by Alex Stewart

The Commerce Commission paid an international expert to assess New Zealand’s telecommunications regime. He warned that the rural market was heading toward a monopoly that the Government would soon find itself unable to regulate. It published his assessment, disclaimed it, and did nothing else with it. By then, officials inside the ministry responsible for policy had been tracking the same dynamics for over eighteen months.

The Rural Blind Spot is a three-part independent investigation into what New Zealand’s government knows about rural broadband monopoly risk, and what it does not.

Part 1: “Nobody Checked” covered the absence of analysis at the two agencies with direct responsibility. This week’s Part 2: “They Were Warned,” covers the risks that officials documented internally but chose to set aside. “Part 3: In Confidence” covers the consolidated dependency, the assessment the Defence Force did for itself, the work nobody did for civilians, and the advice the Minister holds but refuses to release.

About me: I am Alex Stewart, managing director of WombatNET, a rural wireless broadband provider that has never received direct government funding and is a direct competitor to Starlink in the current market.

Three positions, no coordination

On 21 August 2025, the Commerce Commission delivered its final recommendation to the Communications Minister on copper deregulation[1]. Telco Commissioner Gilbertson’s reasoning, quoted in the recommendation, was that most rural copper premises now had access to at least three alternative technologies that were often more affordable, reliable, and better performing than copper: “national mobile, regional wireless (WISP) and global satellite providers[2].”

For context, this recommendation did not arrive in a vacuum. Chorus had already committed publicly to retiring its copper network, with the original retirement window set to close in 2030[3]. The regulatory question was not whether copper would be retired, that had already been decided, but whether the rural alternatives were sufficient to support deregulation in advance of retirement.

Two months earlier, another government agency had reached a different conclusion on a related question. In June 2025, the Infrastructure Commission, Te Waihanga, endorsed a Chorus proposal to extend fibre to 95 per cent of New Zealand, up from the current 87 per cent, as a national priority under its Infrastructure Priorities Programme. The private proposal received a green rating for value for money and a benefit-cost ratio of 6.3, higher than most public infrastructure projects. However, in the same assessment, the Infrastructure Commission marked the proposal down for not aligning with a national rural connectivity strategy. Chorus CEO Mark Aue’s response was direct. “Our proposal was marked down for not aligning with a strategy that doesn’t exist.[4]“

It is worth pausing on each of the three alternatives Gilbertson had named. National mobile, delivered as FWA in rural areas, is often characterised by variable performance, signal fluctuations, and speed degradation during peak periods. These are not simply my own assessments. They are the descriptions Chorus itself used in its September 2024 national advertising campaign, which ran across television, radio, outdoor, and digital channels specifically to warn consumers away from wireless broadband in favour of fibre. The campaign quoted wireless providers’ own fine print back to them under the tagline “Avoid this. Get fibre.”

One of the other alternatives, regional wireless, meaning local WISPs, is the sector this series is about, and the one writing it. Its position under the spectrum and capital settings the government had set was already tenuous. The third alternative, global satellite, at this point, means Starlink.

Eleven months later, the Commerce Commission recommended to the Minister that rural copper deregulation proceed on the basis that regional wireless was one of three sufficient alternatives available to most of those switching from copper.

Forty-six days later, on 6 October 2025, the Commerce Commission published an independent report by Richard Feasey. Feasey described Starlink as heading toward a monopoly that the Commission would be unable to regulate. He described wireless internet providers as unlikely to represent significant future competitors under current settings.

Two of the three alternatives on which the August recommendation was based had, within forty-six days, been described in the Commerce Commission’s own published record as either already non-competitive or heading toward an unregulable monopoly. The other cannot be reasonably relied upon for critical operations, as the country’s largest network operator pointed out. Both positions remain live alongside each other in the Commerce Commission’s published record.

Who Feasey is and what he said

Richard Feasey is not your local pub punter. He has spent more than 30 years working in telecommunications regulation across Europe. He is a panel member for the UK Competition and Markets Authority, which is the equivalent of New Zealand’s Commerce Commission[5]. He was also the Public Policy Director for Vodafone Group UK[6]. If you are a regulator looking for an independent external assessment of whether your telecommunications framework is fit for purpose, he is the kind of person you hire.

So, the Commerce Commission picked him. Paid for his work. And published his report on its own website.

Feasey’s central finding on rural broadband was direct. He concluded that there was “a material risk that Starlink would win the competition ‘for the market’ and become the dominant if not monopoly supplier of broadband services in rural areas[7].“ The phrase “competition for the market” is a term of art. It describes a pricing strategy that clears the field of competitors before consolidating into monopoly pricing. If that occurred, he wrote, “it would be very difficult for the Commission to regulate Starlink or otherwise to ensure competition in the provision of rural broadband services[7].”

He described Starlink and Kuiper (now operating as Amazon Leo) as more closely resembling global digital platforms like Amazon, Google, and Meta than traditional telecommunications carriers[8]. He noted that neither had been designed with regulatory obligations in mind. He flagged “the extra-territorial nature of Starlink’s operations and the resources and mercurial personality of its controlling shareholder” as risk factors the Commission would have to address[9]. He specifically recommended that the Commission or the government act “as a matter of urgency” to address these concerns.

He also described wireless internet providers, directly, as unlikely to represent significant competitors in the future. That is a sentence I have no personal interest in promoting. It is also a finding that directly undermines the Commission’s August copper deregulation recommendation, which had relied on that same sector as one of three credible alternatives.

What the Commission did with it

The Commerce Commission published Feasey’s report on the day it received it. In the same notice, it stated that the contents of his report did not represent the views of the Commission. It added that it was publishing the report “in the interests of transparency” and was “not seeking views in response.”

It is not unusual for a regulator to caveat an independent report. What is unusual is the silence that followed.

No follow-up analysis was initiated. No staff were assigned to examine Feasey’s conclusions. No consultation was opened. No referral was made to the Minister requesting guidance.

When I asked the Commission under the OIA what analysis it held on the monopoly risk Feasey had identified, the response pointed me back to the Feasey report. The document warning that the Commission needed to act urgently was the document that the Commission offered as proof that it had already studied the issue. A warning that analysis is needed is not analysis.

If you appoint and remunerate an international expert to tell you whether you have a problem, and the expert tells you that you have a serious one and recommends urgent action, the usual next step is to do something, or at least start exploring your options. The Commerce Commission commissioned the report, paid for it, published it, disclaimed it, and then returned to its previous position. The August copper deregulation recommendation was not revisited. The competitive alternatives on which the recommendation relied were not reassessed. The monopoly risk Feasey had identified did not appear in any Commission documents subsequently released or published.

The Commission’s own expert had identified a material risk and recommended urgency. The Commission chose to treat his conclusions as something to publish rather than something to act on. Just four days after the report appeared, Evolution Networks, a wireless broadband provider serving rural communities in the Bay of Plenty, entered voluntary liquidation[10]. It was the first local operator to fail since Starlink entered the New Zealand market at scale. The liquidator’s first report cited a lack of profitability and operating cash flows. Industry commentary at the time attributed the failure, in significant part, to Starlink’s pricing and delays in spectrum policy[11].

What officials had already written down

If the Feasey Report were the only evidence that risk had been identified, this series would end here.

It is not.

Documents released under a separate OIA to MBIE show that officials inside the ministry had been tracking Starlink’s market impact for over eighteen months before the Feasey report was published. A weekly report to the Minister for Media and Communications in March 2024, nineteen months before Feasey, logged Starlink’s introduction of a $79 deprioritised service and the WISP industry’s concern that operators would find it difficult to remain commercially viable at that price point[12].

By May 2024, internal MBIE slides had gone further. They state, directly, that “the satellite market is currently a monopoly for mass market residential connections.” The same slides raise questions officials were actively thinking about. Whether Starlink’s $159 standard price and $79 deprioritised price were introductory. Whether Starlink would engage in geographic price discrimination, as it had already done in parts of the United States. Whether service quality would hold as the customer base grew[13].

An MBIE internal slide from May 2024, released under OIA DOIA-REQ-0025973. The visible text records, directly, that the LEO satellite market is a monopoly for mass-market residential connections. Redactions are applied under sections 9(2)(g)(i), 9(2)(ba)(i), and 9(2)(f)(iv) of the Official Information Act.

An undated MBIE draft brief, also released in the same OIA, is even more explicit. It states that there are “inherent price and resilience risks that would be associated with New Zealand becoming over-reliant on a single private provider” and that Starlink “has a track record of increasing prices over time, which could disadvantage consumers, especially in lower-income rural communities.” The talking points attached to that brief are withheld under section 9(2)(f)(iv)[14].

Officials inside government had identified the monopoly risk. They had identified the pricing risk. They had identified the dependency problem. They had written it down. They were continuing to write it down.

Yet, in the same ministry, in the same month, a separate OIA asking for the formal analysis of those exact concerns returned blanks under section 18(e). MBIE conducted no formal analysis of the economic or resilience implications of declining terrestrial broadband in rural areas. It has not advised Ministers on whether infrastructure diversity serves resilience. It had not considered path dependency[15].

The observations existed internally. The translation into formal analysis did not.

A Treasury proposal MBIE pushed back on

Sitting inside the same MBIE bundle is a Treasury-MBIE email chain from October and November 2024.

A graduate analyst at Treasury’s National Infrastructure Unit emailed MBIE on 1 October 2024[16]. They were preparing a briefing for Infrastructure Minister Chris Bishop on whether national rural broadband funding could be repurposed toward Low Earth Orbit satellite solutions. The framing in their email was that this “will dramatically reduce costs to deliver rural connectivity and shorten time to delivery by years.” It could provide, they suggested, “a short path to significant economic growth in rural NZ.” They wrote, “For me, without looking at numerical calculations, it does make sense that Starlink…” and the sentence trails into redaction.

MBIE pushed back. The response noted that RBI funding had already ended, so no appropriated funds were available to repurpose. It then went further. “Starlink currently has a monopoly on mass market satellite broadband services, and has a track record of increasing prices over time. From a policy perspective, we would like to see multiple satellite providers active in the NZ market to manage company specific risks and ensure long term positive outcomes for consumers.”

The email chain itself was eventually released through my OIA. So were the May 2024 slides and the undated draft brief. What was not released, because it does not exist, is the formal analysis behind any of it. The monopoly observation was made in passing to Treasury. It was not written up. It was not modelled. It was not advised on. The concern made it into an inbox. It did not make it anywhere near the public record^*.

The spectrum policy that the industry warned about almost five years earlier

The pattern in the Commerce Commission’s handling of Feasey is not unique to the Commission. A near-identical sequence plays out in rural spectrum policy, two years earlier, inside a different agency.

Radio spectrum is what roads are to a transport operator. Without it, you cannot reach the customer. One piece of rural broadband policy the government did execute during this period was the 3.3 to 3.34 GHz regional broadband spectrum allocation[17]. It was executed in a form that the industry had earlier warned in writing would not work[18]. RSM’s own internal review afterwards agreed with the industry[19]. The management rights run until 30 June 2033, meaning technical changes are not possible until then.

In March 2022, Cabinet approved 100 MHz of spectrum in the 3.3 GHz band for regional broadband and private networks[20]. The spectrum was intended to enable local wireless operators to extend coverage and capacity into rural communities.

Before the allocation was finalised, the industry told RSM what would work and what would fail. In August 2021, WISPA NZ’s technical spectrum working group stated that, while the full 100 MHz was preferred in order to further increase capacity, 80 MHz was the minimum amount of spectrum required to deliver the rural broadband performance objectives set by The National Infrastructure Agency (then called Crown Infrastructure Partners), the Government’s own infrastructure investment vehicle. The same submission warned that the synchronisation pattern being proposed by MBIE would limit communications to a maximum cell size of around 10 kilometres, far below what WISPs need to serve rural customers[18]. When cell sizes are limited, operators are forced to build more towers to cover the same area, thus significantly increasing costs and discouraging, rather than promoting, investment.

Jonathan Brewer, author of the Asian Development Bank’s working paper on Last Mile Connectivity and radio engineer for more than 40 utilities and network operators in New Zealand, submitted a specific technical configuration based on European deployments that would have preserved the 15 to 20 kilometre cell coverage radius rural deployments realistically require[21][22][23][24].

When the process concluded, just 40 MHz was assigned, half of what the industry advised was the minimum required[17]. The remaining 60 MHz has sat unallocated and unutilised since it was set aside in 2023. The 40 MHz was configured with technical parameters that made 5G networks with a cell radius of around 10 kilometres the default for the band, the exact configuration WISPA NZ had warned would not be practical for rural deployment.

RSM proceeded anyway. The March 2023 consultation that finalised the technical parameters was not public. No submissions were released. Two ministerial briefings went to Ginny Andersen, the then Minister for the Digital Economy and Communications. Her handwritten note on the first asks whether “there is an estimate of what the additional uptake and coverage of rural broadband would be if the proposed model progresses[25].”

Hon Ginny Andersen, Minister for the Digital Economy and Communications, on briefing 2223-2369 (10 March 2023): “Is there an estimate on what the additional uptake / coverage of rural broadband would be if proposed model progresses?”

The follow-up briefing did not answer that question in writing[26].

RSM’s own internal post-allocation review, conducted sometime in 2024 and released under OIA, plainly recorded the assessment. Under the heading of whether the process achieved its primary rural connectivity objective, the note reads: “This probably has not significantly increased rural connectivity as the activity has been around urban centres.” Another line reads: “We have delivered the bit we could control but it is then down to the industry to deliver on the ultimate part of this policy goal[19].”

$82,416.98 was expended in external legal and auction costs alone as part of the process. RSM have stated they are unable to determine internal staff expenses as they specifically relate to this work program[27].

The industry clearly warned before it was finalised that the allocation, as configured, would not deliver the expected outcomes. RSM’s own review afterwards said it had not worked. The next opportunity to fix it is seven years away. Further work on the continuation of this band is planned for this year, but it is not a policy response to what has happened since. It is simply the next opportunity for the same system to reach the same foregone conclusion.

The scale of the spectrum asymmetry

What 40 MHz of shared spectrum means in practice is best understood by comparison with what the platform now displacing local operators uses. A single Starlink satellite beam serving a rural area uses up to 240 MHz of spectrum. New Zealand’s entire WISP sector shares 40 MHz of regional broadband-specific spectrum between all operators. Many satellites pass over New Zealand at any given time, each carrying 48 individually steerable beams[28].

Starlink user terminals (dishies) operate in New Zealand under RSM’s General User Radio Licence for Satellite Services. That licence requires no individual application and carries no spectrum fees[29]. The satellites themselves operate under ITU coordination via SpaceX’s global FCC licence, outside RSM’s management rights framework entirely[30][31]. The seven ground stations Starlink operates in New Zealand hold individual RSM transmitting licences, at a cost of approximately $10,000 per year combined[32].

The three mobile network operators collectively paid $259 million for 45 MHz of the 4G spectrum that underpins rural mobile coverage[33]. Domestic wireless operators paid for their 3.3 GHz allocations through a standard licensing process[17]. The foreign provider replacing both pays nothing for spectrum access to New Zealand users. Its downlink rights come through longstanding international arrangements at zero cost to the operator. That describes which kind of provider the current settings favour, either by design or by neglect.

The cost of these settings extends beyond just the WISP sector. Shared spectrum frameworks in comparable jurisdictions have enabled ports, mines, hospitals, universities, transport hubs, councils, and agricultural and industrial operators to deploy their own private wireless networks, boosting efficiency, performance and productivity. Australia has issued at least 79 area-wide licences for private networks since 2023[34][35]. Germany has issued more than 430 in the 3.7 to 3.8 GHz band alone[36]. The United Kingdom has issued more than 1,600 Shared Access licences, supporting users from manufacturing to social care[37]. New Zealand’s older framework, the 2.5 GHz Managed Spectrum Park, has produced one private network, a small link between Christchurch City Council buildings; the rest are public-access wireless internet service providers[38][39]. The newer 3.30 to 3.34 GHz allocation, the framework that replaced it, was assigned at half the bandwidth the industry had warned was the minimum required, and excludes private network use entirely[17].

RSM’s stated objective is “to ensure radio spectrum provides the greatest economic and social benefit to New Zealand.[40]” Under the current settings, a New Zealand port wanting to run its own 5G network for crane coordination, a mining operation wanting safety-critical wireless underground, or a hospital wanting dedicated capacity for clinical equipment must partner with one of three mobile network operators. There is no fourth option, as exists in countless other jurisdictions. Three commercial providers selling the same constrained product to a captive market is not the same as a market in which private enterprises can choose to build their own infrastructure.

None of this is a measure of what enterprises or local operators are capable of. It is simply a measure of what the regulatory settings permit them to attempt.

In March 2025, WISPA NZ’s then Chairman Mike Smith warned publicly of “a serious risk of operators closing up shop if spectrum that is sitting idle is not made available soon[11].” Seven months later, the warning held[11].

When Feasey described WISPs as unlikely to represent significant competitors in future, he was describing the consequence of these arrangements, not a limitation of the technology or the operators themselves.

What this all adds up to

The pattern across these documents is consistent.

The Commerce Commission paid for an expert who identified a material risk and recommended immediate action. The Commission published the warning and then distanced itself from it. No follow-up analysis was conducted, and no revision of the adjacent copper deregulation recommendation, which it contradicted, was made. A predatory pricing complaint sits idly before the Commission, lacking an analytical framework to assess it.

MBIE officials had identified the monopoly and pricing risks internally since at least May 2024 and had put those concerns in writing to Treasury by October 2024. A different OIA to the same ministry, covering the same period, returned blanks on the same subject matter.

A spectrum allocation Cabinet approved in 2022 was implemented in a form the industry had warned would fail, and which RSM’s own review concluded had not worked. Meanwhile, the overseas competitor displacing the domestic industry operates with roughly six times the spectrum capacity per beam, per satellite, at no cost.

The government has invested substantially in rural broadband infrastructure over the past 15 years[41]. By 2026, the networks built through that public investment are being rendered progressively redundant by a foreign provider that contributed nothing to build them and has no desire or legal obligation to maintain what it is replacing.

None of this was unknown inside government. Officials had named the risks. An independent expert, commissioned by the regulator, had named them more sharply. The one thing that did not happen at any point over the roughly two years of active internal awareness was the translation of that awareness into even the most basic formal analysis or action.

Next week, the final part in this series covers the consolidated dependency now running on a single foreign platform, the assessment the Defence Force did for itself, what nobody did for civilians, and the advice the Minister holds but refuses to release.

Read Part 1: “Nobody Checked”. Read Part 3: “In Confidence”

The Rural Blind Spot. An independent-of-government investigative series by Alex Stewart.

References