What a $340,000 Network Refresh Taught Me About ZTE vs Cisco (And My Own Assumptions)

The email that started all of this was three lines long. Our CIO forwarded it from the facilities manager: "Core switches are 7 years old. When are we refreshing?"

That was June 2023. I'm the procurement manager for a 420-person logistics company out of Columbus. I've managed our IT budget — about $1.4M annually, with roughly $340K earmarked for networking — for going on six years. I've negotiated with 30-plus vendors in that window and logged every order in our cost tracking system. I'm not an engineer. I'm the person who has to sign the PO and defend it to the CFO three quarters later.

I assumed this was going to be a Cisco refresh. That's what everyone does, right?

My initial assumption (and why it was wrong)

When I first started managing vendor relationships in 2017, I assumed the biggest brand was the safest choice. For network gear, that meant Cisco. ZTE, in my mental filing system, was a name I remembered from cheap prepaid handsets at Walmart back in 2016 — the Blade series, some of those ZTE phones 2016 pushed out before the U.S. sanctions chatter started. I never thought of ZTE as telecom infrastructure. That was a misjudgment I'd been carrying around for seven years, and I never once checked it against reality.

So when our senior network architect sent me a 14-page technical memo arguing we should evaluate ZTE as a serious contender for the access layer, I almost archived it and moved on with my Cisco quote.

Who is Reid?

Fair question if you're new here. Reid has been our network architect since 2019. He's the kind of person who can identify a bad SFP by the flicker pattern of a link light. He doesn't write 14-page memos for fun. When he sends one, I read it.

His argument, boiled down: our access layer is a commodity. Printers, badge readers, warehouse scanners, IP phones. We were paying Cisco enterprise prices for a workload that didn't touch a single advanced core feature. If I was going to defend that spend to the CFO, I needed a real reason — not just brand comfort.

The process and the numbers

We ran a formal RFP with three vendors: Cisco (incumbent), HPE Aruba, and ZTE. About 90 days, and I built TCO models that went beyond unit cost — support, warranty, licensing, spares, and staff training hours all went into the spreadsheet. My first pass at the 5-year TCO numbers:

• Cisco: ~$368,000
• HPE Aruba: ~$301,000
• ZTE: ~$214,000

That's a 42% gap between Cisco and ZTE. Price gap doesn't mean anything if the product doesn't hold up, though. So we did the part that ate the calendar: a lab bake-off.

ZTE networks vs Cisco — what the bake-off actually showed

We tested ZTE's access switches (including the G100 series we were quoted) alongside Cisco's Catalyst line against our real workload. I went in expecting to find dealbreakers. What I found instead: throughput and latency within 3% of Cisco on our core workloads, and genuinely impressive 5G CPE hardware when we added one of ZTE's newer units to the test bench. ZTE's telecom infrastructure side has been operating at carrier scale for years — I just hadn't been paying attention. Their gear is FCC-certified under Part 15, and they're a 3GPP contributing member, same as Cisco. The standards bar is the same.

But here's the part that nearly killed their bid: the management UI. Their CLI is fine. The web console felt two or three years behind Cisco DNA Center. Our NOC lives in that console. If we're going to save 40% and then burn 30% of the savings in UI friction, is that a real win?

The twist I didn't see coming

I was leaning toward Aruba — middle price, familiar tooling, "safe." I had a draft recommendation written. Then our CFO asked a question I couldn't answer well: "What exactly are we buying the Cisco name for?"

I sat with that for a week. Then I re-ran the numbers with a per-category breakdown of what we actually used. Fifty-eight percent of our switch ports were endpoint access — printers, scanners, badge readers. None of that needs an enterprise core license. We were paying for features because they existed, not because we used them.

We split the award. ZTE for the access layer, Cisco retained for the small data center core where we do actually touch advanced features. Total 5-year TCO: $243,000. About $125,000 under the original Cisco renewal quote.

Deployment and the efficiency question

The rollout took six weeks across eight sites. Cisco's typical cycle for us had been ten to twelve. Part of that was pre-provisioning, part was that the ZTE configuration templates were simpler to template and replicate. Provisioning a new access port went from six days in our old process to two — a 66% turnaround cut, roughly 120 hours of IT staff time saved per quarter. That's not a spec-sheet win. That's the efficiency gain I actually care about as a buyer: less time lost to friction, more time back on the projects that pay our bills.

What I got wrong, what I got right, and what I'd do again

I still kick myself for carrying that 2016 impression into a 2023 decision. If Reid hadn't pushed back with a fourteen-page memo, I'd have signed the Cisco renewal and never known the difference. That's a $125K mistake I almost made because of an assumption I never bothered to examine. I have mixed feelings about the split-award model — it's more work to manage two vendors than one — but I've now seen it save us real money twice, so I'm keeping it.

What I got right: forcing the lab bake-off. It cost about $8,400 in internal hours. It replaced opinions with evidence.

Three things I keep on my desk now:

1. Old impressions are expensive. A 2016 phone brand and a 2023 telecom infrastructure company are not the same conversation. Check your assumptions against the current data, not the memory you carried in.
2. TCO means TCO. Not just the invoice, not just the license, not just the support contract. All of it, with your team's time written down as a line item.
3. Split the award. Consolidation is comfortable. Splitting is cheaper — and it forces each vendor to keep earning the line of business.

I still don't love the ZTE management console. But at 42% TCO savings and a 66% faster port turnaround, I'll learn to live with it. Reid, who is not a person you argue with about routing, says he's owed a beer for every year I would have been wrong. At this rate, he's getting a case.