Elon Musk and his Corporation SpaceX has spent tremendous efforts developing a Low Earth Orbit (LEO) satellite constellation called Starlink. With over 1,000 satellites now in orbit and beta tests starting in multiple countries it is fair to ask whether Starlink is going to be a success. So far, the path of previous LEO initiatives has been littered with failure. Why? Because launching thousands of satellites, connecting them to individual low-cost user terminals and networks of ground-stations to connect to the Internet is intriguing but incredibly challenging. But that is not the most difficult part. The most difficult part is establishing effective distribution around the world and selling the service at an affordable price that is nevertheless sufficient to make money. LEO systems do not just serve one country or even one region. The orbiting satellites spend over 70% of their time over water – and the other 30% over every landmass in the world. They need a global market but unfortunately the world is not one place - it is a complicated myriad of different regulatory environments, different levels of development in terms of telecommunication needs and even more importantly different levels of economic ability to pay for a service.

Of course, Elon Musk has a phenomenal track record in building new businesses in challenging markets and so there is every reason to believe that his commitment and drive to overcome some of these obstacles is going to be successful. But in this case, he is not building a system for market that actually exists. Well maybe it does exist in some more affluent markets but to be able to address the global market a significant amount of work will need to be done to achieve regulatory approval, establish distribution, and indeed more importantly find a price point that will be both successful in a mass market and deliver the financial return necessary to allow the service to be viable.

Let us think about viability a little more. Morgan Stanley have recently put together a model which projects that Starlink has a potential enterprise value of over $80 billion. That potential value is based on a 20-year model that predicts that the business will be able to achieve somewhere in the region of 250M to 300M subscribers in a network of over 40,000 satellites. Clearly there are not enough subscribers in affluent markets to achieve that – no-one with an existing high quality wired Internet connection is going to move to a LEO system so we are dealing with the marginal consumers that do not have a great wired or wireless service. They certainly exist but not in the 10s of millions. That means that Starlink is going to need to penetrate the global market in a big way. This is neither easy nor, frankly, plausible at the level suggested in the Morgan Stanley model. Two of the major markets in the world, China and Russia, will probably be locked from a regulatory point of view - indeed China is talking about its own constellation of LEO satellites. Other large markets like India have significant regulatory barriers. The rest of the global market really does not exist in large chunks - it is fragmented both from a regulatory and from a demographic perspective and it is hard to imagine that hundreds of millions of subscribers will be found easily. By the way, even if a single market of 100M subscribers existed, Starlink would not have the concentrated capacity to be able to serve it – its capacity is spread across the globe. To put this in a different perspective after 20 years Netflix is only just getting north of 250M subscribers and that is with a simple service with unique content at an average monthly price of around $10. The Starlink Internet service is quite different - it is going to be way more expensive and it is going to be way more complex (remember the user is going to have to have their own terminal tracking the satellites). Indeed, it is unlikely that the service can be marketed for much less than $100 since the whole project is going to cost 10s of billions of dollars. At anywhere close to $100/month the market will be very restricted even in economically developed regions.

There are many more challenges for Starlink project. The Morgan Stanley model assumes that the satellites will need to be replaced every five years that is not an unreasonable assumption. This means that to launch and sustain a network of over 40,000 satellites Starlink will need to launch over 126,000 satellites – yes, 86,000 of them will need to be deorbited and replaced in the first 20 years of service. All this means that there are many assumptions in the Morgan Stanley model which, frankly, are unlikely to be achieved to the point of impossibility. That means that the Starlink enterprise value they suggest is wildly overstated nevertheless reports tell us that SpaceX has just raised another $850M in equity to continue to fund its developments. We will watch progress with interest.