Lessons for the aerospace and defense industry
Aerospace was America’s original high-tech industry, and following World War II the prestige and patriotism associated with this burgeoning field captivated the world and attracted its brightest minds. Aerospace—and the closely related defense industry at large—has always been sensitive to fluctuations in government spending, but the post cold war draw down of the early 90s entrenched rather than reformed an industry that remains bloated, overly expensive, and inescapably co-dependent with its dominate customer: the United States Federal Government. To thrive in an environment that will be defined by looming defense sequestration and a longer-term entitlement crisis aerospace must learn to be agile, responsive, and affordable.
How did we get here?
The German V2 first flew in 1944, and by the late 60s America had deployed intercontinental and submarine-launched ballistic missiles as two legs of MAD’s nuclear triad, launched satellites to gather intelligence on enemies half a world away, and landed men on the moon. But by the 70s interest had waned: the last moon landings were cancelled and public attention turned to the war in Vietnam and economic turmoil at home. The 80s-era Shuttle briefly re-ignited enthusiasm, but even prior to the Challenger disaster launches went un-televised.1 Civil spending mirrored public apathy, space was too expensive to entice meaningful or sustained commercial investment, and so it was a large and relatively stable Cold War defense budget that filled the gap and birthed a military-industrial complex that was dependent on the government for its livelihood and upon whom America depended for its security—just as Eisenhower had warned.
Then the Cold War ended, the United States became the world’s only superpower, and even the conservative US President George H. W. Bush and UK Prime Minister Margaret Thatcher promised a peace dividend, a policy that was accelerated by the 1992 election of President Clinton. At a 1993 meeting of aerospace industry executives known as the last supper Clinton’s Secretary of Defense, William Perry, told the aerospace and defense industry that the government wasn’t going to prop them all up anymore, and it was time to consolidate or die—and they listened.
The wave of mergers and acquisitions encouraged by [the Department of Defense] changed the corporate face of the industry… Boeing took over Rockwell and McDonnell Douglas. Lockheed merged with Martin Marietta and absorbed Loral, which previously had acquired the defense electronics businesses of IBM, RCA, and a half-dozen other firms. Raytheon purchased the defense business of Texas Instruments and most of Hughes. General Dynamics—which had sold its missile lines to Hughes, its space launch division to Martin Marietta, and its fighter aircraft business to Lockheed—more recently bought two shipyards, Bath Iron Works and NASSCO, to complement its Electric Boat division, a builder of submarines. Northrop and Grumman merged, bought Vought Aircraft, and later acquired the radar business of Westinghouse. FMC and BMY, builders of armored vehicles and gun platforms, formed a partnership they called United Defense.
Aerospace was ultimately distilled to a trio of primes2 serving a single customer: a textbook example of dysfunctional market that is half monopsony, half oligopoly, and universally difficult to disrupt. The government’s monopsony is protected from competition by regulations that, for example, prohibit the sale of high technology and arms without government consent, while the oligopoly is shielded from new entrants by traditionally high capital costs and even more regulation. Capital barriers include high-value facilities, like anechoic and thermal vacuum chambers large enough to test school-bus-sized satellites and multi-story integration facilities for building heavy-lift launch vehicles. The human capital is a workforce of engineers and technicians with highly specialized skills that have to be home-grown or poached from elsewhere in the industry. Coupled with Defense Federal Acquisition Regulations (DFARS) that make the tax code look quaint by comparison and a promise of ever less lucrative future budgets (assuming Congress starts passing budgets again), the industry hasn’t been enticing to new entrants.3
For the survivors, consolidation also laid the groundwork for a demographic peculiarity that has stunted innovation and disruption within the remaining players. As the companies merged they cut costs—most relevantly by cutting junior staff and freezing hiring—and a quickly growing dot-com boom eagerly snatched up the young, available talent.4 Coupled with the similarly timed retirement of the original generation of engineers that founded the industry in the 50s and 60s, aerospace became a place where the key management and leadership roles (in both government and industry) were staffed by baby boomers, which encouraged many of the remaining Generation X and Y employees who weren’t let go to leave for industries where it’s less difficult to found a start-up and the corporate culture isn’t monolithically defined by a single generation. The Cold War’s existential but non-urgent threat encouraged complex, technical responses, and starved of a source of youthful disruption aerospace’s cultural homogeneity cemented a trend that was neither new nor previously unnoticed. As Norm Augustine, former CEO of Lockheed Martin, presciently joked in his 1983 book:
In the year 2054, the entire defense budget will purchase just one tactical aircraft. This aircraft will have to be shared by the Air Force and Navy 3½ days each per week except for leap year, when it will be made available to the Marines for the extra day.
Hyperbole, but entirely keeping with the theme of a Death Star acquisition philosophy that asks primes for multi-year, multi-billion dollar cost-plus proposals for omnibus solutions like multi-payload weather satellites, Joint-Strike Fighters, and floating cities.
This model of perpetually increasing costs is only marginally sustainable when defense budgets are growing and regardless of what happens with sequestration future cuts seem both likely and long-term. Luckily, a confluence of events mean this new reality represents an opportunity and not a crisis. The answer is for big aerospace to learn a lesson from the high tech start-up industry full of the 20- and 30-somethings it pushed away: release early, release often, iterate compulsively, and innovate like there’s a legion of competitors breathing down your neck.
When too much is asked of a system it’s easy to get trapped in a spiral of increasing complexity, an observation that applies as well to software as it does to satellites.5 For example, it’s commonplace for a large satellite to take a decade to design, build, and test, which causes two problems. The first is obsolescence: follow-on systems almost always have some level of redesign, either to incorporate advancements in technology or to design-out parts that can no longer be procured. The second is that long development timelines are self-sustaining: a 10-year and multi-billion dollar investment only makes sense if the system lasts a similar length of time, but that level of reliability increases complexity, which increases development timelines and cost, thus feeding the cycle.
The response has always been obvious: smaller, less complicated systems developed more quickly for less money. NASA’s faster, better, cheaper was a highly publicized example of this philosophy that has been unfairly panned as a failure, and since then advancements have made this approach even more feasible. The explosion of mobile computing has shifted the focus of Moore’s Law from just increasing raw computational power to balancing computation with size, weight, and power. The result is that flops-per-Watt-per-dollar is finally at a place where micro- and nano-scale satellites can be useful.6 The first signs of this revolution are apparent in the Cubesat phenomenon, which started as a grad school project at CalPoly and Stanford and has spread to universities around the world, NASA, and even Kickstarter (twice).7
The allure is that smaller busses8 shorten development timelines and significantly reduce costs. Cheap busses enable small payloads that don’t generate enough value to warrant their own satellite and currently have to arrange hosting as second-class citizens on larger satellites. Miniature satellites allow aerospace startups to cut their teeth developing flight hardware and software with smaller teams and smaller facilities, and they function as an affordable platform for getting flight heritage on new payloads and components, which increases their all-important Technology Readiness Level (TRL). Small satellites also support redundant architectures, because an entire constellation of small sats can be launched and replenished for the cost of one large one.
Cheap satellites also amortize over shorter lifetimes, which makes it economically feasible to deploy them in more quickly decaying orbits. Earth remote sensing payloads are typically diffraction limited; economics requires that these payloads be placed in relatively high low-Earth orbits and physics says higher orbits means larger apertures, which leads to large satellites. But cheaper, short-lived satellites have a business case at lower altitudes and can achieve the same resolution with a smaller telescope, or survive at higher orbits without expensive propulsion systems. The constellation approach also enables a trade of image quality for latency. The sub-meter resolution of Google Earth and Apple Maps flyover is (mostly) amazing, but imagine the utility if that data could be augmented (on demand) by 1–2 meter images that were only a few minutes old (that is, a trade of spatial resolution for temporal resolution).
And this is only what’s possible now. In time, innovation will respond with super resolution techniques, and perhaps in my lifetime the short-wavelength equivalent of synthetic aperture radar, which will revolutionize what’s possible from a small satellite. Likewise in satellite-based communications: conventional wisdom says communication satellites need to live in geostationary orbit and blast prodigious amounts of power to blanket entire countries with ubiquitous content. Innovation will respond with improved antenna designs and modulation techniques hosted on low-Earth orbit constellations. These architectures will cost less than the systems they replace and replace the broadcast model with one based on individualized distribution that uses myriad directed communication channels instead of a few very high power ones.
These possibilities are incubating a commercial (i.e. non-defense) market that offers retail list prices on everything from piece-part components to full satellite kits, and the promise of a growing market for small satellites is moving things on the launch vehicle side as well. There is renewed interest in medium-sized commercial launch vehicles like Athena, new development of small vehicles like LauncherOne, a growing secondary payload market (ISI Launch, Spaceflight Services) that matches the excess capability of large rockets with small payloads, and even movement within the government to look small with rocket programs like DARPA’s ALASA9 and SMDC’s SWORDS.
The next frontier
At hundreds of millions or billions for a bird and launch, it’s little surprise that large satellites only serve the needs of governments—and large corporations selling services like satellite-based television, radio, imagery, and telecommunications—just as in 1970 the government and large corporations were the only customers for computers. But as prices fell computers became affordable for the masses, and space is at the beginning of a similar trend. As the market grows so will the industrial base, which will attract capital investment and lead to more competition, lower prices, economies of scale, and accelerated innovation.
For too long the aerospace industry has been shielded from this path by the inherent challenge of space and a perverse market, but the long-awaited tipping point is nigh and is fortuitously timed with a politically and economically necessitated austerity that will challenge the status quo and—ironically—foster growth. Now, more than 40 years after humans first landed on the moon space is on the cusp of being accessible to normal people. The transformation will be slow and take longer to develop than the personal and mobile computing revolutions have because space, to an even greater extent than the sea, is terribly unforgiving of carelessness, incapacity, or neglect,10 but the trend is as inevitable as the progression of technology, whereby something that was once only possible with a herculean application of resources eventually becomes trite so that a new impossible becomes the next challenge.
The 32-month post-Challenger drought before the Shuttle returned to flight did nothing to keep the program at the forefront of people’s minds or stoke their interest in space exploration. ↩
There are now only three companies, Lockheed Martin, Boeing, and Northrop Grumman, that are considered aerospace prime contractors; that is, companies seen as credible responders to government requests to develop large integrated systems like James Webb Space Telescope, THAAD, or GPS. Other large aerospace and defense companies like Raytheon, BAE, and General Dynamics are in a continuous hunt to achieve prime status, but have not yet succeeded. ↩
SpaceX is an obvious (but solitary) counter example that’s succeeded for three reasons: it wisely limited its scope to launch vehicles (at least to start; it has since expanded into the Dragon cargo capsule, but only after success with the Falcon 1 and then Falcon 9, and the up-front capital expenditures have still been massive), it eschewed cost-plus contracting and the concomitant regulation in favor of a fixed-price business model, and it was founded by a genius, billionaire, playboy, philanthropist with the financial resources and constitution to finish what he’d started. Bully for SpaceX, but not a particularly repeatable formula. ↩
To this day, Aerospace doesn’t have a large enough pool of Gen X and Y employees to backfill the thousands of management and leadership roles filled with soon-to-be-retiring baby boomers. Succession and leadership planning is a pervasive focus of large aerospace companies, but the planning of top executives is routinely undercut by realities on the ground. The industry size is largely set by the value of government appropriations, which makes leadership opportunities a zero-sum game. Every young-in who is quickly promoted displaces a more experienced boomer, often into a newly created middle-management position. (Which isn’t to say people don’t get fast tracked, but the incentives of the displaced are hardly aligned with the broader strategy.) Over time, the ratio of management to individual contributors (as non-managers are called) has gotten top-heavy enough that it’s not uncommon to see people who are ostensibly Senior Managers with no one reporting to them. ↩
Spacecraft and weapon systems host an incredible amount of largely custom software—millions of lines of code is not uncommon. For comparison, the Linux kernel has on the order of 10 million lines of code, which helps explain why large aerospace systems are so expensive. ↩
Satellites run on solar power, operate in a vacuum (except for Russians designs, which often pressurized their electronics), and have to reject all the heat they generate through radiation, so low-power electronics are important. Satellite electronics are also subject to incredibly hostile radiation environments that require some combination of shielding, derating, developing unique rad-hard boards and components, and redundant architectures that are robust to or can prevent things like single-event upsets, latch-up, and single-event burnout. ↩
I backed this one but haven’t thought of what to tweet from space yet. Send me a reply @wuerl and I’ll pick five of my favorites. Remember there’s a 120-character limit, and leave room for your handle. ↩
The systems on a spacecraft that provide the basic support infrastructure, like power generation and storage, attitude determination and control, guidance and navigation, etc. are collectively known as the bus. The system(s) that provide the primary functionality, such as a telescope and camera for remote sensing or transponders, radios, and antennas for communications are known as the payload. The size and cost of the bus is an important factor in determining what types of payloads can be hosted on a spacecraft. ↩
The Defense Advanced Research Projects Agency (a.k.a. DARPA, who you may remember from the creation of the internet) is one of the more forward-leaning parts of the defense department, but the fact the government is involved—and in some areas leading—is perhaps the strongest argument for its inevitability. ↩
Co-opting the text of a poster in the student lounge of the Aeronautics and Astronautics Department where I did my undergraduate work, which used the same text to compare aviation to seafaring. ↩