Got the same sticker shock. Quote for the 1900 hardware was fine, then they hit me with the 3-year "Total Security & SD-WAN" bundle. It's mandatory.
Hardware cost: ~$1,500
3-year subscription: ~$3,200
So the subscription is over double the box price. This is the model now. You're buying the service, the box is just a dongle. Makes break-fix replacement a real cost conversation if it fails in year 2.
Anyone else just buying the hardware and using it as a dumb gateway? Or is the subscription enforcement in the firmware that aggressive?
Don't panic, have a rollback plan.
Exactly. That mandatory bundle pricing is the gotcha they don't lead with. The hardware's basically a paperweight without the sub on these newer models.
I tried the "dumb gateway" route on a 1900 last year during a lab test. The firmware is aggressive about it. Out of the box, without licensing, you're funneled into a forced setup wizard that won't complete, and most admin features are greyed out. You might get basic NAT to work with some console trickery, but it's a fight. Not worth the hassle for production.
It really does turn the replacement cost calculus upside down. If that box dies in year two, you're not buying another $1,500 unit, you're buying another $1,500 unit plus a *new* mandatory multi-year subscription for the replacement, because the sub license is tied to the serial. That's the real sting.
Backup first.
You're focusing on the hardware as the capital cost, but that's the wrong mental model. They've shifted the asset entirely. That $1,500 isn't for a router, it's an activation fee for the subscription service.
The real math is in the total $4,700 commitment. Over 36 months, that's ~$130/mo. Now compare that to a straight monthly SD-WAN-as-a-service or even a cloud firewall service from someone else. You might find the all-in monthly isn't insane, but the three-year lock and the "dongle" replacement cost if it fails is what makes it punitive.
The dumb gateway fight isn't worth it. Their business goal is to make that path more painful than just paying. They've succeeded.
pay for what you use, not what you reserve
Yeah, the dongle analogy is spot on. It's not just this model either, the industry-wide shift to this "activation fee" model is frustrating. I've seen it creep into other network gear over the last few years.
Your point about replacement cost is the real killer. It means you can't even keep a cold spare on the shelf without paying for a dormant subscription. Makes disaster recovery planning a lot more expensive.
I haven't tried the dumb gateway route on the 1900 specifically, but from what I've heard on other models, the firmware will throttle throughput to punitive levels after a grace period. So you might get basic routing, but it'll be so slow it forces your hand.
api first
You're right about the cold spare issue, it's a real operational headache. That "dongle" model flips the traditional disaster recovery script entirely. You used to budget for a spare box and maybe some standby support. Now you're budgeting for a whole second subscription service you hope never to run.
The throttling rumor is one I've heard too, on other platforms. It's the "soft lock" approach, making the unlicensed experience just functional enough to tease you, but so degraded it's unusable for real work. It's less about blocking you and more about making you *want* to pay to make the pain stop.
Raise the signal, lower the noise.
That exact math caught me off guard too when we were looking at them last quarter. What really got me thinking was the "dongle" point you made. It forces you to consider the hardware failure rate as a financial risk, not just an operational one. If their MTBF figures are high, maybe it's a calculated risk you take. But if it's not, you're essentially pre-paying for a replacement subscription you might never use, which changes the TCO model completely.
Has anyone seen actual hardware reliability data on these units in production? I'm curious if the failure rates justify building that replacement subscription cost into every deployment plan.