Why mmWave Failed to Scale in AustraliaWednesday, 2026-Mar-4

Five years ago, industry leaders like Qualcomm touted millimeter wave (mmWave) as a game-changer, while Analysys Mason—a consultancy for Ericsson and Qualcomm—predicted significant economic benefits. Unfortunately, mmWave has yet to deliver the breakthrough Australians were promised, after growth stalled more than two years ago.

Millimeter wave (mmWave) spans 24-100 GHz and promises ultra-fast wireless speeds. Deployment in Australia began in 2021 with Vodafone at 25.1-25.7 GHz, Telstra at 25.7-26.7 GHz, Optus at 26.7-27.5 GHz, and NBN at 28.5-29.5 GHz.

Active Licensed Sites
3.4 and mmWave (26 GHz) spectrum adoption (Cellular) shows 3.4 growing strong from 2018 to present, and mmWave growth stalling in 2023
Source: ACMA Radiocomms licence data (to 2026-Mar-4)

mmWave Deployment Falls Far Behind Mid-Band 5G

The promise was bold: thousands of mmWave sites lighting up cities from Perth to Sydney. The reality? Just 1,155 nationwide.

Comparing Growth: 3.4 GHz vs. mmWave

Telstra, Optus, and Vodafone's 3.4 GHz deployments vastly outpace mmWave:

  • 3.4 GHz mid-band sites grew from zero to 15,029 nationwide (2018-26).
  • mmWave sites reached just 1,155 (2021-25).
  • In deployment year 5 (2025-26), mmWave site growth slowed to 1% annually, compared to 24% for 3.4 GHz in its fifth year (2022–23).
Annual Increase in Cellular Site Counts
Period3.4 GHzmmWave
2018-19453% n/a
2019-20350% n/a
2020-21127% n/a
2021-22 40%755%
2022-23 24%167%
2023-24 17% 34%
2024-25 13% 4%
2025-26 9% 1%

The 755% mmWave growth in 2021–22 was a low-base surge, not sustained momentum. Since 2023, growth has declined, hitting just 1% this year.

Why mmWave Deployment Stalled?

Potential Niche Applications for mmWave

While widespread adoption has stalled, mmWave remains valuable in tightly controlled or high-demand scenarios, such as:

These use cases suggest mmWave's future lies in targeted deployments, not mass-market mobile coverage.

Conclusion: mmWave Falls Short of Hype

Operators continue to prioritize mid-band (3.4 GHz) 5G, limiting mmWave to niche deployments. In March 2026, just 1,155 mmWave sites exist—dwarfed by 15,029 mid-band (3.4 GHz) sites, a 13-to-1 difference. For now, mmWave’s once-vaunted promise remains largely unrealized outside tightly controlled environments.

mmWave's long-promised revolution has instead become a specialized tool—powerful in pockets but impractical at scale. In Australia, mid-band remains the true workhorse of 5G.

Rebuilding Canada's Wireless Infrastructure DataFriday, 2026-Feb-27

We have updated our backoffice workflow that ingests ISED Spectrum Management System (SMS) data. The result is a stronger, cleaner, and more defensible national data foundation powering our reports and the Canada Cellular Services platform.

Innovation, Science and Economic Development Canada (ISED) manages Canada's radio spectrum. Licensees submit technical RF data into ISED’s Spectrum Management System (SMS), and ISED publishes those submissions as monthly downloads that appear to be national snapshots of Canadian spectrum deployments. They are not. ISED has publicly acknowledged the need to improve data quality, reliability and uniformity within SMS.

In practice, SMS is a cumulative record of submissions rather than a clean monthly replacement of what changed. Older records linger, and updates may appear late or not at all. Months with no visible change do not mean the network is stable — they often mean no new submissions were processed. The transactional UPDATE / APPEND / REPLACE nature of the submission process adds further risk. Treating each monthly file as a clean snapshot leads to unreliable conclusions. Serious longitudinal analysis must reconcile cumulative submissions, upload chronology, and transactional overwrite events.

Our updated workflow is built around that reality. We reconcile cumulative submissions into coherent state transitions, filter stale and superseded records, remove structural duplicates, resolve site relocations, detect anomalous overwrite events, and restore (backfill) emissions when valid spectrum is temporarily removed due to transactional errors. The result is an operational representation of Canada’s wireless infrastructure — not a raw administrative extract.

The workflow is now live and continues to be refined as we identify edge cases and resolve residual inconsistencies.

The ISED SMS files are free. Extracting reliable network intelligence from them is not.

For RF technicians, operators, and site acquisition or investment firms, the risk isn’t downloading the data. The risk is making decisions based on it at face value.