25 Years of CCIE

Cisco sent me a jacket this year for hitting 25 years on my CCIE. If you're reading this, I'm guessing you're not here for the jacket — you're here because you're weighing whether to go after this thing yourself, and trying to work out if it's still worth it.

Here's the honest answer: I can't tell you yes or no in the abstract. But I can walk through how I'd think about it, because there are really three variables in play — what the certification is worth in today's job market, where you are in your career, and what AI is doing to the skill the exam actually tests.

Career timing. This may not be a popular opinion given the CCIE is an expert-level exam, but I think certifications earn their keep in the first third of a career — assuming you're planning something like 30 to 40 years in this field. If you've been doing this job for 20-plus years and you're already the expert in the room, chasing a piece of paper to prove it has diminishing returns. The one scenario where it still moves the needle is a dead heat — you and someone else up for the same role, one of you holding the letters and the other not.

I lived a version of this. At Juniper Networks, nobody seemed to particularly care that I held a CCIE. I've also held the Juniper equivalent — the JNCIE — since 2004, and even that one, I suspect, did more work getting me the interview than it ever did once I was in the building.

Market value. This part is simpler. I was looking at a job posting recently for a role at a company that competes directly with Cisco. That company runs its own well-regarded certification program — you'd expect the posting to ask for it.

It didn't. The preferred credentials listed were CCIE and JNCIE. Not their own.

Sit with that. A competitor built an entire certification track of its own and still reached for a rival's credentials as shorthand for "this person can do the job." That's not nostalgia talking. That's a hiring manager with zero incentive to flatter Cisco or Juniper, deciding those letters are still the fastest way to say someone can actually do the work.

AI and the CCIE

Caveat all of this with the fact that I sat for my lab exams in 2001 and 2009. But from what I understand, today's lab still comes down to how fast a candidate can make a design decision and then configure devices from a command line, under a clock. Plenty of candidates run out of time — which means the exam is, in no small part, a test of how deeply the CLI syntax lives in your psyche.

Will that particular skill matter in ten years?

A decade ago, the going theory was that automation would have already killed it. Automation has come a long way since, and the CLI is still very much alive. But watching AI models operate in the first half of 2026 has changed my answer. Memorizing thousands of syntax combinations is starting to look like a skill on its way out. I got real proof of that the first time I had Claude Code SSH into a router in my lab to troubleshoot and adjust a configuration for a problem I needed to reproduce. It's hard to describe in writing — you have to sit in front of it and watch it happen. It's like science fiction.

For what it's worth, that same session also reassured me a little. I spent a fair amount of it arguing with Claude. Some of its ideas were bad. It went down more than one rabbit hole, burning tokens I was paying for, and I had to keep pulling it back on track. So my working assumption is that we'll still need humans steering the tools for a while yet. And if the CCIE process builds the kind of deep protocol understanding that lets you do the steering, it holds its value in the AI era too.

Blueprint, not brand. One way to assess the potential value for you: study the blueprint for the specific track you're considering, not just the certification's reputation.

After I passed the old Route & Switch lab exam in 2001, I was nowhere near a BGP expert — and I wasn't remotely ready for my first Juniper lab exam, the JNCIP-M, which leaned hard on BGP. Juniper's exam was built for service providers, so BGP carried real weight. Cisco's enterprise-focused blueprint, by contrast, buried BGP as one of dozens of protocols in scope.

If both blueprints looked the same today, with AI advancing the way it is, I think I'd have picked a different track. You want the learning process itself to shape how you think about design, engineering, and operations.

TL;DR: If you're less than 10 years into your career in networking, my advice is simple: pick the hardest target you can find and go learn it. CCIE, JNCIE, or something else. Just go do it.