There is a story that gets told whenever excitement builds around a new frontier. It concerns the great Californian gold rush of 1849 — a frenzied, speculative hunt for wealth beneath the earth that drew tens of thousands of fortune-seekers from across America and beyond.
According to Corlytics, travelling conditions were brutal, disease spread quickly, and the gold in the rivers was panned out fast. What followed was hard digging, grinding uncertainty, and, for most, disappointment. Yet some people prospered enormously. Not the miners. The ones who sold them the shovels.
Corlytics recently discussed the topic of the AI rush, and who were the shovel sellers in this modern gold rush.
The parallel with artificial intelligence is difficult to ignore. AI is the new rush — and, fittingly, its epicentre is California, the very same state where prospectors once streamed into the Sierra Nevada foothills. Today’s headlines are equally intoxicating, encouraging investors and entrepreneurs alike to dream big. But the question worth asking is not whether AI will transform industries. It is the same question that applied in 1849: who actually makes the money?
The companies commanding the most extraordinary valuations in the current AI boom are not, for the most part, the firms that have discovered what AI can do for end customers. They are the infrastructure providers — the shovel sellers of this era. Nvidia is the most striking example. At its peak, its market capitalisation exceeded that of any other company in recorded history. Yet Nvidia does not produce consumer AI products. It manufactures graphics processing units — the specialised chips that AI models require to train and operate. In the most literal sense, it is a shovel company. Its customers are the forty-niners.
Beyond chips, the hyperscale cloud providers — Amazon, Google and Microsoft — are spending heavily on vast data centre infrastructure, which they rent to AI firms to carry out their computations. They function much as the railways, warehouses and supply chains of the original rush did: indispensable to the prospectors, profitable regardless of whether gold was found. Then there is energy. AI models consume extraordinary quantities of electricity, and without power, nothing runs. Energy providers are quietly becoming some of the most consequential beneficiaries of the AI moment.
The historical lessons from mineral exploration are instructive here. In gold fields and oil regions, vast sums are spent speculatively based on geological assumptions and early discoveries. Whether or not significant deposits exist, the companies supplying the drill rigs, the stakes, and the equipment are paid before exploration even begins. AI companies are currently consuming capital at comparable rates — based almost entirely on potential rather than revenue or profit.
This dynamic helps explain the rising interest in what analysts have begun calling HALO stocks — Heavy Assets, Low Obsolescence. These are businesses with significant physical infrastructure that are considered less susceptible to disruption by AI and may, in fact, grow more valuable because of it. The chip, the cloud cable, and the current are the modern equivalents of the shovel, the pick, and the dry goods store.
Yet infrastructure alone is not the full picture. Just as prospectors historically fared better with geological surveys, expert guidance, and reliable maps, the AI sector increasingly requires sophisticated regulatory intelligence. Firms such as Corlytics are filling this role, analysing regulatory behaviour across jurisdictions and helping organisations navigate the evolving landscape — including the EU AI Act, ISO/IEC 42001, and further frameworks in development. In the AI boom, infrastructure powers the search. Intelligence guides it.
History offers a clear verdict on the original gold rush. Thousands of forty-niners died or returned home having earned less than if they had never set out. Those who prospered were the ones supplying the gold pans, the shovels, the hard-wearing denim trousers — most famously those made by Levi Strauss — and the financial services, such as Wells Fargo and American Express, that the prospectors could not do without. Whether or not the miners found gold, they needed these things. The outcome of the rush was quite different from what most miners had imagined when they departed.
The Californian gold rush did ultimately build something lasting: great cities, major infrastructure, and a culture of speculation that endures to this day. The wealth it generated was real, but it derived not from the gold itself, so much as from everything required to extract it. The AI rush may yet follow the same pattern. Those who focus on the tools — rather than only on the pursuit — may find that this era offers something more durable than a seam that runs out.
Read the full Corlytics post here.
Copyright © 2026 RegTech Analyst
Copyright © 2018 RegTech Analyst
