General-Purpose Technologies (GPTs) and General-Purpose Systems (GPSs) are transformative innovations, such as electricity, telephony, the internet, and now AI, that reshape entire economies and societies. Their adoption patterns, often studied through E. M. Rogers’ Diffusion of Innovations theory and the economic perspectives of scholars like Lipsey, Carlaw, and Bekar, involve complex interactions of technological features, infrastructure, market forces, and social behaviour. Understanding these dynamics can help organisations and policymakers foster the rapid and beneficial spread of such powerful technologies.

Mathematical models can capture how an innovation progresses from niche use to the mainstream acceptance that is central to the diffusion of GPTs and GPSs. The logistic growth model, or S-curve, illustrates a three-stage process:

• slow initial adoption

• a period of rapid expansion

• eventual market saturation.

Another foundational principle is the network effect, often described by Metcalfe’s Law. This law suggests that the value of a network rises exponentially as more participants join it.

Yet, value generation goes beyond simple user counts. Economies of scale and the related learning curve phenomenon mean production becomes cheaper per unit as output increases.

Path dependence and increasing returns complicate diffusion further. Once a technology takes root, even if alternative solutions emerge, early success may lock in dominance.

An interesting way to understand how these concepts and GPTs work together is to consider the transformation of society from the end of the steam power era to the computer age. Railroads laid the foundations for corporations and the idea of vertical integration which allows for efficient economies of scale. Countries were urbanised as the railroad systems had the greatest benefit around the terminals and junction hubs on their networks. Electric energy provided lighting and extended the period for productive work. The telephone allowed for instantaneous communication over long distances, facilitating co-ordination of work on a national and international level. The discovery of oil and the invention of the internal combustion engine opened up the possibility for suburbanisation of society, as having access to a car meant workers no longer needed to live at the key locations but could commute to work.

The interplay of these crucial GPTs mutually drove their diffusion. The automotive industry required fuel and roads. The oil industry provided both. Telephony provided the means of communication to build the infrastructure. However, path dependance from the railroad era dictated the size of cars and roads. Where new suburban populations arose, electrical grids were required. Electricity also provided the means to meet increasing demand for products, with the electrification of factories allowing for tools to each have their own electric motor and thus assembly lines could be designed optimally, leading to a 300% increase in productivity over steam power.

Individually, each of these technologies had a slow initial adoption, yet as their mutual benefit became apparent the rapidly expanded and today they are ubiquitous and the structure of society changed with it. We continue to see this cycle with the invention of computers; the internet; email and instant messaging; social media; green energy production; and AI.

Building on these insights, we draw parallels to physical diffusion, using concepts from Fick’s laws to understand how innovations move “through” markets. These analogies highlight that diffusion accelerates when barriers are minimized, “concentration gradients” (or compelling reasons to adopt) are steep, and catalysts, like highly influential early adopters, promote uptake. Accordingly, best practices for designing GPTs and GPSs to optimise diffusion include increasing their relative advantage over incumbents, ensuring compatibility with existing systems, lowering user complexity, and investing in supportive infrastructure or complementary products.

Strategically, organisations can also employ trialability, through pilots and demonstrations, to reduce uncertainty, and observability, through public showcases and testimonials, to clarify a technology’s benefits. Another key consideration is cultivating policy and regulatory support, since favorable standards and consistent legal frameworks often underlie large-scale adoption. An example of this is how telephony received a huge boost after WWI when it’s value to national security was fully realized. Social media and AI are being assessed on this criteria today.

Encouraging open interfaces and interoperability can help avoid lock-in, while continuous iteration and user feedback loops maintain momentum against newer challengers.

In sum, the diffusion of GPTs and GPSs is driven by a blend of economic imperatives, social factors, and mathematical underpinnings. By applying Rogers’ diffusion principles, leveraging lessons from physical diffusion, and considering factors like network effects and path dependence, teams can more effectively steer these transformative innovations toward widespread, sustainable adoption, ultimately fostering growth and sustainable prosperity for every one across industries and regions.