Since the early 2000s, the internet has become interwoven into all aspects of economic life in developed countries, so much so that it is difficult to disentangle online from offline economic activity. The three traditional economic sectors that form the basis of the system of national accounts — agriculture, industry, and services — are all deeply penetrated by and densely connected by the internet. Emerging technologies like cloud computing and blockchain database structures will make that penetration deeper and those connections denser still. Thus, taking a cue from the computer industry itself, it might be time to reconceptualize economic activity from the old trinity of “agriculture, industry, services” to a new, internet-era trinity of “hardware, software, netware.”
“Hardware” is familiar as a term for physical computers themselves, but as the internet of things develops and more and more devices become “smart” devices with some degree of computing power built into them, it makes sense to think of all physical things as forms of hardware. Certainly some physical things will remain “dumb” and chipless, but such things are of declining economic relevance. Rather than persist in using an economic categorization that distinguishes dozens of categories for such dumb staples as nuts, bolts, and screws while lumping all mobile phones together in a residual category of communications equipment, it seems more productive (for conceptual purposes at least) to lump all physical things together as hardware, relegating dumb hardware to a residual category of a larger group of mostly smart (and getting smarter) devices. Think of a screw as a hardware device of smartness level zero.
“Software” is also a familiar term, but it has become to broad to be analytically useful when applied to the internet economy. Operating systems, productivity software (spreadsheets, word processors), and enterprise software (billing systems, customer relation management) are extraordinarily complex structures of computer programming that run into multiple gigabytes of code. It is difficult even to imagine the complexity of a system of several gigabytes of computer code. But many, very valuable smartphone apps are much smaller. Ride hailing (RH) apps, for example, generally take up less than 100 megabytes on installation, and other connectivity apps can be even smaller. That’s still a lot of code, but from a qualitative perspective the “value” in an RH app, however that term is to be defined, is not in the code as such. Users of productivity or enterprise software are purchasing a software service, but users of RH apps are purchasing something else: a kind of connectivity.
“Netware” may thus be used, on the linguistic model of hardware and software, to denote products that generate value by connecting people and/or things. Hardware is tangible and isolable; software is intangible, but also isolable. Netware is intangible and non-isolable. It is a new form of product that derives its value mainly from connectivity. Netware is a loose category of technologies that include social networks, search engines, online marketplaces, entertainment platforms, e-payments providers, sharing apps, ride-hailing services, and even multiplayer computer games. If you are the only person in the world who uses a given model of hardware (e.g., your own custom-built computer), you can still get value from it. If you are the only person in the world who uses a given model of software (e.g., your own custom-programmed enterprise system), you can still get value from it. But you can get no value from being the only person on a given social network. Online dating services live or die based on the availability of potential dates.
Of course, many products contain (and in the future it is likely that most products will contain) a mixture of hardware, software, and netware. But that does not compromise the analytical utility of the categories. For example, Apple sells hardware (the iPhone) that runs Apple software (iOS) that supports Apple netware (the App Store). The App Store is “useful” — it generates value — only insofar as it connects people to other people (or legal persons: companies) who sell things via and App Store. For providing that connectivity service, Apple exacts a fee of up to 30% fee for certain categories transaction. But that fee is not the only value Apple generates from the App Store. A large (though unknown and difficult to determine) portion of the value of the iPhone itself is attributable to its ability to access apps via the App Store. Thus Apple is in some degree a hardware company, a software company, and a netware company.
Though the value of netware may be difficult to calculate with publicly available economic tools, it is certainly large and getting larger. The world’s top five most valuable brands are all, to varying degrees, netware companies. Though the netware share of their brand values is unknown, all five companies are showing via their behavior that they believe netware-type activities to be crucial to their futures. They are all eager to develop, acquire, and/or invest in netware services, and they are all attempting to leverage their existing customer bases into social networks of one kind or another. At the most basic level, Apple, Google, Microsoft, and Facebook all offer short message services as a way to try to tie customers (socially) more tightly into their ecosystems. Amazon is an outlier among the five, but like the others is driving into netware fields like entertainment and e-payments. All are eager to build (and monetize) the social graphs connecting their customers. And all, as facts would have it, have invested in RH services — or are developing their own.
The near future of the personal transportation technosystem
Big technology companies are (in)famous for their incestuously overlapping investments in emerging technologies, so evidence of the “big five” netware companies investing in ride hailing should perhaps be taken with a healthy grain of salt. But it is at least worth noting that Apple has invested in the Chinese RH app DiDi, Google in Lyft, and Microsoft in Uber. Facebook is linked to both Uber and Lyft via its Messenger service, leaving only Amazon out of the running — for now. Amazon has invested heavily in autonomous (self-driving) vehicles (AVs). These may be initially intended to serve as delivery vehicles for Amazon’s online consumer marketplace, but once built they could also conceivably offer autonomous RH services. Of course, Apple, Google, and Microsoft are also developing AVs, as are leading RH services like Uber, Lyft, and DiDi.
In fact, several technologies are simultaneously coalescing around RH apps with an extraordinary level of cross-investment among major players: the apps themselves, AVs, and battery electric vehicles (BEVs). The leading RH apps Uber, Lyft, and DiDi all have AV programs. Tesla most prominently represents the cross-over between AVs and BEVs, though most major automobile manufacturers have programs to develop both AVs and BEVs, and many of the companies engaged in AV research envision their AVs as BEVs. General Motors has taken things one step further with plans to roll out an integrated RH service in San Francisco based on a fleet of autonomous BEVs in 2019. It is this last model, the “all three” approach, that has perhaps the greatest potential to completely transform the personal transportation technological ecosystem.
A technological ecosystem or “technosystem” might be understood as a suite of technologies that reinforce each other through positive externalities in a well-circumscribed positive feedback loop. In other words, the technosystem has positive feedbacks among its components, but does not (at least, not “as such”) expand exponentially outward. For example, before the internet era, personal computers, operating systems, and productivity software formed a distinct technosystem, on which were hung peripheral components like printers and modems. Ultimately, of course, the tail would come to wag the dog, as the modem reduced the PC itself into a peripheral device of a much larger technosystem, the world wide web. The term “technosystem” is intended here to be heuristic, not strictly classificatory, but as a heuristic it is useful for hypothesizing an impending transformation in personal transportation.
Currently, the personal transformation technosystem is centered on individual car ownership, petroleum (gasoline or diesel fuel) power, and widely dispersed filling stations. The rise of RH apps has the potential to displace the first element, BEVs the second, and AVs the third. Individually, these displacements are not particularly transformative, and not mutually reinforcing. For example, RH apps, in principle, offer little more than an alternative to existing taxi services. The switch to BEVs has been driven mainly by environmental concerns and limited by range anxiety; from the perspective of a simple substitution of individual technologies, plug-in hybrids that include a backup petroleum engine have many clear advantages over fully-electric BEVs. And AVs are being promoted mainly as a productivity tool, freeing the owner to engage in other work or leisure activities in the course of moving from Point A to Point B.
Taken together, however, RH apps, BEVs, and AVs have the potential to radically transform the personal transportation technosystem, indeed to replace it with an entirely new technosystem. This is because the use of AVs in a ride hailing context solves the range anxiety, battery capacity, and battery charging problems of BEVs. In the RH-AV-BEV technosystem, individuals would order a car to meet a specific trip need. An AV from among the fleet of available AVs would respond only if it had sufficient battery life to make the trip. When an AV’s batteries ran low, it would return automatically to a charging station to recharge. It would even be possible for AVs to self-coordinate a daisy chain to ferry passengers over long distances, with hand-offs between AVs made, for example, at highway rest areas. Thus although any individual AV might have limited range, the RH-AV-BEV technosystem as a whole would be able to replace the existing personal transportation technosystem in its entirety.
The BEV industry has, to date, been developing on the model of personal ownership, which raises the need for public charging piles where people can charge their cars while away from home. The RH-AV-BEV technosystem obviates the need for public charging facilities — and along with it the need for people to wait while their cars charge. Current charging times of a minimum of 30 minutes are clearly too long for BEV charging piles to operate on the model of petroleum service stations. Thus charging piles have been associated with urban parking spaces. But the need for public charging piles effectively limits BEVs to urban or metropolitan use under the current private ownership model. By contrast, when BEVs are run as AVs in a ride hailing environment, no user ever has to wait for a charge. The car charges; the user moves on.
Similarly, the RH-AV-BEV technosystem solves the most serious challenges faced by the ride hailing industry. Ride hailing faces risks from the potential criminal behavior of its drivers. It also serious regulatory risks relating to its treatment of drivers as independent contractors. Both of these risks potentially disappear with the transition to AVs, as do the costs of compensating the drivers themselves. On the other side of the ledger, the RH-AV-BEV technosystem would dramatically increase the capital intensity of the ride hailing industry, as ride hailing companies develop their own in-house car fleets. Alternatively, the RH apps could contract with regional AV car fleet managers on a franchise basis. Either way, the RH app would remain brand recognized by the consumer and the lead firm in the personal transportation technosystem. And either way, automobile manufacturers would be reduced to something resembling OEMs — unless, like General Motors, they develop RH apps of their own.
Peripheral technologies of the RH-AV-BEV technosystem
Like any technosystem, the emerging RH-AV-BEV technosystem does not exist in a vacuum. Myriad social and political factors shape its prospects and will determine its future. But it also exists within a larger technological world, and parallel developments in closely related technosystems seem likely to favor the transition from a personal transportation technosystem based on individual car ownership, petroleum power, and filling stations to one based on ride hailing, BEVs, and AVs, especially in China. These are the ubiquity of high speed rail connectivity and the transition to smart power grids.
High speed rail (HSR) in China has ballooned from a single 113 km. (70 mile) demonstration line for the 2008 Beijing Olympics into a 25,000 km. (15,500 mile) nationwide system in 2018. The original Beijing-Tianjin line took three years to build, but once the technology had been mastered, the remainder of today’s nation-wide system was built in just ten more. China’s original HSR plans called for the creation of a “4+4” network of four north-south and four east-west main lines; this is now close to completion. As a result, plans now call for an extended “8+8” network of 38,000 km. (24,000 miles), which is projected to be operational in 2025. Later improvements are likely to focus on speed rather than distance.
China does not explicitly define exactly what constitutes “high speed” rail, and aside from occasional press releases, definitive statistics are hard to come by. Until recently, contradictory press releases often reported different system lengths at the same time. But official statistics now seem to designate trains designed for speeds in excess of 250 km/h (155 mph) as “high speed,” while on the major intercity routes, Chinese HSR trains generally operate at 300 km/h (186 mph) with a top speed of 350 km/h (220 mph).
The existence of a nationwide HSR network for medium distance travel, supplemented by a well-developed commercial aviation system for medium and long distance travel, reduces the utility of car ownership for personal travel between metropolitan areas. Simply put, it is faster and easier to take the train. More than that, the development of an RH-AV-BEV technosystem would further reduce the attraction of personal car ownership. Since RH apps can integrate with GPS and train timetables, it is already technically feasible for ride-hailing apps to offer seamless door-to-door services link intercity HSR travel with metropolitan automobile connections on each side (though it seems that, as of time of writing, no major company currently offers such services). The synergies between HSR and ride hailing are even closer than those between air and ride hailing, since trains run on firmer schedules than airplanes and GPS can be used to continuously track terrestrial travelers’ progress. Travelers can already book train rides from their car seat and car rides from their train seats; presumably, in the near future, they will be able to book a car to train to car trip as an integrated service.
More subtly but perhaps even more powerfully, the development of smart electrical power grids is strongly synergistic with RH-AV-BEV technosystems. Smart grids are online systems for continuously managing power usage at a very localized level. They are considered indispensable for the shift from conventional baseload power generation to “green” energy sources that may be intermittent, like wind and solar. They also enable consumers to sell power back to the grid, for example from solar cells on their roofs. Smart devices on smart grids can, in theory, continuously adjust their electricity consumption based on local and system-wide needs; for example, a dishwasher could be set to run automatically overnight at the optimal time of low grid demand, triggered by a signal from the smart grid itself. Similarly, personally-owned BEVs could be programmed to charge at optimal times and rates (amperage) in coordination with the smart grid, and through the smart grid even in coordination with each other. Neighbors’ cars could dynamically “take turns” charging in such a way as to minimize overall stress on the grid.
More broadly, the enormous aggregate storage capacity of all the BEVs connected to a city’s smart grid could be used as a kind of energy reservoir for the whole system. In the individual ownership model, there is little to be gained from this possibility, since energy spikes are not likely to occur in the deep overnight when cars are plugged in at home. The advantage of pooled BEV storage would, practically speaking, be limited to the systematic draw-down of residual BEV charges during the few hours of peak demand when people return home from work with some leftover power in their BEV batteries. But in an RH-AV-BEV technosystem, unoccupied BEVs could be wirelessly signaled to return to charging stations at any time during the day to contribute electricity to the smart grid. The entire BEV fleet would become part of the smart grid, not just when parked overnight, but throughout the day as well.
Though HSR and smart grids are the clearest examples of adjacent technosystems that are likely to intersect with RH-AV-BEV, additional connections are also likely. Delivery services of all kinds are an obvious example. Social networking apps might integrate ride hailing into meeting services. Gamification is always a possibility. Though the full suite of applications can only be imagined, the general implication for value chains is clear. When it comes to netware technosystems, the lead firm is likely to be the app.
The possibility that netware companies might induce a revolution in personal transportation that restructures global value chains in such a way as to turn today’s branded integrated manufacturers into OEMs is not mere fancy. It is already happening: in China, to bicycles. As a form of personal transportation, the bicycle is obviously much simpler than the automobile, but from a technosystem standpoint the principle is the same. In the hardware-based, pre-netware bicycle technosystem, individual riders owned and maintained their own personal bicycles. Public bicycle rentals via automated docking stations made some inroads in a few cities, but overall their impact was limited and certainly not transformative. Then, in 2015, Mobike introduced a netware-based model of dockless bicycle rentals in Beijing. Other companies soon followed, so that within three years most large Chinese cities were saturated with bright orange, yellow, and red bicycles for hire.
As a result of the disruption caused by these netware apps, consumer sales of branded bicycles in China have collapsed. It took just three years for netware companies like Mobike, Ofo, and oBike came to come to dominate self-powered personal transportation in China. These companies still buy bicycles — millions of them — but their suppliers have been reduced from high-margin branded manufacturers to generic OEMs. Ironically, Giant, the world’s leading branded bicycle manufacture, started out as an OEM for Western brands, then developed its own brand, but is now once again a supplier of generic bicycles — to China’s Ofo. Apple’s lead OEM for the iPhone, Foxconn, is also a major OEM for Mobike. When Giant, a manufacturer of bicycles used in the Tour de France, finds itself in direct competition with Foxconn, it’s fair to say that the bicycle industry has been severely disrupted.
Other industries that are ripe for similar netware disruption include retail, restaurants, and hospitality. The e-commerce giant Amazon has long threatened traditional retailing, as has its Chinese competitor Alibaba. The key limitation of e-commerce has always been the cost and inconvenience of delivery. Custom-designed delivery AVs are likely to solve that problem. Restaurants, too, are already threatened by food delivery apps that rely on bicycle messengers for fulfilment. Food service netware companies like Deliveroo and Uber Eats may develop their own centralized kitchens, transforming the restaurant delivery business into a full-menu custom cooking technosystem. Netware app Airbnb has already disrupted hospitality in ways that are transforming entire city centers, their residential patterns, and their real estate markets.
Not for nothing, then, are netware firms able to garner such extraordinarily high stock market valuations. If Apple were priced as a hardware company, Google and Facebook as managers of online advertising, Microsoft as a software company, and Amazon as a retailer, it would be difficult to justify their current stock market valuations, to say nothing of those of loss-making apps like Uber and Lyft. But as netware companies that have the potential to become the lead firms of the 21st century’s dominant value chains, they may be seem more reasonably valued. And if the recent past is any guide, each of the 21st century’s internet value chains is likely to be dominated by just one or two globalized lead firms. Both emerging as well as already existing technosystems are being reorganized around netware, and that netware is usually global. But if you want to track the pace of globalization in the netware age, don’t look for data from the System of National Accounts. Just look at your phone.