While in 2025 the capital expenditure of tech giants (hyperscalers1) was the dream of investors, this year it is causing concern. The optimism generated by the productivity gains expected from the implementation of artificial intelligence (AI) in many sectors of the economy has given way to fears of disruption in several industries.
The concerns began in the fourth quarter of 2025, initially targeting the tech giants specifically. For many years, these companies were able to rely on significant quasi-monopolistic situations in their core businesses, giving them substantial margins and highly predictable profit growth. All this with virtually no debt and relatively few tangible assets, which favored their agility and capacity for innovation. In recent months, however, these giants have undergone a sudden transformation, deploying significant capital in the race for AI and making their future cash flow generation less linear. While this financing was initially mainly carried out using their cash flow, several tech giants have started to borrow more heavily (such as Google and its recent issues totaling $32 billion over several maturities ranging from 5 to 40 years, including a £1 billion issue with a 100-year maturity; Meta with USD 30 billion raised at the end of last year, and Oracle with USD 25 billion raised at the beginning of the year). Since the beginning of the year, AI-related issues on the US Investment grade2 (IG) market have accounted for 40% of all newly issued bonds, compared with 10% in 2023. Added to this is a worrying form of circular economy, where suppliers are financed by their own customers, as well as a decline in cash levels among certain players. Furthermore, these investments in data centers have raised questions about the future profitability of this physical infrastructure, as well as potential compensation by AI players for the rise in energy costs for American households because of their energy-intensive nature.
During this period of great uncertainty about the real monetization potential of all these AI-related expenditures, investors have repositioned themselves in more traditional, cyclical, and lower-valued sectors such as materials, energy, consumer staples, telecommunications, industrials, and small and mid-cap companies. The latter have outperformed the rest of the market since the beginning of the year and, above all, offer better visibility of future earnings.
This trend has accelerated since the beginning of 2026, initially with significant discrimination within the AI value chain, with investments being reallocated upstream of the hyperscalers' value chain to semiconductor manufacturers, who are the direct beneficiaries of this capital expenditure, rather than downstream companies such as software developers. The latter have been shaken by fears of disruption to their business model. This "trade of the month" by hedge funds boosted the performance of semiconductors, with gains of nearly 15% since the beginning of the year and led to a 22% drop in software over the same period.
As a result, almost no software stocks escaped the sector-wide sell-off. As of 16 February, the sector was down nearly 30% from its 52-week high in the United States. Nearly 80% of software companies suffered declines of more than 30%, a phenomenon rarely seen outside of a bear market.
These fears of disruption then spread beyond the technology sphere. Almost every day, a new sector saw its business model challenged by AI. The first to be targeted was insurance broking. Investors began to anticipate future AI capabilities that would give anyone full transparency on premiums and coverage across all insurance policies, potentially driving margins to the floor and disintermediating brokers altogether.
Next it was the turn of real estate, with some stocks in the sector falling by nearly 20%. The investors fear in particular that the automation of complex transactions and lease management will render the traditional role of brokers obsolete, while also fearing a reduction in demand for office space if AI replaces administrative functions on a large scale.
The latest sector to face fears of disruption is logistics. The rise of algorithms capable of optimizing logistics flows without human intervention threatens the margins of freight forwarders and supply chain managers.
Finally, the banking sector has not been spared by this wave of panic. Concerns are centered on retail banking and investment advice, where ultra-personalized generative AI models could offer real-time wealth management and credit services, bypassing physical branch networks. This pressure on intermediation revenues, combined with greater transparency on pricing, is fueling fears of a structural threat to the long-term profitability of traditional banking institutions.
A common thread runs through all these sectors : their business models rely on monetizing an intermediary position. Investors' fears are not so much about the arrival of a technologically superior competitors about AI's ability to give end clients direct access to the service they want, without the need for a paid intermediary such as a broker.
Hopes for productivity gains have therefore given way to fears of disruption by AI. The declines suffered by some players are certainly premature or exaggerated. That said, AI will revolutionize many sectors in the coming years, leading to significantly increase production and a considerable reduction in costs. Some even argue that this could lead to a return to deflation and a wave of significant job destruction. However, such fears should be put into perspective. Historically major technological advances have certainly eliminated certain functions, but these have also boosted productivity and created new types of jobs. The internet, for instance, largely wiped out traditional travel agencies, yet it led to a surge in travel bookings via online platforms, which in turn supported growth in air traffic and employment at airports and airlines.
The significance of market sales was reflected in the fact that nearly 25% of S&P 500 stocks fell by more than 7% between January 1 and mid-February. However, all this has so far only led to a very moderate fall in the leading indices. Despite these events, US equity indices are trading very close to their all-time highs, even though all the "Magnificent Seven" have underperformed the S&P 500 since the beginning of the year.
This sell-off in parts of the technology universe and a handful of specific sectors has not spread to the broader market, and investors remain firmly in a "risk-on" mode. Institutional portfolio liquidity levels are at historic lows, and most equity fund managers are almost fully invested. Private investor flows into US equities have reached record levels in recent weeks, with nearly $50 billion deployed over the last three weeks, nearly 40% more than previous purchase records. Flows into international equity ETFs have also hit a record, exceeding USD 51 billion and marking a 17th consecutive month of net inflows. This rotation into more traditional sectors, rather than into cash or "defensive" assets such as gold is a sign of confidence in the equity market and should even help to extend the rally in equity markets in 2026 on a broader and therefore more resilient basis.
This redirection of investments has also benefited emerging market equities, which have significantly outperformed global equities by more than 10% since the beginning of the year. The combination of expected rate cuts and a weak dollar should allow them to continue this momentum, even if the argument of undervaluation is less decisive, requiring new catalysts.
The bond market reacted by demanding a relatively substantial risk premium on high-quality companies in the technology sector, up to 0.3% above the average for IG2 (compared with historically lower-than-average yields), but has so far had very little impact on other market sectors.
Even if AI-related productivity gains are likely to support higher margins before causing large scale disruption, it will be important to remain vigilant in the medium term to ensure that certain sectors of the economy do not suffer massive layoffs as AI is adopted. Admittedly, a rise in the unemployment rate, which in turn would fuel a decline in inflation, would be conducive to convincing even the most hawkish3 that a rate cut is necessary. Let us not forget that two-thirds of US GDP comes from consumption, which itself depends on employment. A cooling of the labor market, which is not yet visible, should nevertheless be tempered by a smaller immigrant workforce, the key role of investment in the current growth cycle, and the wealth effect generated by financial markets.
And history should not be forgotten: the automobile ruined horse breeders and stagecoach operators, but created jobs for auto workers and engineers, gas station attendants and mechanics. AI is likely to generate the same kind of creative destruction, but ultimately enables a leap forward in productivity, economic growth and corporate profitability, albeit in a more volatile environment.
1 A hyperscaler is an IT service provider that operates very large-scale data centers to offer cloud computing and data management services. These are cloud giants such as Amazon, Google, Microsoft, Alibaba, etc.
2 "Investment Grade" refers to the category of bonds with high credit quality and low default risk.
3 "Hawkish" refers to those who advocate raising interest rates or reducing the money supply in order to slow down the economy and combat inflation.
DISCLAIMER
This document is issued by the Edmond de Rothschild Group. It has no contractual value and is intended for informational purposes only.
This document may not be distributed to persons in jurisdictions where it would constitute a recommendation, an offer of products or services, or a solicitation, and where its distribution could therefore contravene applicable legal and regulatory provisions. This document has not been reviewed or approved by any regulatory authority in any jurisdiction.
The figures, comments, opinions, and/or analyses contained in this document reflect the Edmond de Rothschild Group's views on market developments based on its expertise, economic analyses, and information available to it at the time of preparation of this document, and are subject to change at any time without notice. They may no longer be accurate or relevant at the time of reading, particularly in view of the date of preparation of this document or due to market developments.
This document is intended solely to provide general and preliminary information to those who consult it and should not be used as a basis for any investment, divestment or holding decision. Under no circumstances can the Edmond de Rothschild Group be held liable for any investment, divestment or holding decision made on the basis of these comments and analyses.
The Edmond de Rothschild Group therefore recommends that each investor obtain the various regulatory descriptions of each financial product before investing, in order to analyze the associated risks and form their own opinion independently of the Edmond de Rothschild Group. It is recommended that you seek independent advice from specialized professionals before entering into any transaction based on the information contained in this document, in order to ensure that the investment is appropriate for your financial and tax situation.
Past performance and volatility are not indicative of future performance and volatility, are not constant over time, and may be independently affected by changes in exchange rates.
Source of information: unless otherwise stated, the sources used in this document are those of the Edmond de Rothschild Group.
This document and its contents may not be reproduced or used in whole or in part without the permission of the Edmond de Rothschild Group.
Copyright © Edmond de Rothschild Group – All rights reserved