There's a lot we don't know about Elon Musk's most recently not-announced-but-announced plans to build a huge global satellite network to improve the speed of the Internet. But the scope and scale are huge, bigger than any of Musk's current ventures and will alter the landscapes of the telecommunications and space industries if everything comes together as planned—a BIG "If."

Musk did a big private event in Seattle last week, inviting 400 guests – mostly potential SpaceX (News - Alert) recruits, a couple congressmen, local officials, but no media, reports The Seattle Times – to celebrate the launch of SpaceX Seattle. 

SpaceX's new facility in Redmond, Washington will be the hub to build a constellation of over 4,000 satellites to provide high-speed Internet across the globe.

Let me stop here: Four Thousand Satellites. 

There are roughly 1,100 or so active satellites in orbit, plus another 2,600 inactive ones that are turned off or simply dead. Musk is talking about building and launching more satellites in one project than everything living and dead in orbit today. 

Satellite manufacturers have to be going crazy right now. Analysts expected an average of 115 satellites launched PER YEAR over the next decade—which adds up to around 1,150 total satellites for the decade, or a bit less than a quarter of what Musk wants to do in a single project. Space Systems/Loral booked the largest number of commercial contracts, ten, to build to build a total of 24 satellites. Everyone else in the commercial satellite arena ranges from three to five.

The SpaceX 4K constellation would fly in low Earth orbit at around 684 miles, weigh several hundred kilograms, and use lasers to provide high-speed cross-connections between nodes. Uplinks/downlinks would presumably be conventional high-speed radio frequencies, but there are few details on how that will work.  Musk's estimate is some sort of hardware going up in five years and the full project to cost $10 billion.

"Rebuilding the Internet in space," as Musk puts it, would provide a cash cow to fund SpaceX's ultimate goal of putting a colony on Mars. The goal is to have "the majority of long-distance Internet traffic" on the 4K constellation, and about 10 percent of local consumer and business traffic. People would still have local fiber, but the core network would migrate off fiber onto the constellation.

Google (News - Alert) is going to put $1 billion into SpaceX to make the 4K constellation happen, according to The Wall Street Journal and other sources. SpaceX's 4K provides another way for Google and other Internet companies to reach the "other billion" who either have limited or no Internet access, and do so in a way that doesn't require putting up drones or balloons.

I suspect a lot of long-distance fiber carriers are unimpressed and not losing sleep at this point. SpaceX satellite networking would have to provide highly reliable high-speed connections that at least match those of fiber and are either equal to or lower in cost than existing long-haul broadband connectivity.  Cable and Tier 1 service providers have a ton of fiber in place and are working the technology curve to move from 40 Gbps to 100 Gbps. Five years is a long time and there's a lot of legacy hardware and money in fiber.

Other pieces in play here are SpaceX cranking up a big satellite factory and the implication that the company has 100 percent confidence that it will be dropping the cost of putting things into orbit. Musk basically said if he's building a lot of satellites in a SpaceX factory, he might as well offer to build them for customers. 

Earlier this month, SpaceX was close to landing the first stage of its Falcon9 rocket on a floating barge.  Success brings the potential for reusing the first stage, thereby significantly lowering launch costs.  You'll need a lot of low-cost flights to put up over 4,000 satellites, plus replacements and hardware refreshes.