shift has happened. Increasingly, the same connectivity capabilities are available as subscription services — networking-as-a-service, managed Wi-Fi, software-defined cellular, SD-WAN delivered as OpEx instead of CapEx. The shift is far enough along that "should we own this or subscribe to it?" is now a real strategic question for most enterprises.

The case for OpEx isn't ideological. It's that the underlying economics have genuinely changed.

What changed

Three things, mostly.

1. Software-defined everything. A lot of what used to require specialized hardware now runs as software on commodity equipment. When the value is in the software, it doesn't make sense to think of it as a one-time purchase — software is improved continuously, and the value compounds with updates.

2. Cloud-based management. The control plane for modern network gear sits in the cloud, not in a dusty closet. When the brains are cloud-hosted, the natural economic model is subscription. You're not buying a box; you're accessing an ongoing service.

3. Faster obsolescence cycles. Equipment that was state-of-the-art three years ago may be limiting today. The 5-7 year depreciation horizon doesn't match how fast the technology is actually evolving. Ownership traps you on equipment that's no longer competitive long before it's financially depreciated.

The CapEx model's hidden costs

The conventional wisdom is that buying is cheaper than renting in the long run. For networking gear, this is often not true once you account for what ownership actually involves.

The visible costs of owning network infrastructure:

  • Equipment purchase
  • Installation
  • Support contracts (which most enterprises buy anyway)

The less visible costs of owning network infrastructure:

  • Internal staff time to manage, monitor, and troubleshoot
  • Vendor management overhead across multiple equipment manufacturers
  • The lifecycle hits where equipment is too old to be optimal but too new to justify replacement
  • The "every refresh is a major project" cadence
  • Resale value (typically negligible)
  • End-of-life disposal and recycling costs
  • The cost of running equipment that's no longer being updated by the manufacturer (security risk)

When you total these up honestly over a 5-7 year window, the apparent cost advantage of CapEx often disappears. And it disappears completely if you factor in the capability differences — what you could have been doing with subscription-based infrastructure during that period that your owned equipment couldn't support.

The OpEx model's actual advantages

Subscription-based connectivity has some genuine advantages that aren't well captured in a basic cost comparison:

Continuous capability improvement. The software keeps getting better without you doing anything. The cellular protocols evolve. New device types get supported. Security improves. None of this requires a project on your end.

Aligned vendor incentive. A vendor selling you a one-time purchase has been paid. A vendor running an ongoing managed service is incentivized to keep you happy because you can leave. The economic structure rewards good support.

Predictable monthly cost. Finance teams love this. No surprise capital projects, no end-of-warranty refresh shocks. The cost line is flat and forecast-able.

Scale up or down with the business. Open a new location, add coverage. Close a location, remove it. Owned infrastructure doesn't have that flexibility.

SLA-backed performance. Subscription services typically come with service level agreements. Owned equipment doesn't — when it breaks, that's your problem.

When CapEx is still the right answer

For some enterprises and some situations, ownership still makes sense:

  • Highly customized environments where commodity solutions don't fit
  • Long-amortization investments where the depreciation math genuinely works (campus fiber infrastructure, for instance)
  • Regulatory or sovereignty requirements that mandate physical control
  • Cases where the in-house team is genuinely expert and adds value through deep control
  • Highly stable environments where the equipment will be relevant for its full useful life

For most other situations — most retail, hospitality, healthcare, warehousing, multi-location businesses — the OpEx model has quietly become the better fit.

The decision-making question

The honest question to ask isn't "is OpEx cheaper than CapEx?" It's "are we set up to be a successful owner of this infrastructure?"

Being a successful owner means having the staff to manage it, the discipline to refresh on time, the appetite for occasional project cycles, and the confidence that the technology won't move so fast that your investment is stranded.

For enterprises that can answer "yes" to all of those, ownership is fine. For everyone else — which is most enterprises — subscription is the move.