Writing at the tail end of 2024, a prominent crypto investor laid out the technologies that he argued would define the future of the industry in 2025 and beyond.

He suggested cryptocurrencies would either enter the mainstream, integrating with existing financial markets, or whether bitcoin and other digital assets would remain in the shadows. If the technologies failed, “the historical pattern will probably repeat, and we can expect another crypto winter.”

The forecast was interesting because it didn’t assume a binary outcome: crypto dominance or a wipeout. Instead, it drew a roadmap, offering one route by which the crypto cycle might stabilise.

That has proved wise. Rather than settle its place in the mainstream economy, crypto remains caught between recurring bursts of enthusiasm and lingering doubts about broader adoption.

 The upshot is that crypto remains a significant source of economic activity even as the narrative about its transformative power subsides. As one recent comment put it, “adoption has continued to grow steadily even throughout the current bear market, with no signs of a new bull market in sight.”

Not if, but how

If crypto is a solution in search of a problem, as many observers have wryly noted, the same does not apply to AI.

When trying to predict the economic future of AI the key questions are not whether it is useful, but how transformative it will ultimately prove to be, how much productivity it will generate, and whether the economics of today’s AI businesses are sustainable.

Given the complexity of those questions, it is striking that much of the public debate has focused on binary outcomes: will AI investment continue to accelerate for years to come, or is the bubble about to burst?

There are, of course, important nuances around the scale and second-order effects of either scenario. Yet perhaps because AI promises such a profound technological shift, discussions of its future still tend to default to extremes of boom or bust.

Sea change or ebb and flow?

But what if there isn’t a single transformative moment of boom or bust?

What if, instead, the next decade is defined by years of elevated investment in AI infrastructure? Capital continues to flow into data centres and chip development, but without producing a handful of dominant winners. We can imagine a scenario in which competition remains intense, and AI capabilities continue to improve, becoming an increasingly valuable part of the commercial and technological infrastructure of the economy.

Rather than a definable paradigm shift, productivity gains are a series of incremental improvements. Over time, the transformation could be profound, without a single breakthrough that definitively changes everything.

What if AI turns out to be neither a bubble nor a revolution? What if it is simply the next layer of economic infrastructure: expensive to build and transformative over time. In other words, what if AI is normal?