edit: MASSIVE miscalculation fixed!
Let's say they have 12k satellites. With oversubscription at 10? and a third of time over land? but 20gbps capacity for 200mbps connections, you get 333 users per satellite. 12k gives you 4 million users, at $100/mo or $1.2k/y. I'm going to drop that to $80/mo because Starlink need to pay for backhaul, so let's say $1k/y because it's round. So $4 billion annually. Over 12 years of satellite lifetime that's 48 billion. Satellites cost $300k to make, so subtract another 4 billion for production (44 remaining). Sats are launched in batches of 50, requiring 240 launches. Internal costs are somewhere between 30 to 60 million, it's unclear because of reuse, let's say 50 million for 12 billion launch costs, leaving 3 billion a year.
And then there's some operational costs which they pay from that.
So Iunno, it doesn't seem as extremely profitable as my first calc, but it's still probably pretty profitable. And any improvement they make to the satellites over time just adds on top. (Also I don't think they actually have $50 million internal cost per launch.)
edit: Terminals "reportedly cost around $1000 to make" and are sold for $500, so nbd, subtract another 2 billion per customer-lifecycle.
Here's some of my math:
* Each customer brings in $1.2k/year.
* Satellites have 20 gbps of bandwidth, and consumers today are reporting about 50 mbps links.
* Maximum capacity is then ~1000 customers/sat because a 50 mbps connection takes 100 mbps of capacity to run duplex, and you need to double that again because customer data needs to get to backhaul.
* Satellites will spend significant time over water, and even more time over people who aren't paying - remember, metro areas all have cheaper cable internet. I would give these ~10% capacity at most.
* Satellites have 5 year life, but some fraction of them fail early - let's say that averages to 4 year replacement, and that estimate is at the high end of Musk's original estimate.
* One satellite is capable of $1.2M annual revenue, assuming 100% utilization, or $5M revenue over its life.
* Satellites cost $250k to manufacture, and $1M to launch ($50M per launch of 50 sats, making the math simple). Let's raise the unit cost to $1.5 million assuming a very good yield on satellites making it to orbit.
* The profitability of a single satellite is heavily dependent on utilization: At 10% utilization, each satellite makes $500k of revenue over its life. Satellites have to run at 30% utilization to break even.
* Customers receive a ~$1k subsidy to access the network in terms of the discount on the dish. This means we have 1 year per customer to break even on the customer. Assuming 3 year customer/dish lifetime (and assuming we will subsidize dish replacements when they break), we have only ~$2k margin per customer over 3 years, or $700 per year amortized.
* Now, the per-satellite gross margin drops to around $3-3.5M. This means that you now need up to 50% utilization to break even on a satellite.
What this all means is that starlink, if it is ever profitable, will be profitable based on niche use cases where they can charge extremely high subscription prices. Use cases like airplanes, yachts, offshore oil platforms, and military deployments. Customers who aren't paying a lot per subscription are dead weight.
By making slightly different assumptions, we have dramatically different conclusions.
The biggest factor driving disagreement seems to be satellite lifetime. You also didn't account for oversubscription.
edit: Though-
> Let's raise the unit cost to $1.5 million assuming a very good yield on satellites making it to orbit.
How do you get that?! Starlink don't lose 30% of sats, lol.
edit: I've put up a Guesstimate Model: https://www.getguesstimate.com/models/20609
https://m.youtube.com/watch?v=2vuMzGhc1cg&feature=youtu.be
Looks to me that you forgot about the consumer-side terminal.
https://news.ycombinator.com/item?id=31486083&p=2#31487813
Read around and that thread a bit and you'll see that people confirmed my claims and that those who doubted it and checked his numbers themselves observed a three order of magnitude difference* in his cost estimation versus their estimate.
This isn't a source to get informed by, but one that makes you delusional.
Can you tl;dw? I'm not watching 40 minutes of a video that has DEBUNKED in the title.
Calculation fixed!
