Is a New Technological Revolution in the Making?
Image by TheDigitalArtist on Pixabay

History has shown a pattern of consecutive technological revolutions, each lasting some 50 years, and it’s about time for the new wave to become visible. Even though we should not interpret such historic patterns as a deterministic model, a set of emerging technologies nevertheless shows interesting parallels with the General-Purpose Technologies that drove the technological revolutions of the past.


  • Carlota Perez and others have argued gains in total-factor productivity in the past were driven by so-called General-Purpose Technologies (GPTs) such as the steam engine. One or more GPTs together led to new “techno-economic paradigms” that transformed the economy and everyday life, something we detailed earlier this year in the book From Luxury to Necessity.
  • According to theorists, General-Purpose Technologies are: 1) pervasive and spread to a vast majority of sectors, 2) continue to improve in terms of costs and performance, and 3) spawn a variety of applications and spin-off innovations.
  • The concept of Artificial Intelligence dates back to the early days of the electronic computing, and it has already seen some hype cycles. Even though some results may not live up to inflated expectations, e.g. in healthcare, and it still requires many experts to put AI to use, there is a clear potential to bring about significant productivity gains.
  • Google’s AI program AlphaZero can learn to play new games on its own and proved to be capable of beating human champions in a matter of hours. As such, machine learning reduces the need for human programmers, and allows AI systems to develop insights beyond human logic; AlphaZero not only beats humans, it also employs “alien strategies” that humans never even considered.
  • The 5G mobile communication standard offers a genuine step change in bandwidth, latency, and the number of simultaneous users. Quantum computing may speed up computation exponentially and allow for types of calculations that are fundamentally impossible with conventional digital computers. While 5G will connect smart devices with a backbone, quantum computing may be at the core of the most demanding intelligent systems.
  • Previous technological revolutions took place first and foremost in the hegemon of the time. Given the AI race between the U.S. and China, it can thus also be interpreted as a fight over global hegemony in the coming decades.

Connecting the Dots

The great technologies that reshape the economy, society, and everyday life tend to come in waves. That is, capital concentrates around a small set of core technologies until these GPTs start to run dry and investments no longer result in the desired returns. Investors then turn to alternative technologies and a new wave emerges. While it would be difficult to argue that IT itself is running dry, i.e. digitization is far from complete, and tech stocks are at an all-time high, there are nevertheless reasons to believe that a new wave is in the making and that AI is at its core.

Digitization is far from complete and tech stocks are at an all-time high, but there are nevertheless reasons to believe that a new wave is in the making and that AI is at its core.

No one would seriously doubt whether AI, together with next-gen mobile communication and possibly quantum computing, has the potential to fulfill all the three GPT-criteria. The question is whether AI is really different from conventional IT, and if it can lead to a new techno-economic paradigm.

As with AI and IT, the GPTs of the past also built on each other. Electricity, to a large extent, was an extension of the steam engine and IT itself is an extension of the electric system. Each, however, brought about a sea of change in productivity. The power of steam engines could only be used close to the engine itself, and it required skilled workers to operate it. Electricity made that power available over a long distance with the flick of a switch, and then IT turned electrons into valuable data. Furthermore, AI has the potential to “upgrade” IT by removing the need for manual programming.

In terms of productivity gains, the latter has at least two implications. Firstly, data can be processed much more efficiently and effectively, and exponentially more value can be generated by doing so (cf. the electrification of factories and its role in mass production). Secondly, AI has the potential to accomplish things that IT simply cannot do because human programmers fall short. Something like an autonomous vehicle, for example, would be impossible to code manually, given the complexity of the task, and hence it would require systems that train themselves to recognize situations and act accordingly. Together with next-gen communication networks and possibly quantum computing, our environment will thus become ever smarter and ever more productive.

If AI indeed drives the next technological revolution, and the Perez’s model is correct, there would be a few implications. Public and private R&D funds will increasingly concentrate around AI (along with 5G and quantum computing), which is already visible, and ultimately lead to an investment frenzy akin to the dotcom bubble of the 1990s. It would also raise the question whether today’s tech giants would really be able to lead the new revolution, which seems very much true today, or yet unknown disruptors will eventually take over from them such as it happened in the past.


  • It is not unlikely that the coming techno-economic paradigm shift will coincide with a hegemonic shift at the geopolitical level. Technological revolutions have always taken place at the techno-economic frontier of the time, and China may very well lead the next revolution and become the new hegemon.
  • We have argued before that we should take care not to overestimate the possibilities of AI in the short run, and in any case, the idea of a superhuman general-purpose AI seems misguided even many years from now. From a societal perspective, it is nevertheless worthwhile to consider what kind of policies will be needed in the long run, and how shorter-term policy making term policies, in response to current AI applications, can prepare for that.