Your browser is not supported. Please update it.

6 October 2020

Starlink may struggle to force urban ISPs to lower their broadband prices

A report from Broadband Now suggests that SpaceX’s Starlink satellite broadband offering could force terrestrial rivals to lower their prices by 20% to 40%. It comes as the number of Starlink units in orbit has passed 700, meaning that it has reached the ‘minimum viable product’ threshold, and as CEO Elon Musk kicked off speculation about an IPO.

A Starlink broadband service will be a huge boon for consumers who do not have a competent terrestrial alternative. There are huge swathes of the USA that only have a single high speed broadband provider, and where the only rival is dial-up. Those ISPs should begin sweating now, as a Starlink service would effectively sign a death warrant for them.

The underserved broadband markets are largely rural ones, where the cost of building out terrestrial infrastructure makes it very unattractive to the larger ISPs. This then means that dozens to hundreds of smaller ISPs have cropped up, taking the hit of the upfront costs and then charging much higher fees to customers than these consumers would have to pay in a more competitive market – such as their nearest large town or city.

But SpaceX is not such a concern for the large urban ISPs, because it doesn’t have enough capacity in its satellite network for a large number of customers per cell. There is an upper limit on the network quality, and it is very closely tied to the number of dishes on the ground that are requesting service from the satellite constellation passing overhead. Musk has already warned of this problem, likely to try and take some of the steam out of the hype machine – as this isn’t a problem SpaceX can solve by just throwing more satellites at the problem.

The Broadband Now report estimates that SpaceX could undercut rivals on price by 20%, in markets where it has three or more rival suppliers. In markets where there are 9-10 rivals, the figure is apparently 40%. But given the per-cell capacity constraint, we’re somewhat sceptical about the idea that Starlink is going to slash the average price of broadband in urban areas. Our understanding is that SpaceX wouldn’t be able to poach enough customers in a city, to alter the pricing, because it can’t support those dense customer deployments.

In terms of Starlink’s actual performance, SpaceX had to expand its million-customer licence from the FCC to allow for 5m, due to the popularity of the beta program. It quickly reached 700,000 registrations of interest, and it was at this point in August that Musk explained that “we’ll have some small number of customers in LA, but we can’t do a lot of customers in LA because the bandwidth per cell is simply not high enough. Starlink will serve the hardest-to-serve customers that telcos otherwise have trouble doing with landlines or even cell towers.”

Even with 12,000 proposed satellites, SpaceX can’t support dense customer footprints like those it would find in cities. There would be too many customer end-points for the low earth orbit (LEO) satellite to beam data to and from, and after all, those in cities tend to have multiple options for broadband providers. That is obviously not the case in the countryside.

So returning to the Broadband Now report, it seems clear that the central thesis of a new entrant bringing the price down only really makes sense in rural locations – where the cell density is sufficient to act as a market force. If one such satellite cell could support every home and business within it, because they are thinly distributed in a rural environment, then the incumbent ISP is very susceptible to price cutting and customer defections.

This of course does not apply in cities – where Starlink might only be able to support single-digit percentages of all the available broadband customers within the cell. That’s not enough capacity to influence the market, and not enough to force the pricing down by 40%.

Reddit users have been posting their experiences with SpaceX’s Starlink beta program, with download speeds ranging from 11Mbps to 60Mbps, and an average of 42.8Mbps. The upload speeds, which are crucial for meeting the FCC’s broadband criteria for funding, range from 4.5Mbps, under the threshold, up to 17.7Mbps. Latency ranges from 20ms to 94ms. These are promising early results, but whether they stay this high when the users-per-cell number climbs remains to be seen.

On the IPO front, Musk started dropping hints about what a Starlink IPO would look like – a few weeks after SpaceX president Gwynne Shotwell publicly confirmed the intent to list it on a stock exchange. It appears the plan is to spin the broadband operations out from SpaceX’s main launch business, although this is apparently still several years away. Musk also said that small retail investors would get priority in the listing, perhaps hinting at his frustrations with Tesla’s ongoing share price manipulations. Musk is adamant that Tesla is the most shorted stock in history.

The plan is to IPO once revenues and cashflow are stable, and operational stability is key to this. Starlink is currently working with Washington’s Military Department Emergency Management Division, providing connectivity to both the public as well as emergency services, as they tackle the ongoing wildfires.