Technology sector: bubble or bull?

The alarm bells have started ringing, although not necessarily from the institution one might have expected. In a note published by its economics team, the ECB argues that the stock-market trajectory of the Magnificent Seven is a source of concern for euro-area financial stability, given investment funds’ exposure to the group and historical correlations. The minutes of the Fed’s latest meeting also show that its governors are keeping a close eye on the technology sector, notably because of the inflationary effects generated by AI-related spending. Unsurprisingly, the sector’s weight at both the micro- and macroeconomic levels makes it a central issue for investors, and with developments moving so quickly, an update to our framework seems necessary.

The first update concerns the amounts being spent by hyperscalers in the race to deploy AI infrastructure. The earnings season that has just ended once again gave US mega-caps an opportunity to revise their capex forecasts, from $757 billion to $793 billion according to the consensus estimate for this year. Yet with spending forecasts being revised higher time and again, investor attitudes appear to have shifted, as illustrated by Alphabet, Google’s parent company. Since announcing its first negative quarterly free cash flow since going public, the stock has underperformed the S&P 500 by 5%, highlighting how central the monetisation of these investments has become at a time when technology companies are turning to capital markets on a massive scale to finance their spending.

Indeed, it was Alphabet that brought the summer lull in credit markets to an end by raising $115 billion through a new bond issue. Intel, for its part, tapped the equity market, raising just over $20 billion — a first since its 1971 IPO. Private markets are also being called upon, with Nvidia aiming to raise $500 billion from some of the biggest names on Wall Street* to help its customers finance their investments, particularly in processors.

But this avalanche of paper** comes at a price: investors are demanding greater compensation through wider credit spreads, as shown by Alphabet’s latest transaction at the end of August. The company issued an Australian-dollar-denominated bond, including a 20-year tranche priced at a 180-basis-point premium over the government bond yield, for an overall yield of close to 7%. And the spending trend is unlikely to reverse in the near term, not least because of the rivalry with China.

Performance analysis by the benchmarking website Artificial Analysis of Moonshot’s Kimi K3 model came as a cold shower in the United States. The model ranks just behind the leading proprietary models developed by US players in terms of intelligence. One of the distinguishing features of this Chinese model is that it is open*** and uses a “mixture of experts” architecture****. This allows it to combine enormous capacity with a controlled inference cost.

Beyond the more or less surprised reactions of America’s AI gurus, the US government has also voiced its concern through the White House’s “AI czar”, David Sacks. It therefore comes as little surprise that, having long prohibited Nvidia from selling its most advanced processors to Chinese companies, the US government now wants to ban new Chinese-made components from data centres. Meanwhile, China’s Ministry of Commerce is in talks with the country’s leading domestic players to restrict transfers of training data.

And what about workers? A study covering 22,000 US companies provides some grounds for tempering job apocalypse scenarios: it finds a positive correlation between investment in AI and hiring. Within the sample analysed, the heaviest users of AI increased their headcount by an average of 10.2% two years after adopting the technology. Like earlier studies, this one obviously lacks the benefit of sufficient hindsight to draw definitive conclusions. Nevertheless, investors and governments will undoubtedly be watching future trends very closely, given how socially and economically sensitive the issue remains.

The concerns raised by ECB economists have partly materialised, with the unwinding of the most speculative positions in the technology sector, and particularly in semiconductors. But only partly, as the broadening of market participation has offset the sector’s pullback. This configuration has also enabled the S&P Equal Weight Index to outperform the market-cap-weighted S&P 500 so far this year.

This environment should encourage investors to step up their search for diversification, both across sectors and investment styles, while keeping cash-flow generation as their compass at a time when companies’ cost of capital is once again rising under the impact of higher risk-free rates…

*Apollo, Blackstone, BlackRock, Brookfield, KKR and GS.

**Around $200 billion of issuance by hyperscalers since the start of the year, equivalent to roughly 9% of total global Investment Grade issuance, compared with approximately 2% over the 2022–2024 period.

***The source code can be downloaded, integrated and modified.

****Tokens are processed by specialised subnetworks.

Pierre Pincemaille, General Secretary of Investment at DNCA. This article was finalised in September 4th, 2026.

This promotional document is a simplified presentation and does not constitute a subscription offer or an investment recommendation. No part of this document may be reproduced, published or distributed without prior approval from the investment management company.

DNCA Investments is a trademark held by DNCA Finance