People ask me a lot about what Starlink means for somebody building a rural broadband network. That set me to contemplate the long-term prospects for LEO satellite broadband.

Today, the broadband provided by Starlink is a boon to rural subscribers who have had no alternatives. Hundreds of thousands of prospective customers have gotten onto the Starlink waiting list. It’s not hard to understand why when the rural broadband alternatives are extraordinarily slow rural DSL, high orbit satellite broadband, or cellular hotspots. A rural resident who gets Starlink broadband delivering 50 Mbps to 150 Mbps feels like they won the broadband lottery.

Assuming that Starlink and the other satellites can deliver the same decent broadband to larger numbers of users, the companies should do extremely well for the rest of this decade. That’s still a big if because there are still hurdles for Starlink to overcome. It must find enough backhaul to fully support the satellite constellation  – we know from other broadband technologies what poor backhaul means. There are spectrum battles at the FCC that might limit Starlink’s backhaul. Starlink also still has to launch many thousands of satellites, and according to Elon Musk, still needs $30 billion to finish the build.

But let’s assume that Starlink raises the money, launches the needed satellites, and gets the backhaul bandwidth needed to keep broadband speeds up. I foresee other issues that all of the satellite companies will face if I look out after ten years.

Landline Competition. Assuming that Congress gets its act together and passes a rural broadband bill, then many of the natural rural markets for Starlink will be getting fiber. It’s hard to see satellite broadband keeping much market share if fiber can bring more broadband for a lower price. Fiber is likely to never be everywhere in the country, but it could end up in most places. One only has to look at the Dakotas to see that there is a sustainable business plan for fiber-to-the-home if the initial fiber construction is subsidized. As the richer markets like the U.S., Canada, Australia, Europe, and the far east get fiber, the potential market for selling satellite broadband will be much diminished. We’re not going to see rural fiber everywhere immediately, but in a decade, we might be seeing a lot of it.

Prices. I also wonder about Starlink’s ability to sell broadband outside of wealthy economies.  I’ve read articles by several analysts who say that Starlink needs the $99 price everywhere to be viable. If this is so, then Starlink broadband is not a viable product in much of the world – there are not a lot of people in the third world who can pony up $99 per month for broadband.

Starlink also has to worry about competition from Jeff Bezos. One has to assume that Amazon will offer irresistible bundles of broadband tied to delivery and video services. If Starlink really needs a high price to succeed, it will be vulnerable to a ruthless competitor.

Growth in Broadband Demand. The issue that nobody talks about is the ever-growing demand for broadband. Since 2018 the national average monthly household use of broadband has grown from 215 gigabytes to 433 gigabytes. The growth has been on the same curve since the early 1980s. Even if the Starlink constellation can find enough backhaul to provide adequate speeds today, how will the technology keep up when demand in ten years will be ten times higher or in twenty years will demand will be a hundred times higher?

Viasat, which uses geosynchronous satellites at 22,000 miles above the earth, handles the demand issue by imposing severe data caps. On Viasat, a plan to get just 100 gigabytes of usage for a month is $170, with a 2-year contract required. A household that is limited to 100 gigabytes per month is not participating in the modern broadband world. If low-orbit satellite providers have to implement data caps to remain viable, their days will become numbered.

Constant Need to Replace Satellites. One of the most unusual aspects of low-orbit satellites is that satellites are expected to be replaced every five years or so. That means that the satellite company will face the giant cost of paying for rocket launches forever, just to stay in place. If Starlink ever grows to maturity, it will still have to constantly replace satellites, even if that means no new revenues. Any broadband technology suddenly feels a lot less attractive if the owner has to replace 20% of the network every year in perpetuity.

Summary. Assuming that the satellite companies actually raise the money, launch the needed thousands of satellites into orbit, and keep broadband speeds at a decent level, it’s not hard picturing the companies getting many millions of customers and even being profitable by the end of this decade. But the bigger long-term question is how a satellite company can sustain a business in a world where broadband demand is likely to outstrip the ability of satellites to deliver? The inability to keep up with demand is what ultimately killed DSL and is putting huge pressure today on cable companies to improve upload speeds. I think keeping up with demand will be the long-term Achilles heel for LEO satellites.