The U.S. Air Force’s $150 million F-22 Raptor stealth fighter, which made its combat debut last week, is the most advanced operational warplane on Earth. But under the Raptor’s hood is dated computer hardware that’s creakier than your grandpappy’s cellphone from last century.
Though initially conceived as state-of-the-art technology in the early 1990s, some of the Raptor’s antiquated processors run at 25Mhz. That’s about 56 times slower than the multi-core chips in the iPhone 6.
The Air Force is making upgrades to the F-22 by adding better processors and newer fiber-optic data-buses, but the technology remains hopelessly obsolete compared to current-day consumer devices. Part of the reason for that is it is more difficult to predict how emerging technologies will develop. “Figuring out which technologies pay their way on a commercial product is generally easier than figuring out which technologies offer the best bang for the buck on a military one,” said analyst Richard Aboulafia at the Virginia-based Teal Group.
Not only are tens of billions in Pentagon cash at stake. The F-22, after decades of development, is finally becoming relevant to American security and geopolitics. The Raptor is flying missions over Syria, yes. More importantly, the Raptor—which was designed to be extremely stealthy, fly at very altitudes and at supersonic speeds for extended periods—could play a major role in the faceoff between Russia and the United States. In the case of major conflict, perhaps in Ukraine, the Raptor would be used to clear the air of enemy fighters and to smash enemy air defenses and thereby opening a corridor for friendly conventional warplanes.
The F-22’s archaic processors are just one of the ways that untold billions of taxpayer dollars are wasted by the Pentagon’s ponderously slow development processes. Oftentimes the Defense Department has to special order tiny batches of obsolete parts from questionable sources—which have occasionally been traced to places like China—to sustain these antiquated technologies. That is because those parts are no longer made in the U.S. Further, in most cases a single big contractor that developed the hardware is the only entity that is able to modify or upgrade a system because of its proprietary nature—often at exorbitant cost to the taxpayer. Taken together, billions are squandered every year.
Even the Pentagon’s newest stealth fighter, the nearly $400 billion Lockheed F-35 Joint Strike Fighter, isn’t immune. The F-35 has much better, newer, processors than the F-22, but even before the jet enters service, those electronics are being upgraded. “The F-35 mission systems architecture is far more flexible and upgradable compared to the F-22,” one industry source told The Daily Beast. “That was one of the ‘lessons learned.’ For example, the F-35 will likely get new processors in its mission computers down the road.”
So why doesn’t the Pentagon have always have state-of-the-art computers operating in all of its supposedly high-tech gear? The answer is that the Defense Department has a painstakingly slow process to write up its requirements before it builds anything. “Systems like the Raptor tend to have ‘dated’ components like processors because technology can advance rapidly after requirements are set,” one senior Air Force official told The Daily Beast. “Requirements must be set at some point in order to integrate with the rest of the pieces of a complex system, among many other reasons.”
The problem is that processor speeds tend to double about every 18 months, thanks to Moore’s Law. The Pentagon’s process for drawing up requirements just can’t keep up. “So [the requirements process] can be a pain but the reason it was created was to increase oversight in order to prevent cost overruns and failure to deliver the desired capabilities on time,” the official said. “Unfortunately, DOD [Department of Defense] probably overreacted and we are slowly correcting our course and coming to a more reasonable balance.”
By contrast, the technology industry in Silicon Valley works in a completely different way from the Pentagon. A company like Apple or Google does some market research, figures out where there might be room to develop a new product and then lets the engineers loose to figure out how best to design a new device that costs a certain amount. It delivers products quickly and efficiently, but it’s not the same as developing a new war machine. “Military products often break new scientific ground and technological ground, which is very different than surveying the world of on-the-shelf technologies and deciding what you want,” Aboulafia said.
Both defense industry and military officials told The Daily Beast that Silicon Valley’s approach would not work for the Pentagon. “I don’t think you can draw an analogy between Google/Apple systems and some complex weapon systems DOD acquires... although I would love it!” said the Air Force official. “While there is certainly room for improvement, the integration of platforms such as today’s aircraft are far more complex than what Google or Apple does… Can you image if we lost two Raptors for every 135 flights?”
Nonetheless, even though it took more than two decades and somewhere close to $70 billion to develop the Raptor, the high-flying supersonic stealth fighter has no equals even with the Pentagon’s sclerotic processes. “When you take into account that even though the Raptor was developed over two decades ago, no other nation or industrial base can come close to matching an integrated system like it,” the Air Force official said. “An iPhone has far fewer subsystems and they don’t have to operate at temperature, acceleration, or reliability extremes.”
That being said, the Pentagon can draw lessons from Silicon Valley even if it can’t adopt the tech industry model wholesale. The sheer amount of redundant bureaucracy needs to be eliminated. “Reduce the number of organizations and people involved with acquisition planning, including requirements and specification formulation, as well as source selection,” one industry official said.
Industry officials said the Pentagon needs to let the engineers come up with solutions or what industry calls “capability-based” requirements or specifications. The requirement might state a general purpose like: This aircraft must be able to hit X target though Y anticipated threat. The Pentagon also needs to develop hardware incrementally and have a long-term plan to modify that equipment over time rather than build something in one “big bang.” Further, the Defense Department should give companies “extra credit” for coming up with innovations, the industry official said.
The Pentagon should also adopt “open architecture” standards so that one contractor does not have a complete monopoly over upgrading a weapon system through its lifetime. “In the case of processors, they could be easily removed and replaced by any company... thus increasing competition, lowering price, and speeding up modernization efforts much more affordably while keeping pace with our potential adversaries,” an Air Force official said. “I think you'll see this trend significantly increase over the coming years. In fact, some future Raptor upgrades could include primes other than Lockheed leading the effort. Breaking the proprietary grip on DOD systems will be a huge leap in the right direction.”
Some industry officials would go further and mandate common interfaces for equipment—similar in concept to how USB ports work on home computers. “[I would] have ‘orchestrated diversity’ in defense systems—share common subsystems and interfaces--to avoid ‘winner take all’ industrial base issues,” an industry official said.
Ultimately though, “figuring out military needs is somewhat harder than figuring out commercial market wants,” Aboulafia said.