Could private markets be the key to selling shovels in the AI gold rush?

It has been just over 175 years since gold was first unearthed in Sutter’s Mill.

At the foothills of the Sierra Nevada, this water-powered sawmill was the site of a discovery that triggered the infamous California gold rush, and whose lessons remain prescient still to this day.

Once news of gold spread in 1848, prospectors flocked to makeshift mining towns to pan for gold in the hope of returning with a fortune, and for the most part, their ambition was richly rewarded. The earliest gold-seekers were able to collect large amounts of easily accessible gold, earning 10 to 15 times their daily wage and often retiring after just six months in the region.

Yet this exuberance was short-lived, and in the years following its first discovery, nearly 300,000 prospectors from as far as Australia soon flocked to the region. Mines were quickly industrialised, and those who arrived in the latter stages of the rush were taxed heavily and unable to earn a living. Stories soon circulated of those who had amassed great fortunes in the gold rush, not by engaging in the mining euphoria, but by selling the equipment needed to participate. The lesson of the gold rush was simple: “Don’t dig for gold, sell the shovels.”

As we enter what appears to be a digital gold rush, this wisdom could be foremost in the minds of any would-be prospector.

Much like the late arrivals to Sutter’s Mill, investors are coming to terms with the fact that the perceived easy money has likely already been made from AI applications.

Early investors in the sector have been rewarded over the last few years, with valuations reaching record levels in some cases, and a handful of startups rocketing towards unicorn status thanks to the letters ‘AI’ being added to their names. Most strikingly, private US startup Unconventional AI burst onto the scene promising to enhance the energy efficiency of machine learning systems, raising $475 million in a seed round and reaching a $4.5 billion valuation within two months of launch, a far from typical result.

Yet, cracks have started to show, and those on the wrong side of the previous software trade have found themselves experiencing greater levels of price volatility. For instance, Microsoft’s share price has retreated around a third from its peak in late 2025, despite being viewed as an early beneficiary of AI adoption, possibly due to questions mounting about the extent to which AI integration will drive future earnings growth.

In Europe, private equity firms like EQT and Brookfield are already engaging in multibillion-euro sales of their data centre holdings, capitalising on the growing demand for this critical infrastructure.

Similarly, alternative asset manager DigitalBridge is quietly amassing a $115 billion data centre and digital infrastructure portfolio, situating itself as one of the largest global investor-operators in the sector.

Many of these deals are characterised by long-term commitments with high-quality counterparties, such as Amazon, acting as exposure to the forefront of AI development, away from the software layer

By taking an integrative approach, private markets could help bridge this gap between AI-related demand and the insufficient existing energy supply, with single firms developing dedicated infrastructure funds that could capture value across all of the sector bottlenecks.

One of the largest is Brookfield Asset Management, which has launched a $100 billion global AI infrastructure programme, securing partnerships with energy providers to deploy data centres with dedicated on-site power facilities.

Similarly, a consortium of alternative investment managers including BlackRock’s Global Infrastructure Partners, completed a $33.4 billion acquisition of AES, a global utilities provider operating facilities close to a number of data centre hotspots, marking the largest PE-backed deal in the power generation sector.

This joins a steadily growing list of private markets firms announcing acquisitions of US power companies, with five of the seven largest acquisitions in the industry being announced in the last three years.

For those keen to heed the lessons of the 19th century shovel sellers, the advent of artificial intelligence has brought with it a similar case of a market characterised by fortune or folly.

In these conditions, it may be the case that a prudent source of returns could be found in the areas most instrumental in laying the physical foundations for the artificial intelligence age. Private markets could provide an alternative route for sophisticated individual investors to move beyond the speculation over which company has built the most adept chatbots. Through a combination of equity, infrastructure, and credit, these markets could offer alternative exposure to more fundamental elements of the forthcoming period of industrial capital deployment.

Historically, these private market vehicles would have been off-limits to anyone outside of the ultra-wealthy, sovereign wealth funds or institutions, prohibiting individual investors from going any further than the exposure offered by public equity markets. However, the emergence of semi-liquid funds has provided a new vehicle for experienced individual investors, offering an alternative potential way to benefit from the AI tailwinds.

Funds such as these, from some of the largest Private Markets managers, can be accessed with a minimum £10,000 investment through Wealth Club’s Private Markets Platform. This allows eligible investors to reach funds across Private Equity, Private Credit, Secondaries, and Infrastructure, with each fund reviewed by our in-house research team.

It must be noted, of course, that identifying the shovels to sell is only half of the battle, as this strategy relies on the continued solvency and demand from those wishing to buy gold mining equipment. In the case of AI, if that demand proves less durable than expected, the performance of these investments would be materially affected.

Furthermore, while there may be a strategic case for using private markets to capture the wider artificial intelligence transformation, they come with considerable risks of their own. Private market investments are long‑term and illiquid, and returns may take many years to materialise, if they do at all. They are not easily realisable, so eligible investors should only use money that is not needed for at least five to ten years. As with all private market investments, there is a real risk that investors could lose some or all of the capital committed.

They are much less transparent and have heightened levels of pricing uncertainty compared to publicly traded alternatives.