Starlink fills gaps but doesn’t fix NZ rural coverage

Looked at from one perspective SpaceX’s Starlink could be the missing piece of New Zealand’s rural broadband jigsaw puzzle.

The Commerce Commission thinks it is. This week it said its “latest Measuring Broadband New Zealand report shows the potential of low Earth orbit satellite (LEO) broadband internet as an option for the 13 per cent of Kiwis who are not able to access fibre.”

There is no question that Starlink reaches areas beyond the reach of today’s fibre and the better performing fixed wireless towers. It outperforms legacy copper DSL broadband connections and the more feeble rural fixed wireless offerings by a substantial margin.

Starlink doesn’t require much infrastructure. It has ground stations, but that’s nothing compared to the UFB or RBI networks with fibre cables and cellular towers.

Impressive

There are people who think having the internet beam in from space is glamorous. Other people are upset about their view of space being disturbed by those satellites.

Regardless of these points of view, Starlink represents an impressive technical and engineering achievement. Getting hundreds of satellites into space and developing small, consumer-grade dishes were impressive feats.

Starlink needed vision and an heroic risk appetite to get the project off the ground.

Many New Zealand Starlink customers are happy with the service. A handful are delighted. Many customers have contacted me to tell me how happy they are. Only a single unhappy customer has been in touch. It’s hard to write about rural New Zealand telecommunications issues without someone explaining that Starlink has the answers.

Starlink’s promise

Starlink’s promise is that it delivers fast, affordable broadband from the skies. There is no need to dig trenches across paddocks or string fibre cables on power lines.

Nor do customers in more remote, hilly parts of the country need to worry about what sits between their home and the nearest fixed wireless tower.

There isn’t the nonsense you get with older satellite technologies where you need to aim a dish near the horizon - which can be a struggle near mountain ranges.

Starlink bypasses many obstacles. It does this by offering a large number of satellites. The idea is that there will always be at least one available satellite in a section of the sky that your dish can see. This isn’t perfect yet, people still experience short outages, but I’ve yet to see many Starlink customers complain about this.

If demand rises, Starlink can launch more satellites. It could end up with tens of thousands.

For a section of the population, Starlink is without doubt the best way of getting a decent connection.

Shortcomings

After reading about Starlink’s positives, let’s look at the negatives. The biggest downside is cost. At the time of writing, a basic Starlink plan is NZ$160 a month for New Zealand customers.

Put in simple terms, New Zealand Starlink customers get a fifth of the performance of a fibre max connection for twice the price.

On top of this, unlike fibre for the majority of New Zealanders, there’s an additional upfront cost: you have to buy the dish.

Deep discounts

At the time of writing the dish is temporarily discounted to NZ$200, it was previously $520, a year ago it cost $1040.

Officially the $200 dish price is “for hardware in rural areas”, which may be a form of price discrimination that would concern the Commerce Commission.

For the dish price to be that low, it has to be subsidised.

When Starlink was new, stories appeared in the US media saying the dishes cost close to US$3000 to manufacture. That cost will have fallen, economies of scale kick in and manufacturers are able to refine designs for cost savings over time. Yet, however you cut the numbers there is a substantial subsidy in a $200 dish that has to be recovered from the monthly subscription.

Starlink’s prices look like a bargain for well-heeled customers who are off the fibre map and can’t get a decent fixed wireless broadband connection. It may prove too expensive for less well off rural families.

As mentioned earlier, many rural Starlink customers are delighted. Yet, over time, they may feel disgruntled that they pay a hefty premium for, when compared to fibre, a second class service.

Starlink for business

If you think the characterisation of Starlink costing twice the price for a fifth of the performance is harsh, consider this. Starlink’s business-class option comes with an eye-watering price tag.

Starlink’s business class dish costs a whopping NZ$4,374.80 up front. Plans start at NZ$426 a month for a terabyte of data, a six terabyte plan costs NZ$2507 a month. This is more than 20 times what a user would pay for a fibre connection with the same performance.

There are roaming and maritime options for people on the move or those at sea.

The company’s website says the business class dish offers a higher level of reliability and users can expect speeds of “up 220 Mbps”. That’s about a quarter of the speed a fibre customer might see on an $85 a month Fibre Max plan.

Starlink business

Starlink’s marketing for the business class service says: “With a higher gain antenna, additional throughput allocation, and better extreme weather performance, Starlink Business helps ensure bandwidth for critical operations 24/7.”

This is something fibre customers never need to worry about.

When you connect to the internet by fibre, you have all the bandwidth on that strand of glass to yourself. Any form of wireless broadband involves a degree of shared bandwidth. This applies to Starlink. If everyone connected to the same satellite is streaming HD video at the same time there will be congestion.

Congestion

As mentioned earlier, Starlink can fix this by adding more satellites and it plans to do exactly that. Yet, for now congestion is real. You don’t hear New Zealand customers complain about congestion, overseas it is a different story. Canadian Starlink users report speeds far lower than those reported here by Measuring Broadband New Zealand.

There are parts of the world where Starlink is not accepting new customers.

Late in 2022 Starlink announced a fair use policy which allows residential customers 1TB of data each month. It says it reserves the right to throttle speeds if it needs to reduce network congestion. This appears to be a region-by-region policy.

Let’s put a terabyte of data in perspective. It is a lot of data, but it is possible customers will rub up against the limit. In February of 2023 the average data use per customer on the Chorus fibre network was 567 GB. The company expects average monthly data consumption to hit a terabyte in 2026, three years from now.

By then Starlink will have next generation satellites and more satellites in the sky. It will have more customers, but it’s realistic to assume the data cap will either be higher or will be forgotten.

At the time of writing, Starlink’s New Zealand terms and conditions page has removed mention of data caps and replaced it with the message shown here. As you can see, Starlink reserves the right to throttle data speeds.

The fact that Starlink announced then withdrew its data cap and throttling plans in a short space of time is a warning that these plans could be dusted down and revived at any moment. Yet as we shall see, that looks increasingly unlikely.

Competition

Starlink adds another layer of competition to New Zealand’s telecommunications market. While competition is good, it’s not as if the market wasn’t competitive before Starlink arrived.

New Zealand has one of the world’s most open and competitive telecommunications markets. The local telecommunications service providers make slender margins from selling broadband connections, compared with many overseas markets. High margins have been competed away.

The bigger telcos can cut deals with Starlink, 2degrees and One New Zealand have done this in recent weeks. They’ll be able to reach accommodations.

Starlink competition could be brutal for the smaller regional Wisps (Wireless internet service providers), in part because of the lower capital cost of getting a connection to a remote site and lower maintenance costs. Firing up a Starlink customer involves couriering a satellite dish. A Wisp might need to build towers and connect cables, truck rolls cost money.

Other LEOs

Next year Starlink will face competition from Amazon’s Project Kuiper LEO network. Oneweb, another LEO, says it will have coverage by the end of this year. Other LEO projects, including one from the Chinese government along with Iridium Certus, Telesat, and Facebook Athena, are at various early stages.

Not all these projects aim to compete with Starlink for residential and business customers. For example, OneWeb says it will partner with existing telcos.

When these alternative LEO networks come online Starlink will face direct price competition for the first time. Moreover, it will find it harder to implement data caps or using throttling if customers can decamp to a rival.

It’s hard to know if Starlink can sharpen its monthly subscription prices. The company is privately held, which means its finances are not transparent and its business model is not clear. What we do know is that Starlink is not yet profitable.

Launching and replacing satellites is expensive. Satellites need to be replaced every five or six years. The cost of building and launching a satellite is roughly in line with the cost of constructing a cellular tower, but towers do not need constant rebuilding. This suggests that Starlink and other LEOs will always be more expensive than fixed wireless broadband.

Starlink parent company SpaceX is developing larger recyclable rockets to launch more satellites at the same time and reduce the launch cost per ‘bird’. Last month the test rocket exploded on take off. There’s the cost of subsidising the residential dishes to take into account.

Does Starlink meet New Zealand’s rural broadband need?

The simple answer is that Starlink does a good job filling the gaps for farmers, rural businesses and professionals living outside the UFB fibre footprint. For people beyond the reach of fibre and fixed wireless networks it is the only plausible option for now.

Yet Starlink is too expensive for a lot of rural New Zealanders. It would be best for these people if New Zealand can follow what is happening in Europe and much of the rest of the world where fibre networks are extended further into unserved areas. We need the fibre network to reach the other 13 per cent of the population.

Starlink beats rural fixed wireless broadband by a long way on performance even if it costs twice the price. There’s a danger that rural broadband users will split along class lines with better off customers enjoying satellite broadband while less well off families are stuck with poor quality fixed wireless broadband.

Because Starlink gives the impression that New Zealand rural broadband is fixed, it could mean less well off rural families will go without broadband for far longer.