“We view AI as the new digital.” This phrase from the foreword to Accenture’s Technology Vision 2025 report immediately caught my attention. From the perspective of a global IT consultancy, it’s probably true. From an even broader perspective, it may not be big enough. What’s certainly true is that CEOs are replacing ‘digital’ with ‘AI’. But is this more than just a change of buzzword?

“Digital transformation” is shorthand for the implementation of digital technology and the resulting transformation of the business. It is an ongoing process that is often referred to as “the fourth industrial revolution” or “industry 4.0”. As a digital technology, AI has already been part of this transformation, in the form of expert systems, machine learning, and deep learning.

With the move to generative AI, or perhaps general intelligence, we’re witnessing a paradigm shift. AI is now what economists call a general-purpose technology:

There aren’t that many general-purpose technologies in the history of tech. The computer is one good example, the internet is another. For earlier generations, think electricity, or the steam engine. What all these have in common is that they impact almost all parts of our collective and individual lives: our societies, economies, and the way we work, shop, live, or entertain.

In general, it takes time to apply general-purpose technologies to businesses and bring them to consumers. Electricity, which has been around for more than a century, is just now entering into a new growth phase to electrify everything, as the catchphrase goes. The energy revolution is similar to the digital revolution to the extent that it replaces atoms (and molecules) with electrons, which are easier to move, and cheaper to produce from renewable sources than from fossil fuels.

Just another general-purpose technology?

For all these revolutions, huge investments are needed:

With previous technologies, such as electricity and IT, growth rates were below those attained in the decades immediately preceding their arrival.¹⁹ As Brynolffson et al. explain, substantial complementary investments were required to realize the benefits of general-purpose technologies, where productivity emerges after a significant lag.⁴ With electricity, for example, it took decades for society to realize its benefits, since motors needed to be replaced, factories needed redesign, and workforces needed to be re-skilled. IT was similar, as was the Internet.

But let’s get back to AI. Is it just another general-purpose technology fueling digital transformation? In a way, it is, just like computers and the internet.

AI is similarly in its early stages, where businesses and governments are scrambling to reorganize processes and rethinking the future of work. Just as electricity required the creation of an electric grid and the redesign of factories, AI will similarly create an “intelligence grid” requiring a redesign of processes to realize productivity gains from this new general-purpose technology.⁵ Such improvements can take time to play out.

The question is: Will AI continue the trajectory of the digital revolution, or will it lead to a new technological paradigm? In a recent paper on that question, the authors state that

AI is fundamentally different from ICT technologies in that its core capability – making predictions – differs from those of the computer – computing – and the internet – i.e. managing information. Obviously enough, these distinctive features of AI technologies are even more evident after the arrival of the generative AI algorithms, such as large language models (LLMs).

The rise of a new paradigm

The Tech Vision report sees the rise of a new technology paradigm defined by abundance, abstraction, and autonomy:

• Abundance: the creation of digital systems is getting a lot cheaper and faster.
• Abstraction: agentic or autonomous systems can unbundle the function/data/interface paradigm of software.
• Autonomy: systems can build and execute code on their own.

This goes beyond features and capabilities.

Abundance means we’ll get even more digital systems, which, on its own, would lead to a new wave of digital transformation.

Abstraction won’t just give us new, natural-language interfaces – it will fundamentally restructure the inner workings of computing, the bundling of functions and data into applications with graphical user interfaces. We’ll need new interfaces for agents and autonomous systems that work both ways. What will the world look like beyond today’s app paradigm?

Finally, autonomy brings systems that make decisions beyond what they are programmed to do. They can programme themselves – or other systems. This will add a new level of recursive and reflexive structures to our IT landscape, to put it mildly.

How big can it get?

A new round of IT growth is likely to emerge as the new paradigm unfolds. Given that the Big Tech oligarchy that is investing heavily in AI is already made up of the biggest empires in history, how big can it get? Regarding the market cap, this amounts to the question of whether we’re in a bubble or not. That’s a question that is notoriously hard to answer before the bubble finally bursts. With Nvidia, AI has already brought a new player into the Big Tech league.

DeepSeek, at least, is “demonstrating that others can also leapfrog the big boys with clever thinking”, as Om Malik has put it. So, it’s not game over yet. And OpenAI (which is not part of Big Tech, at least not yet, although it’s affiliated with Microsoft) isn’t tiring of pointing out that AI is the fastest deflating technology in history:

The cost to use a given level of AI falls about 10x every 12 months, and lower prices lead to much more use. You can see this in the token cost from GPT-4 in early 2023 to GPT-4o in mid-2024, where the price per token dropped about 150x in that time period. Moore’s law changed the world at 2x every 18 months; this is unbelievably stronger.

So yes, it can be big. But it’s not a foregone conclusion that today’s Big Tech boys will control and oligopolise AI as they did with the internet. Despite the rush to build AI and the heavy investment in hardware, there doesn’t seem to be a moat. At least not yet. This may change over time, but:

The potential size of this market is hard to grasp — somewhere between all software and all human endeavors — so we expect many, many players and healthy competition at all levels of the stack. We also expect both horizontal and vertical companies to succeed, with the best approach dictated by end-markets and end-users.

Which brings us back to one of our evergreen pieces:

The internet and its technology have done many things to many industries and continue to do so. But what really strikes me is the observation of both vertical and horizontal integration at the same time and in the same industries. This dual strategy and its success is probably one of the reasons why Big Tech has become extremely dominant over the past decade.

Big Tech is likely to try this approach again under the new paradigm of AI. Whether they succeed or not will have huge consequences for our future. But that’s another story.

Image by Solen Feyissa / Unsplash.