Elon Musk wants to surround the Earth with a constellation of internet satellites, providing fast, low-latency broadband internet service to every person on Earth who wants it. New data from Speedtest.net suggests that he's well on his way to doing just that. 

Earth surrounded by an image representing an Internet satellite network

Image source: Getty Images.

Facts and figures

Musk has already made progress toward this goal. SpaceX's 11th "Starlink" launch Tuesday put 58 more internet broadband satellites in orbit, growing the number of functioning satellites to probably 650 or 660 sats. (Each Starlink launch can carry 60 satellites, but the company has recently been giving up a few slots on each mission to customers seeking rideshares to orbit.

Tuesday's launch marked SpaceX's 98th launch, and its 95th Falcon 9 launch. After delivering its payload, the Falcon 9's first stage landed back on Earth -- SpaceX's 57th such successful landing. (The specific booster that launched and landed has now made six such round trips.) SpaceX also successfully "netted" both payload fairing halves -- a $3 million piece of equipment that can now be reused on a future launch. 

Musk wants to put anywhere from 12,000 to 42,000 satellites in orbit eventually. Just 800 or so, however, should suffice to offer moderate coverage to most places on the globe -- so Musk is now more than 80% of the way to his goal of blanketing the Earth in broadband. 

Before he does that, though, he's going to want to focus on delivering internet to America.

A $16 billion subsidy for SpaceX

Rural America, to be precise.

As we discussed back in June, the FCC has given SpaceX the green light to compete for all or part of the $16 billion up for grabs under its Rural Digital Opportunity Fund (RDOF) competition. Under RDOF, the FCC will pick one or more companies to provide "gigabit speed broadband networks" to "unserved rural areas" of America.

To win even a piece of this business, SpaceX must prove it can provide internet service from orbit on par with service run through terrestrial cable services such as Comcast. Specifically, SpaceX must demonstrate that signals sent through its Starlink satellites have less than 100 milliseconds of latency (aka "lag" or "ping") between when a command is sent (for example, "buy it now") and a response received: "you won!"

60 down, 940 to go

As Ars Technica reported earlier this week, internet customers chosen to beta test SpaceX's Starlink network have been pinging speedtest.net to learn the speeds of the connections they're getting from Starlink, and posting the results on Reddit. 

What have they found out?

So far, results show that Starlink users can expect internet download speeds of anywhere from 11 megabits per second (Mbps) to 60 Mbps, and upload speeds of 5 Mbps to 18 Mbps. At best, those numbers are just 6% of the 1 gig (1 Gbps) speeds the FCC wants to be able to promise under RDOF. But considering that Starlink is still only 80% of the way to having enough sats for "moderate" capability, it's still impressive. (And as Ars notes, even 60 Mbps is much faster internet than many rural customers are able to access today.)

Perhaps more importantly, SpaceX and Starlink seem to have cleared the FCC's bar for competing in the RDOF contest. If you recall, the FCC was skeptical of SpaceX's ability to provide sub-100 lag speeds from space, and promised to apply "very close scrutiny" to SpaceX's performance. Ars, however, says Starlink is delivering ping rates of anywhere from 94 ms (already below 100) all the way down to just 31 ms.

What comes next

As far as qualifying to compete for RDOF and the FCC's $16 billion, this looks very much like "mission accomplished" for SpaceX. Going forward, the company will be aiming to drive its lag rate down below 20 ms, while boosting its internet download speed toward 1 Gbps. And over time, SpaceX hopes that these levels of performance will win it as many as 5 million broadband satellite internet customers in the U.S. At a rumored monthly service cost of $80, times 12 months in a year, Starlink therefore looks to be building toward a $4.8 billion annual revenue stream for SpaceX. 

What does it all mean for SpaceX?

When you consider that it would take SpaceX launching nearly 80 rockets a year at a launch cost of $62 million apiece to generate similar revenue from space launch, it's easy to see how Starlink could soon become SpaceX's primary revenue driver -- and a great candidate for an IPO