@Pixabay.
Artificial Intelligence (AI) could be said to have come of age in 2025 with the complete acceptance of its place in the investment ecosystem.
However, continued questions about its efficacy, environmental impact, and regulatory boundaries persist, which have caused some to question whether it is necessary.
But what is obvious is that it is now entrenched in the insurance investment landscape. Investment portfolios now use AI for a variety of analysis, operational, and allocation preliminary work.
“As cloud data centres demand more electricity, fibre, and cooling, the grid and construction sectors could benefit."
There is also heavy interest in the sector from insurers. For instance, last week, managing general underwriter Advanced Technology Assurance Limited (ATA) launched a $750 million insurance facility to underwrite the global AI infrastructure build-out, which is a market it said is forecast to attract nearly $7 trillion in investment by 2030.
This is part of a huge influx of capital to the market. “The digital boom complements the physical one,” Andrew Ye, Investment Strategist at Global X ETFs, on where the cash is heading.
“As cloud data centres demand more electricity, fibre, and cooling, the grid and construction sectors could benefit. Investment in both physical and digital infrastructure has become strategic. Cement and silicon, once separate worlds, now share a purpose: securing resilience in an uncertain age.”
The investment buildout has, for the most part, been concentrated in the hyperscalers (Amazon AWS, Microsoft Azure, etc.): those global data centre operators that serve companies needing vast computing power, data storage and infrastructure.
Consensus estimates that this group will spend $450 billion in capital expenditures (capex) for 2025 and $600 billion for 2026, which would account for more than 10% of all capex in the US economy, according to David Byrne, Senior Equity Fund Manager at Setanta and Lenny McLoughlin, Chief Investment Strategist at ILIM in Keyridge Asset Management’s 2026 Outlook.
“A key consideration is what return these companies are getting from their spend,” they said. “Meta revealed in its latest earnings call that its Reels product, powered by AI, has an annual revenue run rate of over $50 billion.”
Alphabet is seeing similar success in Google and YouTube, making it clear that both companies are generating immediate returns from their AI investments. Another area of AI success has been in computer coding, which is seeing efficiency improvements as large as 30% in coding time.
“We can think of many applications, from mortgage applications to truck and trailer routing, that are likely to see AI integration in the very near future,” they said.
AI’s role in infrastructure means it continues to stand out as a resilient and dynamic asset class, with a continuous strong outlook for 2026, according to Gilles Lengaigne, Managing Partner, Head of Origination and Corporate Development, at Infranity.
“Institutional investors are increasingly drawn to infrastructure for its distinctive characteristics: stable, inflation-linked cash flows, low correlation to public markets, and downside protection,” said Lengaigne.
“This innovation cycle is extremely capital-intensive, unlike the platform-based, capital-light tech revolutions of the past."
“Strong data growth, cloud computing, and artificial intelligence are driving exponential demand for data centres, fibre networks, and 5G connectivity,” he said. “Sustainable digital infrastructure is now as critical as transport or power grids once were.”
The spill-out effects of AI are also being considered for investment opportunities.
“This innovation cycle is extremely capital-intensive, unlike the platform-based, capital-light tech revolutions of the past,” said Giordano Lombardo, CEO and Co-CIO, Plenisfer Investments, in the Generali 2026 Outlook. “Hundreds of billions are being invested each year in hardware, chips, data centres, and increasingly, energy. For that reason, access to low-cost electricity is becoming an overlooked but critical competitive advantage.”
Lombardo said rather than chasing the most crowded megacap names, which are companies with market caps above $200 billion, he said he prefers the enablers: “semiconductor manufacturers, chip testers, and the broader supply chain that typically benefits from AI capex without the same valuation risk”.
The epicentre of the AI revolution for investment opportunities is the US. However, various indicators around the US economy show different stories for where it is heading in 2026 and what AI can offer it.
Whilst the majority still believe the US economy is in relatively good shape, recurring areas of weakness are mentioned, and many are talked about in relation to AI.
Some of those include the over-reliance on the so-called Magnificent Seven and other tech –primarily AI companies – on Wall Street.
“Overreliance on this single driver of growth is our first concern for the US economy in 2026 should heroic levels of AI-related investment not be sustained."
“A key component of the US economy’s resilience in 2025 has been investment connected with the AI build-out,” said Impax Asset Management in its 2026 outlook. “‘Hyperscale’ data centre operators were expected to invest $342 billion in capital expenditure (capex) in 2025, up 62% year-on-year. Based on estimates by JP Morgan, AI-related capital expenditures contributed 1.1 percentage points of US GDP growth in the first half of 2025 – more than consumer spending.”
Impax said that “Overreliance on this single driver of growth is our first concern for the US economy in 2026 should heroic levels of AI-related investment not be sustained”.
They added that the recent shift in AI build-out financing, from balance sheets to debt, is significant and that hyperscalers have issued more than $100 billion in bonds in 2025, up from $20 billion in 2024.
“While not an immediate concern, it brings into sharper focus questions around the financial sustainability of the AI-related investment cycle.”
Byrne and McLoughlin said that, historically, substantial productivity improvements from new technologies have only become apparent once adoption rates exceed 50%.
"Productivity gains of 0.2-0.3% per annum could be achieved, with improvements of 0.8-1.3% possible by the 2030s, providing a foundation for growth.”
“Current reports indicate that the US is approaching this threshold, with an AI adoption rate of over 45%, compared to just 25% in Europe,” they said. “However, in many cases, adoption is not yet fully scaled. If these trends continue, productivity gains of 0.2-0.3% per annum could be achieved by the end of the decade, with improvements of 0.8-1.3% possible by the 2030s, providing a strong foundation for future growth.”
While AI does present risks to employment – particularly in certain roles – evidence of widespread job losses remains limited.
There has been some increase in unemployment among younger workers in the technology sector, but little impact has been observed in other sectors considered most vulnerable to AI.
Research suggests that ultimately 6-7% of workers could be displaced by AI, though estimates vary widely. It is important to note that the introduction of new technologies often creates new types of jobs; for example, 60% of today’s workforce is employed in roles that did not exist in 1940.
There could, in the short-to-mid-term, be some backlash against AI.
This has been highlighted, primarily, by the environmental concerns, namely the huge water usage of data centres.
There is also the social aspect. In mid-January, the UK banned deepfakes, and X owner Elon Musk entered into a war of words with the British government over free speech around the use of sexualised images on X’s AI chatbot Grok.
Several other countries are blocking certain aspects of AI and investigating ways to limit access. This could then bleed into both the companies that develop AI and their attractiveness to investors, as well as corporate uptake of their products.
