Quick answer: 3CX is a capable SMB PBX, and a team that wants telephony as a managed appliance with extensions on a call tree can still run it fine. The 2026 question is different: after the ownership changes that reshaped on-prem and 3CX Cloud packaging, the renewal a buyer re-signs is per-seat subscriptions plus an SBC license to carry their own carrier trunks plus a stack of plugin-based CRM integrations, and the honest comparison is what that stack costs against a platform priced on usage. Devotel Orbit is the different shape: SIP trunks that keep your carrier or replace it, call flows and IVR re-authored as configuration instead of a PBX call tree, AI voice agents on the same path, and a published per-minute bill. This post works the architecture, the cost model, the receptionist question, and the migration order.
The 2026 trigger: re-signing a PBX you no longer recognize
The 3CX evaluation of 2026 starts at the renewal, not at a feature gap. Three things changed the math for the installed base:
- Ownership and packaging changes. The ownership changes of the last few years moved 3CX through new hands and new go-to-market packaging, and each move re-papered the same estate: on-prem deployments, 3CX Cloud hosted instances, and the edition ladder between them. Buyers who signed under one owner now re-sign under another, and the upgrade path they were promised has been re-packaged more than once.
- Per-seat licensing plus SBC licensing. 3CX prices by concurrent users or seats per edition, and connecting your own carrier SIP trunks runs through session border controller licensing or a certified-SBC list. The bill is seats times edition plus the SBC layer plus the host you run it on if it is on-prem.
- Plugin-era CRM integrations. The CRM story is a plugin and CTI bridge layer: screen pops and click-to-dial arrive through per-CRM add-ons that track 3CX version support on their own clock, not through a native API surface a team can build against.
None of those is a failure. They are the cost structure of a PBX vendor, and 2026 is simply the year a lot of installed bases re-sign it and ask what the alternative shape costs.
Architecture: a PBX call tree versus SIP trunks, flows, and agents
3CX organizes telephony the way a PBX always has: extensions hang off a call tree, the call tree lives in an admin console, and the carrier leg terminates on an SBC in front of the box, whether the box is your hardware or 3CX Cloud. Every behavior, forwarding chain, office-hours rule, and ring group is an entry in that tree, administered per deployment.
Orbit inverts the stack. The carrier leg is a SIP trunk you bring or one Devotel terminates: bring-your-own-carrier keeps your existing carrier relationship and re-points its registration or peering at Orbit, while Devotel-terminated trunks route over Devotel's own wholesale softswitch. Above the trunk, call handling is configuration: a visual IVR and flow builder for routing, business hours, and fallbacks, extensions with softphone and hot-desking for the people side, and AI voice agents on the same path for the calls nobody should wait on hold for. There is no call tree to administer per site; the routing contract lives in flows that apply to every trunk and number on the account.
The practical difference shows up at version time. A PBX call tree migrates through vendor upgrades; Orbit flows are tenant configuration that does not change when the platform ships.
The cost model: seats and maintenance versus published usage
3CX's model is the PBX model: an annual per-seat subscription by edition, the SBC layer to carry own-carrier trunks, and, for on-prem estates, the host hardware and the maintenance window that comes with it. The bill scales with headcount whether or not the headcount makes calls, and the true-up conversation happens at renewal.
Orbit prices on usage: published per-minute voice rates per destination and per-message rates on the same account, one pay-as-you-go wallet, no platform fee, and no seat count on the voice path. A reception desk that answers 4,000 minutes a month and a sales floor that answers 40,000 pay for what they use, not for every seat in the building. The vendor-vs-vendor pricing register works the same usage-based frame across the wider market, and the pricing page publishes the rate table the bill is graded against.
The honest caveat: a team with very high, very steady minutes and its own carrier contracts can beat usage pricing with wholesale plus an engineer. That team is not the 3CX installed base; the installed base is paying seat licensing on top of wholesale either way.
Receptionist parity: AI receptionist and attendant console
3CX's answer to the front desk is a receptionist client and attendant console: a human operator with presence, transfer, and queue visibility. Orbit ships the same surface for humans and adds one for the hours humans are not there. The attendant console operator workspace covers the human side: presence across queues, transfer and hold controls, and the queue visibility an operator works from. The AI side is a voice agent that answers the main number after hours or at overflow, resolves the routine calls (bookings, directions, balance-of-account lookups against your data), and hands the rest to a person with the transcript attached.
The parity question is therefore not whether Orbit can put a human at a console; it does. It is whether your front desk needs to be staffed at 7 p.m. on a Saturday, and on Orbit the answer can be no without the call going to voicemail.
Migration: port the numbers, re-point the trunks, re-author the IVR
A 3CX exit is the same arc as any PBX exit — the Avaya PBX migration runbook walks the larger version of it for bigger on-prem estates. The 3CX cut is three moves in order:
- Stand up SIP trunking beside 3CX first. BYO-carrier re-points your existing carrier trunks at Orbit while 3CX keeps serving production traffic, so the new path validates on a test number before anything customer-facing moves.
- Port the numbers in waves. The number porting guide is the mechanics: bulk portability screening, in-platform LOA signing, and staged waves that start with a non-critical line and move the main number last. 3CX holds no number ownership your carrier will not release; the losing side of the port is the carrier, not the PBX.
- Re-author the IVR and call tree as flows. 3CX's call tree does not export into any other platform, and no migration tool changes that. Each office-hours rule, ring group, and forwarding chain becomes a flow, a business-hours rule, or a ring group on Orbit, re-authored from an export of the current behavior so a regression after cutover is diagnosable against a baseline.
The PBX retires after the last wave validates, not before the first one starts.
Where 3CX still wins
- Air-gapped and sovereign-media sites. A deployment where media must never leave your network, no external SIP allowed, is a PBX job by definition. Orbit is a cloud platform; the media residency requirement that bans cloud media also bans Orbit.
- Sunk hardware and perpetual deployments. An estate that already owns the host, the SBC, and a paid-up edition, with no renewal event in the window, has no trigger. This post is for the teams that do.
- The appliance preference. Some teams want telephony as a box a vendor packages, upgraded on the vendor's clock, with no configuration surface beyond the admin console. That is a legitimate operating preference, and it is the opposite of a programmable platform.
The win list is real and short: media residency, sunk cost, and appliance preference. Everything else on a 3CX renewal, seats, SBC licensing, plugin integrations, upgrade re-packaging, is exactly what the usage-priced platform absorbs.
Frequently asked questions
Is Orbit a good 3CX alternative for an SMB?
Yes, for an SMB whose trigger is the renewal rather than media residency. The comparison is a per-seat subscription plus SBC licensing plus plugin integrations against published per-minute rates with SIP trunking, flows, softphone extensions, and AI voice agents on one account. The voice alternatives hub runs the same question across the wider voice shortlist.
Can I keep my carrier when I leave 3CX?
Yes. Bring-your-own-carrier re-points your existing carrier SIP trunks at Orbit; the carrier relationship survives the PBX change. If you would rather consolidate, Devotel-terminated trunks route over Devotel's own wholesale softswitch. The choice is per tenant and does not change the cutover order.
Do my 3CX extensions and IVR import into Orbit?
The behavior imports; the configuration does not. 3CX's call tree and extension definitions do not export into any other platform, so each extension becomes an Orbit extension with a softphone, and each call-tree entry is re-authored as a flow, a ring group, or a business-hours rule, validated on a test number before cutover.
What about the attendant console and receptionist?
Orbit ships an operator workspace for human attendants with presence, transfer, and queue controls, and AI voice agents cover the hours a human console is unstaffed. Most 3CX receptionist deployments map to one of those two, or to after-hours overflow from the console to the agent.
Where does 3CX remain the right call?
On-prem sites where media must stay air-gapped, estates with sunk hardware and no renewal event, and teams that prefer a vendor-packaged appliance over a programmable platform. If any of those is your constraint, stay; the comparison only moves when the renewal does.
Where to go next
- The wider shortlist: voice alternatives and alternatives for 2026.
- The heavier PBX exit: the Avaya PBX migration runbook.
- The carrier and number mechanics: the SIP trunking BYO-carrier guide and the number porting guide.
- The human front desk: the attendant console operator workspace.
- The bill: pricing and the vendor-vs-vendor pricing register.
Published 28 October 2026.