A prediction made seven years into the future, for a service which didn't exist then, only off by an order of magnitude doesn't sound too bad imo.

Additionally, holy crap has a lot changed in that time. Falcon 9 didn't complete its first landing until the very end of 2015 with Flight 20. What's now routine and key to their costs is a feat less than a decade old. SpaceX was still talking about the Mars transport/BFR at that, which were very different in a variety of ways from the Starship SS/SH system we see now. What they've got now feels a lot more practical and sustainable long term, moving to steel for example was a great change.

So as far as long term forward looking plans go I'd agree that execution here was pretty darn good, and they've got a very clear roadmap for the next 5 years as well with full size v2 sats waiting in the wings and Starship proceeding along at a solid clip. It was a curiously negative feeling piece particularly for the WSJ. How many companies execute ambitious highly capital intensive global scale infrastructure rollouts using never been done combos of tech at never been done price points and pull it off like that?

Sporting a growth rate that, if it holds, means the prediction was off only by a couple of years.

Quite correct. Less than 2 years off from that timeline if it continues with its current growth rate.

60 million people live in rural areas in the US, over 40 million of those already have broadband at home[1], and less than 2 million people subscribed to satellite internet before Starlink launched[2]. His projection of 20 million subscribers would have meant that literally every single rural individual who didn't already have broadband would sign up for Starlink within a couple years of being available. Okay, that's not entirely fair because Starlink does business in more than just the US, but still, that was an unrealistically aggressive projection even 7 years ago.

I think Starlink is doing great, but Musk's projections are notoriously bad, and there is no reason to defend them.

[1] https://www.pewresearch.org/short-reads/2021/08/19/some-digi...

[2] https://web.archive.org/web/20181020011542/https://www.stron...

This is misleading for several reasons. The FCC defines broadband as 25/3 mbit up/down, which is not fast. Also it doesn't take into account latency which is a major problem with geostationary satellites. So plenty of those 40 million with "broadband" are going to want to switch to better service anyway.

Your first article concedes as much:

> Even though rural areas are more wired today than in the past, current infrastructure does not support consistently dependable broadband access in many rural areas. This lack of reliable high-speed internet access has come to the forefront of discussions about navigating remote work and school during the coronavirus pandemic.

Your second link states that there are 1.7 million satellite internet subscribers in the US. But Starlink already has that many subscribers, so it is clear that the market size is much more than just existing satellite users.

Even if 100% of rural customers were dissatisfied with their existing service, projecting that you will capture a third of the potential market within a few years of providing service is unrealistically aggressive.

The world is a big place. With billions of people

You always make pie in the sky projections when raising money in the very early days.

Still a $1.4 billion (in revenue) dollar company with a real business model and market dominance shrugs

It's cool, for sure. But until they stop selling a dollar for 90c, then they don't have real business model.

> until they stop selling a dollar for 90c, then they don't have real business model

Starlink is glued to a transport monopoly. It doesn't take much imagination to see how economies of scale on production and marketing make those numbers work.

Amazon sold a dollar for 90 cents for 20 years, and I don't see anyone laughing now. It's not a bad business model for a massive fixed cost but minimal marginal cost business like Starlink.

It does have minimal marginal cost, as long as you live in the right place on earth, without many other customers next to you.

Once you exceed, fairly limited amount of people that can use it one cell, you need to grow massively your fixed cost, as you need to grow the whole constellation or just won't be able to grow.

Time will tell if that will add up. I don't think it's obvious it will.

You nailed a big difference between Starlink and Amazon: if all my neighbors become heavy users of Amazon, Amazon becomes more profitable.

The starlink satellites decay after 5 years. The constellation costs $40B. That's an $8B annual burn rate just to maintain the current constellation.

Where did you get this $40 billion idea?

Given my observation of the space and EV industry in the last 5 years, this seems downright reasonable.

The criticism chain for Elon Musk's ventures is usually "It's impossible! This is a scam!" -> "He promised he would have X five years ago!" -> "He has too much money / power / a monopoly! Something has to be done!"

Cool. Did you buy the $10,000 feature that will let you summon your Tesla from across the country sometime in 2018?

https://www.consumerreports.org/cars/autonomous-driving/time...

"He promised he would have X five years ago!"