Thoughts
on Commercial Space, Part IIC
Commercial
Space 2.0, Big LEOs and EELV
The next wave of
commercial space began in 1991 with the establishment of Iridium. The idea was conceived in the late 1980’s by
engineers at Motorola to provide satellite phone service to the globe. Iridium is named for the 77th
element in the periodic table because the constellation would have 77
satellites. It was subsequently reduced
to 66, six orbital planes of eleven satellites each. [The sixty-sixth element is called dysprosium,
perhaps a more apt name in hindsight!] The
project attracted investment from many industry giants including Lockheed
Martin (LM), Sprint and Sony. But it was
Motorola that footed most of the bill. Iridium was the first of the low earth
orbit constellations to be proposed and one of the few to be fielded.
The
first launch of five satellites was in May 1997 on board a Delta II rocket from
Vandenburg AFB. The launch campaign was an international effort with the
Russian Proton and the Chinese Long March each getting some of the
launches. The system was declared
operational in December 1998. As a major
space project, Iridium was a technical success.
Seven years from conception to operational service is respectable even
by today’s standards. But nevertheless,
the company filed for Chapter 11 bankruptcy in August 1999. What went wrong?
Lots
of things. Fundamentally, Iridium was
unable to sell enough subscriptions. To
pay off the seven billion in initial investment, Iridium needed to sell one
million subscriptions within the first year.
They managed only fifty-five thousand.
It turns out that while Iridium was spending billions and taking years
to deploy a sixty-six satellite constellation, other companies like Verison and
AT&T were installing cell towers fast and cheap and building a customer
base with small pocket sized phones and low cost deals. In contrast, Iridium’s phones were large and
clunky and the cost per call was many times that of the cellular
competitors. True, you could take a call
using Iridium in a remote canyon in a remote region of the world, but most of
the customer base didn’t need that capability.
I was
at Lockheed during the Iridium years and I remember the excitement. Lockheed built a special purpose factory to
assemble the satellite bus for Iridium.
Lockheed lost close to one hundred million dollars when the wheels came
off, a significant sum, but far less than the billions lost by Motorola.
The
excitement generated by Iridium attracted many copycats and competitors. Globalstar was also formed in 1991 as a
partnership between Loral and Qualcomm. It
launched its first satellites, built by the Italian Aerospace company Alenia,
in 1998. I remember touring the Alenia
plant outside Rome sometime in the late 1990s.
Globalstar production was already complete, but the idea of rate
production of dozens of identical satellites and the manufacturing efficiencies
gained thereby has stuck with me ever since.
Sadly, Globalstar also filed for bankruptcy—in 2002.
Motorola
had grand plans for more Iridium-like constellations. M-star and Celestri were the most
ambitious. I remember hosting several
Motorola executives at the Lockheed facilities near Denver in the late
1990’s. I was working on the Atlas V
development program at the time and we couldn’t afford to ignore a potential customer
needing to launch hundreds of satellites. But the Motorola representatives insisted
on pricing far below market pricing at the time. I remember being lectured that Motorola’s
buying power would force us to play ball or go out of business. Atlas V is still going strong. All of Motorolas’ space ambitions are
ashes—just saying…
Perhaps
the most ambitious of the big LEOs of the nineties was Teledesic. An
independent company unassociated with any large aerospace prime, Teledesic
nevertheless had some impressive backers including Craig McCaw, Bill Gates and the
Saudi Prince Alwaleed Bin Talal Bin Abdulaziz.
Their concept went beyond satellite phones to broadband services, the
internet in the sky. The Teledesic
constellation was originally 840 satellites, not including spares. The figure below shows what this might look
like. We used to joke that Teledesic
would “darken the sky.” In July 1997,
Teledesic selected Boeing to be the prime contractor, beating Lockheed, and
Boeing engineers quickly resized the constellation to 288 satellites.
For a time, I was the Lockheed
capture manager for Teledesic launch services.
This was before I was drafted onto the Atlas V development program. Our solution was a modernized version of
Titan, a configuration that was much like the never fielded heavy version of
Atlas V. One feature of our vehicle was
the composite payload fairing (PLF) supplied by the Swiss company Contraves
(now part of Ruag). That feature did
eventually get incorporated into Atlas V.
I took many trips to Teledesic’s headquarters on the shores of Lake Washington
near Seattle.
A
unique aspect of launching these large constellations was the requirement for a
dispenser, the structure within the PLF to hold and deploy multiple
satellites. In one of our briefings of
different concepts we showed a graphic of a huge pez dispenser (remember pez
candy?) with Bill Gates’ head on top. Got
big laughs. I wish I had saved it.
Once
I moved to Atlas V, the marketing responsibility moved to Lockheed
International Launch Systems (ILS) and the solutions shifted to Proton (ILS,
since sold by Lockheed, continues to be the marketing arm of Proton) and Atlas
V. To reduce the cost per satellite as
much as possible, we created a whole new version of Atlas V consisting of a 5
meter PLF (from Contraves) and strap-on solid rocket motors (SRM). The original configuration had up to 4 SRMs,
but in a frenzy of last minute marketing/engineering we increased it to 5 to
squeeze a few more satellites and lower our cost even further. Increasing to six would have required
wholesale design changes due to the location of the engine feedline and
avionics pod. My systems analysis team
was responsible for figuring out how to fly such an asymmetric configuration,
and convincing management that we weren’t completely nuts. Furthermore, I convinced them that we could
fly with any number of SRMs, up to five.
To date, all versions: 0, 1, 2, 3, 4 and 5 SRMs have successfully flown—although
I admit it is somewhat disconcerting to see the single SRM version crab off the
pad. (I know it will work
just fine…)
Teledesic
raised a billion dollars (of the estimated $9B cost) and managed to launch one
test satellite (in 1998). However, faced
with the direct evidence of the business failures of Iridium and Globalstar,
Teledesic folded in 2002. By then, ILS
had managed to sell them a single launch service, which was eventually executed
by ICO in 2008, a venture founded by Craig McCaw from the wreckage of Teledesic.
The
Evolved Expendable Launch Vehicle program
The US Air Force (AF)
EELV program began in 1994 with the release of the Moorman Report. A team, led by General Tom Moorman, conducted
a study to determine what to do about the military’s aging and increasingly
expensive launch fleet. At the time, the
military used three separate systems for launch depending on the size of the
spacecraft. Delta II was used for the
smaller satellites like GPS. Titan was
used for the largest satellites including many of the most sophisticated spacecraft
for the Intelligence Community. Atlas
was used for spacecraft in between the capabilities of Delta II and Titan. Moorman presented four options: 1) maintain the existing systems; 2) evolve
the existing systems; 3) develop a new expendable system; 4) develop a new reusable
system. The Air Force chose option two
and in 1995 released the first Request for Proposal (RFP). Four companies were awarded 15-month study
contracts: Lockheed Martin, Boeing, McDonnell
Douglas and Alliant Techsystems, with the idea that just one would be the
eventual winner.
Phase II of the program
began in December 1996 with the competitive selection of Lockheed Martin and
McDonnell Douglas for $60M Pre-Engineering & Manufacturing Development
(pre-EMD) contracts. It was at the
beginning of the pre-EMD phase that I joined the program. The circumstances were interesting and worth
relating. The Lockheed Martin concept was
a family of vehicles based around the Common Core Booster powered by the
Russian RD-180 engine, Atlas V. During phase I, the LM team had struggled to
meet the mass to orbit requirement of the heavy (three core) version of the
system. However, the severity of the
performance shortfall was not fully communicated in LM’s proposal for the
pre-EMD phase. After award, the AF
discovered the issue, was justifiably furious and threatened to cancel the contract. LM’s response was to completely change the
management team while assuring the AF that the issue would be resolved
ASAP. I was part of that “second wave”
as we called it. The “first wave” was
shown the door. My first assignment was
to lead an emergency performance improvement program.
In 1997, in the face of
the general industry optimism with respect to the commercial launch market and
the big LEOs, the Air Force decided that they would keep two winners. Furthermore, instead of a single winner receiving
a $1.6B development contract, each winner would receive only $500M and had to
make up the balance of the cost with private investment. Clearly, this was far less attractive to
Lockheed Martin and Boeing (who had just bought McDonnell Douglas). A senior Lockheed executive who complained
was told to play along or else risk other more lucrative AF business like the
F-22. Those of us in the trenches had
drunk the big-LEO Kool-Aid and wasted no time building business models to
justify the investment.
In June 1998, two
development contracts of $500M each were awarded to Lockheed Martin and Boeing.
These contracts were under the Other
Transaction Authority (OTA) of the FAR. This innovative acquisition approach
levied top level performance requirements but left the details up to the
contractors. Insight was achieved by a
small government team. (This approach was much like what NASA did later for
COTS). In addition, the companies
submitted proposals for the initial 28 missions. These were awarded in October 1998, nineteen
going to Boeing and only nine to Lockheed Martin. Those of us on the Lockheed side were shocked
and bitterly disappointed. With both
Atlas and Titan, we were the incumbents and had the expertise for the big
complicated missions. The other guys had
only ever launched the small Delta II.
It turns out Boeing had
offered very low prices predicated on huge production rates of dozens per year
serving the commercial market. Lockheed
had done the same but was a bit more cautious.
For example, Boeing built a new manufacturing plant in Decatur Alabama
sized to build 40 Delta IV cores per year.
Lockheed facilitized for “only” 19 Atlas cores. The precision of the number 19 is an
indication of the false rigor we had put into our market analysis.
Much drama ensued over
the next few years. Having won no heavy
missions and having no commercial demand for the heavy, LM renegotiated its
development contract to eliminate both the heavy and the west coast capability. Boeing received a heavy demo launch and LM’s
west coast launches in consideration. Later,
Boeing was found to have obtained and used LM pricing data in its
proposal. This violation of the
Procurement Integrity Act cost Boeing some of its west coast launches and a new
contract for LM to develop west coast capability. As you can imagine, animosity between the two
contractor teams was intense.
Nevertheless, the
technical work went well. The first
launch of Atlas V occurred in August 2002 from complex 41 at the cape. It was one of the highlights of my
career. By then, I had become the Chief Systems
Engineer and Director of the Systems Engineering and Integration Team. It was my responsibility to ensure all the
myriad different pieces worked together as intended. We launched on the first launch attempt at first
second of the launch window. The Eutelsat
commercial communications satellite was deployed perfectly in its intended
orbit. Amazingly, there were no
significant flight data findings.
Everything worked just as predicted.
I was at the press sight giving interviews during the launch. My most memorable quote was “Whoo hoo” as we
got word of the successful spacecraft separation.
I could keep going. The history of EELV continues to this day
amid even more drama, some of which I will cover in subsequent posts. But let
me take stock of where the program stood at the time of the first launches in
2002 (Delta IV first launched in November 2002). From the perspective of the US government,
the program was an amazing success. It
got two independent launch systems, each capable of launching most or all of
the required missions. They invested a
total of $1.2B in the development, including the initial phases. In return, they extracted about $4B in contractor
investment. All the cost risk was born
by the contractors. Each system was
fielded on time and worked perfectly the first launch (and have been
essentially 100% successful since). The
prices for the initial 28 launches were less than the 50% cost reduction target
from the heritage systems. A great deal for the taxpayers.
From the contractor
perspective, things were not so rosy.
Both had invested very large sums in the expectation of a burgeoning
commercial market. By 2002, however, the
big-LEO bubble had collapsed—in conjunction with the dot com bubble. Iridium and Globalstar had declared bankruptcy,
Teledesic had folded its tent. Both
Boeing and LM counted on this market to recover their investment and to support
the prices they bid on the initial 28 AF launches. Making matters worse, the 28 launches were
already starting to slip as government satellite development programs experienced
massive delays. Eventually, some of
these slips exceeded 10 years!
In summary, both
companies were hemorrhaging money with no clear way to stop the bleeding. It took four years, threats to exit the
business, wholesale contract restructuring, and the shotgun marriage of the two
company’s launch divisions into United Launch Alliance (ULA) before the
business situation stabilized. Even
under ULA it was only in the last couple of years that the Atlas V investment
was recovered. With ULA’s plans to
retire Delta IV, Boeing’s full investment will never be recovered.
Coincidently, 2002 was
also the year that SpaceX was founded.
Over time, that small startup became the next disruptive force in the
launch market. But that is the story of
the next post, Commercial Space 3.0.
Lesson
Learned #3: Beware the terrestrial competitor who has a
scalable business model. Competing with terrestrial businesses from space is fraught
with peril. The terrestrial business
model is generally scalable. The cell
phone company can put in some cell towers, gain some customers (revenue), then
put in some more—on a timescale measured in months. The space business (in the mode of Iridium) must
get everything fielded before any revenue accrues. The timescale is measured in years.
Lesson
Learned #4: Be deeply suspicious of your own market
analysis. You should start from the
premise that all commercial space business models suck (See Lesson Learned #1). I still shake my head at the wild market
analyses we used to justify the investment into Atlas V and Delta IV. We were
assuming that the worldwide launch market would grow by 4 to 5 times its then size
in just a few years. The fact that everyone got it wrong is little
consolation. If it seems too good to be
true, it probably is. Compounding the
problem was that each side assumed they alone would garner the lion’s share of
that new market and priced accordingly. Obviously,
in the real world, the market will be shared among many competitors, some of
whom don’t have to recover investment or make a profit. (See Lesson Learned #2).