By Peter Norton
What’s at stake:
The semiconductor industry faces mounting problems in sustainability, affordability and reliability in chip supply. We can tackle each problem individually, but we should also look for a much more consequential solution by going back to Econ 101: What’s the opportunity cost?
Chip shortages come and go, but one thing stays the same. Chip manufacturing is energy intensive. Its enormous environmental costs include “a huge carbon footprint.” While chips have countless energy-saving applications, from process optimization to remote work, many other uses are themselves energy-intensive.
In this latter fact lies an opportunity for a future of fewer chip shortages and more sustainable chip manufacturing. If we limit the most energy-intensive demands for chips, we can ease supply constraints while reducing the environmental burdens – on both the manufacturing and the application sides. And we have ample precedent for such limitations.
Faster, better, cheaper
The search is on for ways to improve sustainability, affordability, and reliability in chip supply. Like similar triads, this combination of goals is a daunting ambition. In the 1990s, when NASA told engineers to get “faster, better, cheaper,” some warned the agency that it could have any two, but not all three. NASA’s reluctance to accept such warnings was a contributing factor in the 2003 Columbia shuttle disaster.
Similarly, if we want (a) more sustainable chip manufacturing, and (b) affordable chips, and (c) a chip supply sufficient to meet soaring demand, we will be forced to choose any two. Avoiding the choice, or choosing to hope that amazing new techniques will let us choose all three, are in themselves bad choices.
Choosing two is tough, but there is a right answer. If demand accelerates much less than anticipated, we can forego (c). Supply will still have to grow, but at a more manageable pace. And choosing (c) yields a double payoff. By easing pressure on supply it limits the total environmental costs of chip manufacturing, and – with the help of some achievable public policy – it can do so in ways that also limit the environmental costs of some chip-intensive systems.
There are some obvious opportunities for demand reductions. For example, given the climate emergency, the energy demands of cryptocurrency mining are unjustifiable. Crypto’s contraction gives us cause to hope that this source of demand for chips may be under control. Planned, programmed obsolescence in phones or other devices is also indefensible, as are design techniques that needlessly deter repair or recycling. Both of these practices waste working chips. Italy has sued manufacturers over such practices, and France may soon follow suit.
Numerous other chip-intensive ventures have been poised to proliferate, threatening to intensify demand to the point that none of the three supply values – sustainability, affordability, or reliability – could be ensured. But the good news is that we’d be better off without some of them. If we forego or limit them, or at least ensure that public policy does not promote them, we spare ourselves from some of their undesirable effects while also relieving the stressed supply chain.
The greatest opportunities for win-win reductions in demand for chips lie in the automotive sector – where supply constraints have also been tightest. Automakers account for only about 15 percent of the world chip demand, but the share is rising and – with vehicle electrification and automation – may rise steeply. Some transportation futures are being sold to us in which demand for chips would be at least an order of magnitude greater.
Even ordinary cars without advanced driver assist (ADAS) features use a lot of chips. A car of recent vintage is likely to have over 1,000, many of them serving essential purposes in systems such as emissions control. ADAS that improves safety can entail hundreds more.
But much of the rising demand for chips in the auto sector is due to onboard systems that do not make cars safer or more fuel efficient. Some even make driving more dangerous. David Zipper has called attention to the fact that utilitarian touchscreens in cars are less useful to most drivers than simpler interfaces – and can be dangerously distracting. Even automakers have recognized the problem, and gone back to using ordinary buttons and knobs instead. Far worse offenders are in-car entertainment systems that drivers can access. They are opportunities for monetizable data collection that contribute nothing to passenger transportation. Some demands for chips actually make driving more dangerous. We would be foolish to try to manage extreme supply constraints while neglecting such needless demands.
Such status-quo chip demands, however, are negligible next to those of the transport futures that automakers and tech companies are trying to sell us. Setting aside the techno-utopian fantasies, engineers today are hard at work developing the systems necessary for a future of connected automated cars. The onboard automation systems in each car would alone demand hundreds or thousands of chips. Far more chips would be necessary to give the cars the connections necessary to making it all work together. Vehicle-to-vehicle (V2V), vehicle-to-infrastructure (V2I), and vehicle-to-everything (V2X) are all essential to the idea, and all require chips. To automate driving, cars would be equipped with numerous sensors, all feeding onboard navigation systems. Every car would maintain and constantly update high-definition (HD) maps. The possibilities and safety benefits of such connected systems have been oversold, and they would demand vast quantities of chips.
What’s the opportunity cost?
In assessments of future mobility, somehow an elementary economics concept has been neglected. We hear a lot about what’s technically possible. Is level-4 automation in a crowded cityscape possible? Can Cruise serve San Francisco safely? Can robotic vehicles learn to interact with pedestrians? Framed this way, each of these questions is answerable with a qualified “yes” – and this answer, in turn, looks like a green light to proceed. But before we go, we must first ask the question that Econ 101 would have us ask: “What is the opportunity cost?” Given limited resources – including limited chip supplies – what are our alternatives and which delivers the most transport for our transport dollar? And among these alternatives, which ones offer us the best returns in terms of sustainability, health, and inclusiveness?
The opportunity cost of pursuing high-tech driving has been dismal. For well over a decade, the excuse has been that the future payoff will be so great that the losses would eventually prove to have been worthwhile. This unfalsifiable claim was far-fetched a decade ago. By failing to consider opportunity costs, we have put ourselves at risk of committing a second Econ 101 error: spending still more to chase irretrievable sunk costs.
Of course the private sector is free to commit such blunders. In several respects, however, this wasteful dead-end has been a public sector failure borne by the public generally. The early lead of the US in robotic car development began with public sector money, in the form of the DARPA Grand Challenges. Since then, the US Department of Transportation and other federal and state agencies have consistently celebrated and generously funded ambitions to automate driving, much as they did for an earlier failed effort to automate highways. Competing for business, governors have invited companies to try out automated driving services on public roads.
Above all, the futures we are considering are shared futures about a public service as essential as a safe water supply: everyday transport. The public support that has been extended to expensive but disappointing driving automation efforts has been justified on the grounds that better transport for all is in the public interest. This principle also justifies public efforts to ensure that the companies that stand to gain from this support do not squander it.
Precedents of government regulations under emergency conditions
If the US has set an example of supporting automotive enterprises’ ambitions, it also has a long record of regulating them in the public interest. Automakers have long had to meet federal fuel efficiency standards, emissions limits, and safety regulations. Despite numerous loopholes and inconsistencies, these regulations did press automakers to make safer, cleaner, and more fuel efficient cars – and in every case the public interest principle was sufficient to justify the intrusion. Indeed automakers sometimes welcomed regulations because in a competitive environment no one company could risk acting first – even in those cases in which all companies stood to gain.
Under emergency conditions, government intrusion has gone even further. The national 55 mph speed limit began as an emergency response to the OPEC oil embargo. In many states, to conserve fuel, access to gas stations was limited to certain days. And most extraordinarily, just a month after the attack on Pearl Harbor in 1941, the US government ordered an end to civilian automobile production as an emergency measure – effective just two months later. By then the US was already a car-dependent country; it had about two thirds of all cars worldwide. But Congress and the president, responding to an emergency, ordered Detroit to stop making passenger cars. And it did. Dealers had no new cars to sell for almost four years. To reduce fuel consumption, a national speed limit of 35 mph was introduced. These were emergency responses to emergency situations. The climate emergency we face today warrants at least a comparable emergency response.
The problems of chip supply and sustainability are of course global problems. Yet for these, too, we have ample precedents in the form of international agreements. The threat to the earth’s ozone layer led to the Montreal Protocol of 1987, and ultimately to a successful international effort to shift industrial practices into a safer path. It sets a precedent for an international agreement to ensure that chip supplies are not squandered on practices that endanger the public.
If we reviewed opportunity costs with the public interest in mind, and if we resist the temptation to chase sunk costs, we will find attractive paths to a more affordable, healthful, and sustainable mobility future – and one that requires even less chips than the automobile sector demands today. We have abundant precedent for the regulations necessary, provided we do not lose sight of the fact that basic transport is an essential service like other basic utilities. Through responsible public policy and international agreements, we can limit the proliferation of superfluous and sometimes dangerous onboard vehicle electronics. We can redirect public funding away from connected and automated cars and instead toward modes of transport that require vastly less chips – and which are also more affordable, healthful, and sustainable. We can relax the zoning rules and highway engineering standards that preclude proximity between destinations, thereby compelling people to drive.
Above all, we can recognize that good public policy in transport would give people choices – driving, transit, electric bike riding, or walking – rather than compelling most to use the most chip-intensive, energy-hungry, and dangerous surface mobility mode yet devised.
Peter Norton is the author of Fighting Traffic: The Dawn of the Motor Age in the American City, and of Autonorama: The Illusory Promise of High-Tech Driving.
Copyright permission/reprint service of a full Ojo-Yoshida Report story is available for promotional use on your website, marketing materials and social media promotions. Please send us an email at firstname.lastname@example.org for details.