SpaceX's next big commercial customer

[Robotbeat]

Planet Labs is awesome for checking on construction status. Particularly their new Skybox which has 0.9m imagery.

I can check to see if Raptor or BE-4 has had any test firings and when based on ground burn markings. I can watch as Tesla expands the Gigafactory, places solar panels, etc.

If you want to know what your competitors are up to or if some company you’re considering investing in is BSing you, Planet Labs provides pretty valuable insight. The higher resolution Skyboxes are a real game-changer, too. You can see exactly what stage of construction something is in. You can measure stockpiles of coal or other commodities.

[rory]

Well, FarmersEdge is a customer now....What he uses it for is kinda silly.

That's my point.  I've yet to hear about a non-silly need for Planet Labs style imagery.  The kind of need where lots of folks in the same situation say, yeah, that's a good idea, and buy as well.

Who's to say the FarmersEdge daily satellite imagery is silly? They've been resellers since at least early 2016, along with FarmLogs. It's important enough to their product that it's featured on the top bar of their website — in fact it seems to be the basis of all their mapping features.

FarmersEdge has raised over $60 million. FarmLogs has raised another $37 million. I think it's safe to say their Planet imagery is providing real value to many paying customers.

[Kansan52]

Your right. Only meant the one example was fun and silly. It was meant to say I don't know how the data is used by the customer.

[niwax]

This is an interesting recent video covering aerial photo applications and companies:
For example, the Gates Foundation relies on low-cost satellite imagery to count huts in Africa for population estimation to guide humanitarian efforts.

[First Mate Rummey]

I don't think SpaceX is willing to launch OneWeb satellites at this point. Higher launch costs or longer time to deploy OneWeb with non SpaceX launcher will make Starlink more competitive. This may change, for example if OneWeb will become a lot more or a lot less competitive than Starlink, than SpaceX may be fine with launching if it doesn't directly impact its constellation business.

Maybe also true for DreamChaser. Each of the three companies has a minimum of 6 launches each. Using F9 (rather than a costly other launcher) for DreamChaser would make it more competitive than F9 + Dragon for eventual additional launches.

[wannamoonbase]

I don't think SpaceX is willing to launch OneWeb satellites at this point. Higher launch costs or longer time to deploy OneWeb with non SpaceX launcher will make Starlink more competitive. This may change, for example if OneWeb will become a lot more or a lot less competitive than Starlink, than SpaceX may be fine with launching if it doesn't directly impact its constellation business.

Maybe also true for DreamChaser. Each of the three companies has a minimum of 6 launches each. Using F9 (rather than a costly other launcher) for DreamChaser would make it more competitive than F9 + Dragon for eventual additional launches.

I'd bet they would. 

I understand your logic, but launch revenue could fund the deployment of Starlink.  And I'm sure Starlink will get a better price per launch.  Each launch of the Falcon family makes the next launch cheaper and improves the overall position of SpaceX in the marketplace.

They have the launch sites and Block 5 could be the vehicle that makes it possible to do 30-40 launches a year.  That's a lot of launches.

[Lar]

Can't see SpaceX turning OneWeb away once they have room on their manifest. Doesn't mean they give them a great price though....

[AncientU]

I don't think SpaceX is willing to launch OneWeb satellites at this point. Higher launch costs or longer time to deploy OneWeb with non SpaceX launcher will make Starlink more competitive. This may change, for example if OneWeb will become a lot more or a lot less competitive than Starlink, than SpaceX may be fine with launching if it doesn't directly impact its constellation business.

Maybe also true for DreamChaser. Each of the three companies has a minimum of 6 launches each. Using F9 (rather than a costly other launcher) for DreamChaser would make it more competitive than F9 + Dragon for eventual additional launches.

I'd bet they would. 

I understand your logic, but launch revenue could fund the deployment of Starlink.  And I'm sure Starlink will get a better price per launch.  Each launch of the Falcon family makes the next launch cheaper and improves the overall position of SpaceX in the marketplace.

They have the launch sites and Block 5 could be the vehicle that makes it possible to do 30-40 launches a year.  That's a lot of launches.

Getting these payloads on Falcon takes them off the table for other (more expensive) launch services suppliers, which is as important to long term goals as is making money on Dragon 2, for instance. 

Starlink will always be flown at cost vs OneWeb being at market price, so that dollars per sat on orbit advantage is guaranteed.  Starlink needs to win the global internet market competition based on satellite capabilities and system software, though, more than launch expense, IMO. 

[oldAtlas_Eguy]

I don't think SpaceX is willing to launch OneWeb satellites at this point. Higher launch costs or longer time to deploy OneWeb with non SpaceX launcher will make Starlink more competitive. This may change, for example if OneWeb will become a lot more or a lot less competitive than Starlink, than SpaceX may be fine with launching if it doesn't directly impact its constellation business.

Maybe also true for DreamChaser. Each of the three companies has a minimum of 6 launches each. Using F9 (rather than a costly other launcher) for DreamChaser would make it more competitive than F9 + Dragon for eventual additional launches.

I'd bet they would. 

I understand your logic, but launch revenue could fund the deployment of Starlink.  And I'm sure Starlink will get a better price per launch.  Each launch of the Falcon family makes the next launch cheaper and improves the overall position of SpaceX in the marketplace.

They have the launch sites and Block 5 could be the vehicle that makes it possible to do 30-40 launches a year.  That's a lot of launches.

Getting these payloads on Falcon takes them off the table for other (more expensive) launch services suppliers, which is as important to long term goals as is making money on Dragon 2, for instance. 

Starlink will always be flown at cost vs OneWeb being at market price, so that dollars per sat on orbit advantage is guaranteed.  Starlink needs to win the global internet market competition based on satellite capabilities and system software, though, more than launch expense, IMO.

Its a competition thing and the "Cost of Money" item is a major hidden cost in the business case. This is how much the capital outlay at the beginning of the effort costs before the effort makes money. Starlink gets rides at SpaceX internal cost. One Web would get rides for the Internal cost + some profit %. That would be some where above a 20% value. So if the internal cost is $.75M/sat then the profit value could be $.25M for a total of $1M per sat for a much lighter sat. But note this is a competitive price per sat for One Web to other providers. It would make on the launch every 100 sats $25M in profits for SpaceX. For every 1000 sats $250M in profits. This is a significant customer and would not be ignored by the space access transport arm of SpaceX even if it is a competitor to Starlink. Starlink would still have the advantage because of the "Cost of Money" issue because of a lower cost per sat for deployment and a lower capital expenses prior to revenue generation.

[speedevil]

Getting these payloads on Falcon takes them off the table for other (more expensive) launch services suppliers, which is as important to long term goals as is making money on Dragon 2, for instance. 

Starlink will always be flown at cost vs OneWeb being at market price, so that dollars per sat on orbit advantage is guaranteed.  Starlink needs to win the global internet market competition based on satellite capabilities and system software, though, more than launch expense, IMO.

Starlink needs to win (from a Musk perspective) for any reason at all.
This can include the satellites being a really nice shade of purple, if that turns out to be important.

Rather more likely to be important is a six or twelve month in getting a full constellation deployed. If Starlink gets up and working in the minimal usable configuration six months earlier, due to other providers not being able to launch as fast, that could rapidly drive revenues which dwarf the launch profits.

These revenues can be directly plowed into finishing BFR and BFS and incidentally killing off any temporary benefit to other launch providers.

[Lar]

I think you're both right but I have to give a lot of credit to AncientU for the 'starve the other providers" aspect which is perhaps more easily overlooked.

on the other hand, Speedevil's "get there first" aspect has a lot of merit.

[Jcc]

It's possible for both Starlink and OneWeb to be successful, as broadband becomes truly ubiquitous and cheaper, demand for it will go up everywhere. The difference between 800 sats and up to 12000 says that Starlink will have far greater capacity even if OneWeb is first to market, the best scenario for OneWeb is they saturate their system with customers and traffic, but Starlink could undercut their price and win customers from them.

Anyway, the willingness to launch a competitors satellites would be a defense against charges of monopoly and regulators going after them, so they should want OneWeb to succeed.

[oldAtlas_Eguy]

From an anti-trust standpoint SpaceX has to offer the same price for launch services to their competitor OneWeb that they offer other customers. So if OneWeb wants to ride a F9/FH/BFR then all they have to do is contract with SpaceX. SpaceX is not really in a position to refuse.

As total number of flights per year increases then the margins on profits per flight increase because there is less fixed costs applied to each individual launch. By OneWeb contracting will SpaceX it may actual reduce the launch costs for Starlink because flight rate increases.

Now back to who or what may be SpaceX's next big commercial customer. We have stated the obvious of Starlink and OneWeb but there could be a new application or business case that closes with lower launch costs than what is currently available. One such item that keeps being kicked down the road is Space Based Solar Power Satellites. To close the business case the $/kg needs be be < $500/kg. The best currently is FH with a possible of as low as $2,000/kg. But if the business case closes the launches is not 50 to hundred its more like hundreds to a thousand launches of something as big as the BFR.

In the future I see two different large customer types. Large sat constellation operators and bulk cargo for very large on-orbit construction or propellant for interplanetary flights. At the moment it only seems like the first type have business cases that currently close (maybe). But a bulk cargo application could show up based on providing an alternative to the high individual costs of launching to common orbits such as an on-orbit (probably LEO) satellite assembly and deployment facility that accepts bulk shipment of assemblies that is them fashioned into final satellites that is then using a refuelable tug (prop depot required) to do the deployment. This business case has been studied but there is yet to be enough monetary incentive to implement mainly a lack of low $/kg bulk transport.

The final leg in this is that although the FH is a beast it is still not really large enough or cheap enough for the bulk transport business cases to close so that by default leaves only large constellation operators as the possible next big customer. But we can still be surprised if NASA chooses on a fast trac BEO HSF support flights for its new Lunar focus. 2 to 4 FH launches each year is still significant even if the number of launches is low. Yearly contract value $200-500M depending on circumstances and number of launches. Such contracts would be multiple years on the order of 3 to 5 or a total contract value of $.6B to $2.5B. So don't count the US government out quite yet as a player in this game of who and what will be the next big customer.

[Mader Levap]

There is very simple reason for SpaceX to launch OneWeb. These sats will be launched either way by someone, so why not get something out of it, even if it is for competing constellation?

[IainMcClatchie]

For example, the Gates Foundation relies on low-cost satellite imagery to count huts in Africa for population estimation to guide humanitarian efforts.

Not an example of a trillion dollar business though.  Humanitarian efforts are tens of billions a year, the imagery part of that might be tens of millions a year at best, more likely a million a year, as they are going to use older imagery that hasn't been shot to spec and so is much cheaper.

Cellphone and broadband service is a large scale business.  Oil is a large scale business.  The business of space-based imagery seems mostly to be building satellites for spooks who use it when they can't get a drone in.

[docmordrid]

For example, the Gates Foundation relies on low-cost satellite imagery to count huts in Africa for population estimation to guide humanitarian efforts.

>
The business of space-based imagery seems mostly to be building satellites for spooks who use it when they can't get a drone in.

Except for the large number of farmers who use satellite imagry to keep track of their crops*, businesses, cities, and villages for muicipal needs etc. etc. My family includes farmers in WI, PA and MI and every one of them now uses satellite imagry.

* ex: Land O’Lakes (the dairy & butter co.) bought GEOSYS, and resells imagry to farmers through agricutural retailers.

[Mader Levap]

  The business of space-based imagery seems mostly to be building satellites for spooks who use it when they can't get a drone in.

Ahahahaha. No.



How you can write that just after posting this very video? Did you even watch it?

[IainMcClatchie]

That video is supposed to be thrilling, but it doesn't say much quantitative, except in the title.

At 3:10, it says that the GSD of WorldView 4 is 30 cm, which means that it "can clearly see a single sheet of paper placed in a parking lot".  No.  A single sheet of paper would be barely resolved.

At 3:35, Dan Jablonsky talks about how his satellites, at 30 cm resolution, can see lane and turn markings on roads.  Supplying that imagery might be a real business, except all the companies that do road mapping do it from cars, not satellites.  This Vox article says that Google and competitors must spend $1-2 billion/year on road mapping.  That seems a bit high.  This NY Times article mentions that the Google team in Hyderabad that converts imagery to the lane graph is "over 2000" people.  2000 entry-level workers in India is more like $10m/year.  Obviously the overall program costs more than that but not two orders of magnitude more.  [I led the team that built Google's StreetView cameras.]

At 9:20, DG is giving imagery to Gates foundation at a huge discount because otherwise the satellites would be unused over those areas.

At 11:00, Dan Jablonsky says "Our largest customer is the US Government."  Which is what I said.  The next biggest customers are other governments.

After that is mapping companies.  The mapping companies do not pay the same as USG does for imagery, because they typically buy older imagery and don't specifically task the satellites.  This Forbes article says Microsoft spent $130 million getting a 30cm GSD basemap of all of Europe and the US, and that was very expensive because they did it from airplanes.

At 13:18 comes the most important quote in this video.  "Our customers are demanding more actionable information..."  That's the business.  To make money in the imagery biz, you have to deliver information which makes it possible for your customers to make even more money than you are.

And then at 19:00, this becomes a banner for Maxar and a plea for the public to pay for lots of programs that will benefit Maxar.

In short, this video does not suggest how space imagery can be even a multibillion dollar a year business, let alone a trillion dollar a year business.

[Mader Levap]

At 11:00, Dan Jablonsky says "Our largest customer is the US Government."  Which is what I said.

Nope. You said it is for spooks. Big difference.

[IainMcClatchie]

At 11:00, Dan Jablonsky says "Our largest customer is the US Government."  Which is what I said.

Nope. You said it is for spooks. Big difference.

You're right, that is what I said.  To my mind, spooks pretty much are the USG, from an overhead imagery revenue perspective.

The largest single non-spook US government customer that I know of is the National Agricultural Imagery Program [Edit: looked at old notes, forgot about USGS, duh].  NAIP shoots the entire US every three years from airplanes, and contracts with the six largest aerial imagery collection companies.  They shoot leaf-on, which is the opposite of what everyone else wants.  Because they shoot when most everyone else is idle, and because they are so large in scale, they pay less per area than every other (non-spook government) customer and (to date) the imagery shot is promptly put in the public domain.

Interestingly (for me), the various government agencies that NAIP coordinates with are having trouble working together (article).  The result is that NAIP is probably going to move to a model where they pay providers less, but the imagery is only usable to the US Department of Agriculture.  The other government agencies will have to pay the providers for their own licenses, and the overall revenue for the same imagery collected will go up.  Folks who used to use the freely available imagery are out of luck.

Edit: added USGS budget.  This got much bigger in 2017 due to $22m of Landsat and Sentinel 2 budget added.  I'm not sure where that was before, maybe NASA?

NAIP 2017 budget:               $24.1m.
USGS 2017 imagery budget: $55.1m
NGIA 2017 budget:           $4900m (estimated, classified)

Obviously just some of the NGIA budget is spent on collection, but it's a sizeable fraction.

My estimate is that spooks spend an order of magnitude more on imagery than other USG customers.  Budgets for non-spook imagery are decreasing rather than increasing, and I suspect classified imagery budgets are down from the all-time high of the 1967 - 1984 KH-9 program, whose budget and some of the imagery collected are the only things still classified about that program since my friend Phil Pressel got it declassified.