Ultra-wealthy investors are doubling down on artificial intelligence, increasing allocations to private assets to gain early access to innovation, diversify risk and capture long-term growth.
Artificial intelligence (AI) is fast emerging as one of the most compelling long-term investment themes, drawing significant capital from the world’s wealthiest families.
While many portfolios already have meaningful exposure through large-cap US tech stocks, leading families are accelerating their push into AI by directing investment into private equity, venture capital and infrastructure.
Private markets are increasingly viewed as a key route to investing in earlier-stage innovation, and a way to reduce reliance on public tech giants.
According to PWM’s 10th annual Global Asset Tracker survey — which draws insights from chief investment officers (CIOs) and heads of asset allocation at 50 major private banks managing $25tn in client assets — AI and automation is the most popular investment theme, now featured in 80 per cent of client portfolios (see chart).
In addition, more than half of family offices have already allocated capital to generative AI, according to Citi Private Bank’s 2024 Global Family Office Survey. While 53 per cent include this latest technology in their portfolios, a further 26 per cent are considering an investment.
Although public equities were cited by around a third of family offices, 30 per cent preferred private equity funds and 21 per cent opted for private direct investments in their pursuit of AI-related opportunities.
But as adoption deepens, questions remain over whether the sector is entering speculative territory or undergoing a structural shift.
Grace Peters, global head of investment strategy at JP Morgan Private Bank, leans firmly towards the latter. “Our conclusion is that we are not in a bubble,” she says. “The economic benefits of AI, particularly its potential to enhance productivity, could boost GDP growth and help offset labour market pressures, from ageing populations to tighter immigration policies.”
Software holding the edge
As deployment costs fall following breakthroughs such as DeepSeek, AI penetration is set to accelerate across industries. But not all companies will emerge as winners, warns Ms Peters.
The wider ecosystem spans ‘hyperscalers’ — large-scale data centres and cloud service providers — infrastructure and power providers, and application developers.
With uncertainty over where the greatest value will emerge, Ms Peters advocates broad exposure and dynamic allocation across the sector, while highlighting a preference for software companies over hardware manufacturers.
“As AI becomes more affordable and accessible, we expect a wave of new applications and use cases, with software providers likely to benefit most,” she notes.
Despite recent pullbacks among the so-called ‘Magnificent Seven’ — Apple, Microsoft, Nvidia, Amazon, Alphabet, Meta and Tesla — Ms Peters believes large-cap stocks are still likely to outperform the broader market, albeit with a slimmer margin than in recent years.
“These companies remain among the most reliable earnings generators, and over time, strong earnings growth tends to translate into stronger share price performance,” she says. US tech groups are responding actively to competitive pressure. In some cases, such as Alphabet, capital spending has exceeded expectations.
She favours four of the group, particularly those more exposed to software rather than hardware. “These are highly cash-generative businesses. If AI-related capital spending slows, we anticipate more capital to be returned to shareholders through dividends and buybacks.”
The S&P 500 is forecast to carry out up to $1.2tn in share buybacks this year, with more than a third expected to come from the largest seven tech stocks. That, she says, adds to their appeal for investors.
Growing allure of private markets
AI is not just redefining industries but transforming how the world’s wealthiest families allocate capital.
"Some of our largest families now have up to 45 per cent allocated to private markets, not just private equity, but across a diversified mix," says Ms Peters. “Private assets are crucial for meeting return targets in a risk-aware way, helping manage volatility and potential tail risks.” Entrepreneurial clients are particularly drawn to innovation themes, she adds.
Trish Halper, CIO for the Family Office Group at Northern Trust, echoes this sentiment. “We continue to see higher expected growth in technology-enabled opportunities and venture capital over the next decade.”
With their tolerance for illiquidity and longer-term horizons, family offices are well placed to tap into AI-driven trends, investing directly in AI businesses and using the technology internally to boost operational efficiency.
“Family offices tend to be more fully allocated to private markets than core wealth investors and continue to reallocate as investments mature,” notes Ms Halper. Diversification across vintage years, sectors and geographies remains essential to managing cyclicality.
AI could drive long-term productivity gains and impact asset classes worldwide, according to Northern Trust’s analysis. In public markets, while valuations are a “terrible predictor of near-term returns,” the spotlight is on companies deploying AI strategically and at scale. Those are likely to pull further ahead of competitors lacking the resources or expertise to keep up, benefiting from productivity-driven margin boost, says Ms Halper. “We have to stay hyper-focused on fundamentals,” she adds.
Energy and infrastructure
Infrastructure, particularly in the electricity sector, is another major beneficiary due to rising power demands of AI-driven data centres. “Even a simple AI search consumes far more electricity than a traditional internet search,” says Willem Sels, global CIO at HSBC Private Banking. The trend is fuelling demand for both traditional and renewable energy sources.
Cooling data centres is also energy-intensive, prompting fresh investment in mobile nuclear units and clean energy. “The need for diversified energy sources is critical. We anticipate continued reliance on gas, nuclear, and clean energy, along with significant expansion of the electricity grid,” says Mr Sels.
Key beneficiaries are also industrials, buoyed by AI-led innovation and government-backed reindustrialisation efforts.
While the US remains a global leader in AI, investor interest is increasingly international.
Advances by China’s DeepSeek have highlighted the country’s technological progress, contributing to a rotation away from the US tech giants, says Mr Sels.
This shift reflects stretched valuations in the US, changing macroeconomic conditions and growing AI momentum in China and Europe. In the old continent, Germany is investing in digital infrastructure, while France has partnered with the UAE to develop data centres. These initiatives have led some investors to rebalance portfolios in favour of AI beneficiaries across Asia and Europe.
“As costs fall and adoption broadens, the value is shifting from ‘enablers’ to ‘adopters’, companies using AI to unlock tangible growth and efficiency,” says Mr Sels. Private markets are central to this change, with smaller, often overlooked firms using AI to challenge larger rivals. Software companies, he adds, are a particular focus, given their strong presence in the private sphere.
Robots rising
AI is already having a tangible impact across industries. In healthcare, it is revolutionising diagnostics, pharmaceutical development and hospital operations. “AI accelerates drug discovery and enhances diagnostic accuracy,” says Mr Sels. “In many cases, AI works alongside doctors to boost outcomes — we call them ‘co-bots’.”
Automation is another frontier. AI-powered robotics are thriving in tight labour markets and high-wage environments, particularly as firms look to reshore production to the US or Europe, he explains. These systems are making independent decisions and improving efficiency.
“As working age populations shrink, competition for skilled talent will intensify, spurring investment in automation and productivity-enhancing technologies,” says Jean Chia, global CIO of Bank of Singapore, echoing the optimism at HSBC. The next wave, she predicts, will include AI-robots and ‘humanoids’ increasingly active in daily life.
Private equity and venture capital firms are also turning to AI to sharpen operational performance, says Anastasia Amoroso, chief investment strategist at fintech firm iCapital. Beyond improving due diligence, AI is being deployed within portfolio companies to streamline processes, a key lever for value creation.
“AI and machine learning are still in early stages of adoption. Despite the hype, very few companies have adopted them at scale,” she notes. “There is a massive opportunity for buyout managers to embed AI into portfolio companies and unlock earnings growth.”
With higher interest rates limiting the scope for financial engineering and multiple expansion, she expects returns to depend more on operational gains. Over the next decade, iCapital forecasts average annual returns of 9.7 per cent, roughly 300 basis points above US large-cap equities.
Private markets remain key to financing US tech innovation. While public markets are dominated by a handful of large firms, private assets, she says, offer scope for higher returns and lower volatility. “Private markets invest more in innovative sectors, there’s simply a higher share of growth companies,” says Ms Amoroso. “Also, these businesses tend to grow earnings at a faster clip, which is crucial for driving long-term returns.”
Navigating risks
The AI opportunity is substantial, but so are the risks. Manuel Villegas, next generation research lead at Bank Julius Baer, warns of geopolitical tensions, monetisation challenges and regulatory hurdles.
“Everyone wants the highest-performing chips, but sharing technology doesn’t come for free,” he notes. US export restrictions on high-performance computing chips have disrupted global supply chains, not just in China, but worldwide.
Valuations pressures are also a concern. “While the market expects major efficiency gains, some AI model providers may struggle to monetise their products in an increasingly competitive and commoditised market,” he says.
Moreover, incumbents are working to protect their attractive margins, while challengers are eager to disrupt them.
Regulation, adds Mr Villegas, is likely to intensify. As the power of AI models grows, so does the need for oversight, especially in a world where vast amounts of public information can be aggregated rapidly and efficiently, lowering traditional barriers to entry.
“For long-term investors,” concludes Northern Trust’s Ms Halper, “the priority is separating hype from fundamentals and backing businesses embedding AI in ways that drive real-world outcomes.”
