From the studio

Podcast: Signal over Noise with Ulrike Hoffmann-Burchardi (7 mins)
Podcast: Jump Start | AI selloff, US data, and global central bank decisions (8 mins)
Video : Mark Haefele’s five New Year resolutions for 2026 (3 mins)
Video: CIO’s Leslie Falconio on credit opportunities for 2026 (6 mins)

Thought of the day

Weakness in tech stocks at the end of last week disrupted a rally supported by the Federal Reserve’s interest rate cut. The Nasdaq fell 1.7% on Friday despite Broadcom’s solid quarterly results that included better-than-expected revenue and sales guidance. Instead, investors were concerned over a potential consolidation in the chipmaker’s gross margins, while Oracle’s higher-than-expected capital expenditures for fiscal 2026 added to market unease over AI profitability.

But while we expect global AI capex to continue to rise over the coming years, we do not see evidence of an investment bubble. In fact, we have said that the race to artificial general intelligence could trigger a capex cycle where the capex of the enabling layer is dissociated from the near-term monetization potential of the application layer. And this pattern is consistent with previous innovation cycles.

We look at where we are in the AI investment cycle, and address some of the market concerns to help investors navigate the path forward:

Demand for AI compute remains strong. Without taking any single-name views, Broadcom’s AI chip backlog and recent updates from major tech companies point to continued strong demand for AI compute, with surging usage of tokens—the basic unit of compute in AI models—a positive driving force for the sector. In fact, as AI adoption expands from consumer chatbots to broader enterprise and industry use cases, we estimate that the required compute capacity could be orders of magnitude greater than today’s installed base.

Value creation is increasingly shifting beyond the enabling layer. While both Broadcom and Oracle reported healthy results, their share price reaction suggests that capex increases and order backlogs no longer automatically translate to stock performance. A key lesson from transformational innovation historically is that the majority of value captured over time accrues to the application layer—companies that use the new technology to generate cash flows. As the AI story advances into its fourth year, we believe the beneficiaries of AI are broadening out— not just within the technology industry (to platforms and companies with robust AI use cases), but also to other sectors of the economy such as health care and financials.

AI’s return potential is large, and company fundamentals in aggregate remain robust. While revenue diffusion at the application layer has lagged the rapid pace of infrastructure build-out, our conviction in the structural return on AI investment remains strong. We expect the total addressable market for AI to be USD 3.1tr in revenue potential by 2030, suggesting a compound annual growth rate of 30% over the next five years. Looking at fundamentals, we expect major tech companies to collectively maintain stable margins, while valuations remain reasonable compared to historical bubble peaks.

So, we believe the AI story remains intact, and expect a more widespread capture of AI value creation to support a broadening of leadership in equity markets. We recommend diversification beyond the enabling layer and into the application layer, and suggest limiting exposure to companies trading at elevated price-to-earnings multiples.