AI boom stirs fresh market risk concerns

An accelerating wave of AI-related investment is reshaping global markets, with large technology platforms committing very high levels of capital expenditure to data centers and computing infrastructure. Combined spending by major hyperscalers is estimated to reach about $725 billion this year, reflecting intense competition to build AI capacity. This capex is flowing into chips, memory, power and networking, helping to drive sharp moves in parts of the technology sector.

The scale and speed of this buildout have turned AI into a dominant market theme. Strategists note that AI now influences both equity valuations and macro expectations, as investors assess how the investment boom could feed through to growth, inflation and financing costs.

Risk specialists are warning that AI is also enabling new forms of financial crime and operational vulnerability. At a recent banking and insurance summit, a consultant highlighted that deepfake scams have risen dramatically over the past three years, describing an increase of more than 2,000%. The spread of highly realistic synthetic media raises the risk of fraud in customer onboarding, payments and remote authentication.

The same speaker pointed to adversarial AI techniques such as prompt-injection attacks in documents, where hidden instructions are embedded in financial reports or forms to manipulate AI-based processing tools. As more institutions rely on AI assistants and automated document workflows, these techniques could compromise internal controls, data integrity and decision-making processes.

Banks and insurers are increasingly deploying AI agents to scan market signals, rebalance portfolios and hedge risk exposures. These systems can act at machine speed and are often trained on similar datasets or external feeds, potentially leading to clusters of similar behavior across institutions. The concern is that this could concentrate risk and amplify market moves if many models respond in the same way to new information.

Experts also highlight an accountability gap when AI systems take decisions with limited human intervention. If automated agents accumulate or shed exposures before traditional risk and governance frameworks react, losses or liquidity strains could emerge faster than legacy controls can detect. This raises questions about how responsibility is assigned when AI-driven strategies misfire.

Market strategists are tracking a sharp divergence between companies building AI hardware and memory and the large technology groups financing much of the spending. The Philadelphia Semiconductor Index has risen about 87% in 2026, reflecting enthusiasm for chipmakers that supply AI processors. A memory-focused exchange-traded fund launched in April has climbed roughly 141% since inception, highlighting strong demand for memory suppliers tied to AI workloads.

By contrast, the large technology companies committing the biggest sums to AI infrastructure have not matched those gains. This split between equipment suppliers and heavy spenders has been compared to late-1990s market dynamics, with some analysts warning that such a pattern can leave sentiment vulnerable to reversals if expectations change.

Views on how to position for the AI wave remain divided. Some high-profile equity strategists continue to back the theme, recommending exposure to AI infrastructure and power-related plays as well as to the major platforms funding the buildout. They argue that, despite volatility risks, the structural demand for AI capacity and energy remains strong.

Policymakers, meanwhile, are monitoring the AI-driven capex shock closely. A central bank chair recently said the investment surge has "huge implications" for interest rates and noted that its effects are appearing first on the demand side of the economy. With upcoming policy meetings in focus, officials are assessing how AI-related spending might affect growth, inflation and financial stability, including potential feedback loops through equity and credit markets.