But there's one more thing about Starlink unclear to me. What about costs of rocket launches, which are delivering satellites to the orbit? IPO doesn't reveal it, but most probably they are not included in Starlink costs. Either by depreciation write off, mentioned in the article, or simply by putting them to the space business (highly unprofitable), not communication. Having in mind that about 80% of Falcon launches were delivering Starlink satellites, it is, in fact, Starlink cost. And not just one time investment, but fixed cost for the lifetime. .
Is there anyone who could clarify what is that cost, actually? Official price is 74 M$, but that's sales price. Internal cost must be different. I've seen estimates (do not remember where) that it is 30 - 35 M$. Sounds reasonable, but it would diminish Starlink's profit significantly, if included in its costs. I've seen also other estimates, claiming that launch cost is only 10 - 15 M$. It would be great for Starlink, but would also question the role of Starship. Why to spend enormous money on the development, if the launches are already very cheap, so payback time counts in tens of years.
For depreciation rules, companies can have different rules for tax and regular GAAP rules. The one time bonus depreciation allowed for tax purposes is allowed for tax but not GAAP. Musk is most likely using straight line depreciation for general accouting practices. Lots of tax deductions in the first year, but for his books it is 20% per year if the satellites are good for five years.
I live in a rural part of Switzerland without phone or fibre service. So for internet I have to rely on something else.
That something else is 5G. It gives me reliable internet, at about 600Mb/s down and 200 up. Those limits are probably down to my hardware, which is already 4 yo, and which I need to replace. I download about 1.5Tb a month. (Working from home). Cost 35 CHF (about 40$) a month.
Starlink starts here at 35 CHF a month. Starlink is heavily discounted in Europe compared to the US, as it would otherwise not get any takers. For the Max plan I would pay 70 CHF /month.
Were I to live in the city I could get fiber at speeds up to 25Gb, with unlimited traffic, for 77,- a month. Starlink has a problem... Its product is just not that attractive.
I've been getting a Starlink postcard in the mail about every week for the last 2 months or so.
I live in a very rural area, but I'm in town, where we have fiber optic internet.
People out here in the country (outside of town) have used satellite internet for years and they HATE IT. I don't think he's going to find many customers here.
Thank you for this. The life of a LEO satellite seems to be about 3 to 5 years, max. Ironically, the life of an AI server is probably 3-5 years. We are facing a future with long term investments into short term tech gear.
Meanwhile, no one is talking about the bandwidth constraints on the AI / server / user connection.
The coming months and years will be brutal as this shakes out.
I waited for this article, thanks! 🙂
But there's one more thing about Starlink unclear to me. What about costs of rocket launches, which are delivering satellites to the orbit? IPO doesn't reveal it, but most probably they are not included in Starlink costs. Either by depreciation write off, mentioned in the article, or simply by putting them to the space business (highly unprofitable), not communication. Having in mind that about 80% of Falcon launches were delivering Starlink satellites, it is, in fact, Starlink cost. And not just one time investment, but fixed cost for the lifetime. .
Is there anyone who could clarify what is that cost, actually? Official price is 74 M$, but that's sales price. Internal cost must be different. I've seen estimates (do not remember where) that it is 30 - 35 M$. Sounds reasonable, but it would diminish Starlink's profit significantly, if included in its costs. I've seen also other estimates, claiming that launch cost is only 10 - 15 M$. It would be great for Starlink, but would also question the role of Starship. Why to spend enormous money on the development, if the launches are already very cheap, so payback time counts in tens of years.
For depreciation rules, companies can have different rules for tax and regular GAAP rules. The one time bonus depreciation allowed for tax purposes is allowed for tax but not GAAP. Musk is most likely using straight line depreciation for general accouting practices. Lots of tax deductions in the first year, but for his books it is 20% per year if the satellites are good for five years.
I live in a rural part of Switzerland without phone or fibre service. So for internet I have to rely on something else.
That something else is 5G. It gives me reliable internet, at about 600Mb/s down and 200 up. Those limits are probably down to my hardware, which is already 4 yo, and which I need to replace. I download about 1.5Tb a month. (Working from home). Cost 35 CHF (about 40$) a month.
Starlink starts here at 35 CHF a month. Starlink is heavily discounted in Europe compared to the US, as it would otherwise not get any takers. For the Max plan I would pay 70 CHF /month.
Were I to live in the city I could get fiber at speeds up to 25Gb, with unlimited traffic, for 77,- a month. Starlink has a problem... Its product is just not that attractive.
I've been getting a Starlink postcard in the mail about every week for the last 2 months or so.
I live in a very rural area, but I'm in town, where we have fiber optic internet.
People out here in the country (outside of town) have used satellite internet for years and they HATE IT. I don't think he's going to find many customers here.
Thank you for this. The life of a LEO satellite seems to be about 3 to 5 years, max. Ironically, the life of an AI server is probably 3-5 years. We are facing a future with long term investments into short term tech gear.
Meanwhile, no one is talking about the bandwidth constraints on the AI / server / user connection.
The coming months and years will be brutal as this shakes out.