Comment by xoa | Hacker News Reader

Comment by xoa | Hacker News Reader

Parent Full thread

The beta terminal cost is $500, and beta service is $100/month (at up to ~150 Mbps or so right now?). So the money would cover the terminal cost to those customers and then full freight 8 months of service at current pricing. Alternatively/in addition, and more similarly to traditional in-ground service, it may also simply help fun actual infrastructure in the form of more/faster deployment of ground stations and sats. SpaceX may also (and it seems likely) have more standard service tiers for full general release, so something at 25 Mbps for $25/month say, and then working their way up. If they use this money to cover the minimum tier and then let people choose to pay the difference for more if they desired, then that'd bump the coverage to close to 3 years.

For the level of service they're offering that seems pretty reasonable honestly, and a much more efficient use of dollars than typical efforts (though WISPs might do even better in certain areas, they have different trade offs). Since it appears how the companies make use of the money is fairly flexible SpaceX will probably deploy it towards a few different metrics. The $500 upfront capex is probably the biggest impediment for a lot of people so getting that down is important, but if internally they had any issues with getting their sat shells filled out that'd be a good upfront usage too.

This is also spread out over 10 years, so where the most bang for buck is delivered will probably vary over time, by 2025 let alone 2030 SpaceX's economics will certainly look quite different with Starship deployment and filled shells/done ground stations.

I don't believe the subsidies are meant to subsidize the customer. Rather they are subsidizing the upfront capital cost to build out infra.

In other words, they get all the money, even if no one subscribes.

From the article:

>FCC funding can be used in different ways depending on the type of broadband service. Cable companies like Charter and other wireline providers generally use the money to expand their networks into new areas that don't already have broadband. But with Starlink, SpaceX could theoretically provide service to all of rural America once it has launched enough satellites, even without FCC funding.

>One possibility is that SpaceX could use the FCC money to lower prices in the 642,925 funded locations, but the FCC announcement didn't say whether that's what SpaceX will do.

SpaceX is still in the painful bootstrapping phase of things, and their cost structure right now is still higher then what they ultimately hope for. But from the sound of it they've got a lot of flexibility in how they deploy this money so long as they ultimately meet service targets, and they're only one player amongst many even for this subsidy money. If they're satisfied with service people don't tend to casually switch providers, so there is real value in being first. I think therefore SpaceX may consider using some of the money at least to lower the bar of someone signing up and try to get currently underserved households onboard as customers faster.

Additionally, don't forget this is only the first round, there is even more money available next time. Given the nature of bidding, I assume strategic calculations about how to win more next time will also play a role.

According to some leaked info SpaceX is paying $2400 per terminal.

The leaked info is sort of wrong in both directions. (Or, the info is correct, but is widely misinterpreted.)

The $2.4B they paid to ST Micro is not for terminals, as ST Micro doesn't manufacture electronics, they are a foundry and so they only make the chips. The $2.4B includes the costs for the chips for the first million receivers, but the assembly and non-chip costs add a lot on top of that. So in that sense, $2400 per terminal is a serious underestimate.

But it's also wrong in the other direction, because the $2.4B didn't just include the chips, it also included the development and mask costs for those chips. This should be considered to be NRE and not considered as part of the price of the terminals, as SpaceX will not have to pay that for the next million terminals.

So the real number is above or below $2400 depending on what costs more, the motors, pcbs, the antenna layers, the chassis and the assembly of a terminal, or 1/1000000 of the NRE of developing the ASICs for Starlink. Who knows.

Reply on news.ycombinator.com