Interview: two guys at the vanguard (part 1)
by Sam Dinkin
|It doesn’t matter what you design to. The shuttle is designed to five nines. They have lost two in a hundred.|
Renaud: How many times, at $100,000 [profit] a ticket and $50,000,000 in development costs, do you have to fly before you get your money back? How many people do you have to fly? You have to fly 500. If you’ve got a 1% chance of failure and carry five people at $100,000 a pop, you have a 100% chance of losing your money before you pay off your investor. Investors in risky ventures look for a 10x return on their investment. They look for tens of dollars on their investment dollar. They are looking for the big win. You have, in contrast, a pretty good probability of losing everything before you break even.
TSR: If you’re going to do tourism, you have to design to five nines?
Renaud: It doesn’t matter what you design to. The shuttle is designed to five nines. They have lost two in a hundred. They have a one in fifty demonstrated success ratio. All that matters is demonstrated. When people say they are going to design to five nines, the only way to do that with a straight face is to test every single component to that level of reliability. That’s not a $50- or $100-million program, that’s a multi-billion dollar program.
TSR: Now you are talking about aircraft design.
Renaud: No, you’re not, you’re talking about spacecraft, launch vehicle design. Aircraft are not designed to nine nines of reliability. They are designed to a very low reliability levels, but fault tolerance. It’s a very different kind of problem.
TSR: Can you build in the fault tolerance?
Renaud: Absolutely. You can’t say to somebody, “Here’s our reliability number,” because you haven’t tested for it.
TSR: The insurers are only willing to trust a study that says it’s better than everything that’s come before it so far.
Renaud: I don’t think they even are. Insurance is a numbers game. You base everything on size of probable loss, and probability of loss. Probability of loss is actuarial tables. We went to Aon Space and we said, “How much for liability insurance for our first flight?” They said, “The said maximum probability of loss times—”
Renaud: [Demonstrated] probability of failure. Which is why it drives you to flying in the middle of nowhere where your maximum probable third-party liability is really, really low.
Renaud: That’s just the way it works. So the way we looked at the market is that we just can’t make this business plan close. So we never really went after the tourism market. If somebody can show that they can overcome all these hurdles and break into the market, go for it guys. I went after a market I thought I could close the business plan on.
TSR: You are going after popping up—
TSR: Take some pictures, or do some science and come back down.
Renaud: We found a lot of people in the DoD who have a need for that mission.
TSR: You are accessing a new market?
Renaud: Nobody does imagery right now from a suborbital vehicle—very little.
TSR: For a while they had the SR-71 Blackbird going because they decided the Keyhole spy satellites weren’t enough. You’re saying, “Hey, you need real time imagery.”
Renaud: Well, let’s do it: When you want overhead imagery right now, if it’s over denied territory, the only way to get that today, is with a satellite. A satellite costs $1 billion. It’s a $200-million Delta launch. It’s a couple hundred million dollars of satellite development minimum. Plus the insurance.
TSR: You need a bunch of satellites.
|It’s no big shock that people would like a better way to get imagery. Something that doesn’t cost $1 billion and doesn’t have a very low yield.|
Renaud: Just figure one. It’s $500 million to $1 billion to launch one imagery satellite to get you sub-meter imagery resolution on the ground. You put that thing up in space. It goes around the planet. It circles a planet that’s covered in 70% ocean, 40% is in cloud cover, and 70% of the land mass is absolutely uninhabited. Only about 2% of the time is that satellite ever over something even remotely interesting—it’s a very, very low yield—and you only come over that spot with a satellite once a day if it’s in a polar orbit. Combine that with the fact that anybody with a laptop can get on the Internet and download the ephemeris [a table giving the coordinates of a celestial object at a number of specific times during a given period] for our entire constellation. They know exactly when the satellites are coming overhead. Denial and deception is a real problem.
TSR: Even thirty years ago people knew how to do that.
Renaud: It’s no big shock that people would like a better way to get imagery. Something that doesn’t cost $1 billion and doesn’t have a very low yield.
TSR: Yours would have a high yield.
Renaud: A high percentage yield. You only fly a suborbital imagery mission when you can get the pictures, when you want the pictures, and where you want the pictures. So your yield is 100%.
TSR: It’s also a lot cheaper so the interest on your billion dollars will pay for one of these suborbital flights whenever you need one. What is NRO’s budget anyway?
Renaud: That’s a classified number, but it’s a big number.
TSR: So there are a lot of satellites—
Renaud: In fact there are very few. When there’s a regional conflict that we are about to be involved in, we retask all the satellites: we move them around in their orbits. We have coverage, almost continuous coverage or high percentage coverage over one place, but that means you have very low coverage over every place else. So it’s no surprise that while we retasked all our satellites to look at Kosovo and the Balkans, that’s when North Korea restarted their nuclear program. They knew nobody was watching. They’re not dummies. What we have is national assets that are conflicted. They are at least at the theater level. Low intensity, theater-level operations are generally shortchanged on the national assets.
TSR: You are saying something different, too: “Even if they got all the assets, that still would not be as good,” as your product.
Renaud: During the Gulf War, how many Scuds did we get on the ground? How many Scud launchers did we get after they launched one? After we did photoreconnaissance of the area, they were gone! Because there’s a huge amount of latency, even if you could get the satellite overhead, the satellite imagery is not beamed down to the guy on the ground. The satellite takes the imagery and stores it, and then it passes over a digital download place, like Guam or White Sands or something. It goes to Chantilly [Virginia] and is printed on these giant printers, driven out to Dover [Air Force Base in Delaware], then put on a C-5 and flown to the theater.
TSR: So you can find out what happened last week.
Renaud: You’re not getting stuff that is minutes old. Even if you can improve it to where there is network connectivity, gigabit over the long haul is lot of bandwidth. You are suffering a bandwidth choke. It is very tough to get that in a timely manner—and I am talking minutes—into a theater commander’s hands.
TSR: Suppose instead you had suborbital over the theater. You could just have an optical link to headquarters or just burn a DVD while you are up there.
Renaud: It’s a ten-minute mission, bring it back and hand it to the guy. You slide it into the jacket and hand it directly to the analyst. Ten minutes. We are talking a ten-minute latency.
TSR: Once you get the optical link going, you can have real-time data. You can launch your next strike before you land your suborbital rocket.
Renaud: So that is something that a theater-level commander would love to get his hands on.
TSR: So we are spending what, $50 billion, $500 billion in Iraq?
Renaud: Oh God, we just spent… isn’t the supplemental for $82 billion? And that’s just for this year.
TSR: If you took $2 billion to spend on this kind of intelligence, how many flights would it be?
Renaud: A lot. Thousands, tens of thousands.
TSR: So you could have snapshots every ten minutes during every critical mission. Tora Bora, looking straight down at the mountains.
|From 80 kilometers up, you can see almost 1,600 km, which means from Taiwan you could see the majority of the populated areas of China.|
Renaud: You don’t really need imagery that often. When something changes, you need imagery of it. When somebody is going to mass troops and do an attack, you need imagery of it. If you fly a strike and drop weapons on somebody, you want to be able do rapid battle damage assessment so you know whether you have to send more pilots back into harm’s way. That kind of stuff where it’s not really intelligence that the strategic people need to worry about. It’s purely tactical. The theater-level commander has a need for local reconnaissance that he has a tough time filling.
TSR: You say theater, but we are spending a lot on theater this year.
Renaud: There’s more theaters than just Iraq. Don’t you think that PACOM [US Pacific Command] would like to have some imagery of North Korea?
TSR: It would certainly be nice to send one of those once a month.
Renaud: From 80 kilometers up, you can see almost 1,600 km, which means from Taiwan you could see the majority of the populated areas of China. Don’t you think the Taiwainese would really like to know what’s going on in China right now?
TSR: What you’re saying is that the guys who did these orbital satellites had this low tech low cost option that they just could have taken their X-15 or their sounding rockets, and made something reusable out of it and done this instead of the national fleet of satellites? Wouldn’t that be cheaper for all the theater support?
Renaud: Not really. There’s a problem with glass. To get decent sub-meter resolution imagery from 80 kilometers up you are not talking about something the size of a coffee can. A Black Brant can launch a coffee can. You’re talking about something the size of your desk or two desks. You’re talking about a meter aperture piece of glass—high gain optics. A large sized CCD, solid-state storage, power—
TSR: You are basically lofting an observatory.
Renaud: You are lofting a ton of stuff, literally a ton of material.
Renaud: There isn’t anything that will launch that in a cost-effective manner to 80 kilometers. There was—it was called the DC-X. The DC-X did not have inflight restart capability and was limited to about 45 kilometers. So you tend to get a little more performance with that capability.