Hacker News Reader

Comment by gonehome | Hacker News Reader

$500 deposit, $2500 total cost for the dish, $500/mo for the service.

Looks like it's targeting the enterprise market.

The dish cost is probably closer to the actual value of the dish. The original dish we bought for $500 I think I read actually cost closer to $3k for them to manufacture (I'm not an expert on this)?

Regular service is $500 for a dish and $99 a month which is a great deal for 50-200mbps at pretty low latency. Particularly if you're in an area where your only other option is geostationary (which sucks and costs a lot).

View on HN

Why does this seem aimed at the enterprise market (besides the price tag)? From what I've seen, Starlink uses CGNAT (carrier grade network address translation) so it's hard to handle incoming connections (given the lack of a public IP address). CGNAT seems like it would be a problem for enterprise customers. Would this be enterprises in cities where they'd be looking for a backup for a wired connection? I know that a lot of companies have wireless back-ups for their wired connection so I'm not sure that Starlink would be a better enterprise option for them.

The question in my head is whether Starlink is going to slow down their $100-tier signups if they start seeing enough $500-tier signups. People have been waiting for years to get Starlink. If they start getting lots of $500-tier interest, it seems like they'd prioritize those users for new signups.

From the page:

> Unlimited Service Locations: Starlink is ideal for rural and remote locations. Order as many Starlinks as needed and manage all of your service locations, no matter how remote, from a single account.

This screams corp fleet management of the equipment, similar to SSO price segmentation.

Ah I missed this bit, yeah that’s definitely what they’re going for.

The cost paired with this copy: “enabling high throughput connectivity for small offices, storefronts, and super users across the globe.”

It’s 5x the price of standard Starlink (both for the hardware and the monthly cost). Won’t be worth it for most regular users imo when standard Starlink is quite good.

I’d pay $150/mo easily and maybe $200/mo for premium, but not $500/mo.

There’s also currently no way to order one from your current account if you have regular Starlink, you have to use a different email (also suggests targeting different market).

Why is CGNAT an issue ? Enterprise here doesn't mean on-prem public facing hosting. It just means rich entity that can afford $500/month per account. It could be anything ranging from a government office to a small business. You just need the ability to connect to your actual servers in AWS or wherever and transfer at acceptable rates and latency.

They say in their support topics that they may support publicly routable IPs in the future, so it looks like they may change that. But not all enterprise use-cases require inbound traffic, or they may require the bandwidth and low-latency even more, and can workaround the CGNAT in the meantime.

How many enterprises are running their own DCs these days? Very few from what I see. Cloud for your services, StarLink for individual office premises. Sounds fine to me.

Many small enterprises probably aren't. But the CGNAT problem can be solved by a VPN seamlessly connecting to the rest of the enterprise network, for those companies that need incoming connectivity (might be for remote maintenance, for example).

I was nodding along until I got to the $500/mo. service charge. I'd feel pretty good about $150/mo. or $200/mo., but I guess I'm not their target market since I have decent wired broadband. I was hoping that this would give Spectrum some competition in our area, but I'll have to wait a little longer.

Starlink won't really ever be designed to compete with wired internet. There is limited bandwidth per satellite, and the density of most places that have good wired internet far exceeds what is reasonable on the current generation of satellites.

It's designed for people who are currently suffering on geostationary satellite internet, or some other low-bandwidth/high-latency setup.

Exactly. I'm typing this out over a first-gen Starlink connection, and the value is a lot better than most people realize. My other options include:

- 4g hotspot, 150ms latency, ~15gb/month, speeds of ~10-20mbps, costs $50 + $15/gb overage per month

- WISP network, 400ms latency, unlimited data at 500kbps-6mbps, ~95% uptime, costs $80/month

- HughesNet satellite, 600ms latency, 50gb/month, less than 10mb/s down, 90% uptime, $150/month

Or I could get Starlink with ~50ms of latency, unlimited data, ~100-150mbps down, 99% uptime at the fair price of $100 per month. Once you've lived away from broadband for long enough, you can bet everyone is fed up with the other 3 options.

With the spread of 5G mid and high band home internet service I wonder if that might not cover enough ground at some point to make it worthwhile. Verizon is actually selling it as unlimited internet for a decent price. I don't know about other carriers.

It certainly won't cover anything out where I live. Even if they did put up a 5G antennae, it would cover maybe 2 houses on my street if they placed it strategically.

you're getting better service than the others because they are not at capacity, and won't be until the chip shortage is over. enjoy it while you can, because speeds will get worse soon, and what you thought was a great service will be just like the rest.

*In countries with reasonably good wired internet

I live in Australia, a 30 minutes drive from the centre of a state capital and I've had a near 10x increase in my download speed and a 4x increase in my upload speed by moving to Starlink. I'm paying $40 AUD extra per month ($130 for Starlink, $90 for pleb tier wired) for the privilege. My wired link will not go any faster, 5G still isn't available at my house, and even if it was it's metered so borderline useless for a primary home link with a family watching Netflix, YouTube for kids, etc.

The regular dish is great for normal personal use already in remote areas (and the costs are very reasonable).

It’s not competitive in (most) areas with options from cable providers, but that’s also not its purpose.

Starlink will give us high speed low latency internet anywhere on earth, it moves the lower bound up to 150mbps with pretty low latency which is great.

That's sticker shock for me... I'm paying $50/month for 1Gig in Rural Utah... I'm good.

It cost them around $1500 to manufacture.

> The original dish we bought for $500 I think I read actually cost closer to $3k for them to manufacture

$3k? Were they mining the metal by hand and then using only the most hipster blacksmiths shipped in from Portland to hand press the steel for the dish?

If I remember correctly there’s a fair bit of expensive electronics and fancy materials. The phase steered arrays I worked on in college were expensive and they were tiny. https://hackaday.com/2021/01/11/starlink-satellite-dish-x-ra...

It's not really the dish that's expensive (and it's not even a "dish antenna"). It's the integrated electronics and phased antenna array (which takes expensive tricky electronics to make it work).

There are 1500 antennas, and accompanying electronics for each, plus main board to deal with all that, GPS, motors. You also get a Wi-Fi router. Even $3000 is mindbogglingly cheap once you see what's inside of that "dish" and what quantity these are produced at (in 100K vs millions like smartphones).

Reply on news.ycombinator.com