Ookla, the network analytics firm behind the Speedtest platform, has published a commentary arguing that Voice over New Radio (VoNR) does not yet consistently deliver a better calling experience than the voice architectures it is designed to replace in 5G networks.

The piece, titled "The 5G Voice Paradox: Why VoNR Isn't Always Better (Yet)," challenges the assumption that moving voice traffic natively onto the 5G radio interface automatically improves quality. Ookla's analysis lands while operators across several markets continue to mix deployment modes, with subscriber devices often dropping to 4G LTE or to circuit-switched 2G/3G paths for voice.

What does Ookla mean by the "5G voice paradox"?

Ookla frames the paradox as the gap between VoNR's design promise and the real-world experience users see today. In standalone 5G, voice can travel end-to-end over the 5G NR radio and the 5G core. In non-standalone (NSA) 5G, where the 5G NR runs alongside an LTE anchor, voice typically stays on VoLTE or falls back to LTE entirely through the EPS fallback mechanism.

The Ookla commentary points out that because VoNR requires both 5G NR coverage and a 5G standalone core, real-world availability remains patchy. Where VoNR sessions are short, fragmented, or constantly handed back to LTE, the user perceives no upgrade — and may encounter longer call-setup times or dropped sessions during transitions.

How does VoNR differ from VoLTE and EPS fallback?

VoNR (Voice over New Radio, also called Vo5G) is the 3GPP-defined IMS voice service carried natively over a 5G standalone access network. It supersedes VoLTE, which runs over 4G LTE, and EPS fallback, which keeps voice on LTE while data uses 5G NR.

The practical difference for the operator is spectrum and core efficiency: a VoNR call keeps traffic inside the 5G system rather than requiring a 4G anchor. The practical difference for the user is, in principle, faster call setup and the ability to combine voice with enhanced 5G features such as network slicing or low-latency applications.

Why has VoNR rollout been cautious?

VoNR depends on three conditions aligning: a deployed 5G standalone core, sufficient 5G NR coverage to keep voice sessions alive, and device support across the subscriber base. Ookla's underlying point is that even where these conditions exist, the perceived voice gain over VoLTE has so far been marginal.

This explains the slow commercial rollout. Several Tier-1 operators have launched VoNR in limited markets or in single-mode 5G SA islands, while keeping VoLTE as the default voice path. Vendors and chipset makers have added VoNR support in waves, meaning older handsets still default to 4G for voice regardless of the network in front of them.

What should operators watch next?

Ookla's title — with its "(Yet)" — implies the verdict is provisional. The economics of VoNR will tighten as more operators retire 3G and 2G circuit-switched cores, removing the legacy fallback that has propped up voice continuity.

Expect the next phase of the debate to be measured, not declared. Voice-quality benchmarks drawn from Ookla's Speedtest Intelligence dataset, device-by-device IMS attach statistics, and operator-by-operator SA coverage maps will all shape whether VoNR moves from "technically available" to "consistently preferred" over the rollout cycles ahead.