In an era where technology markets are defined by sharp volatility and analysts are questioning if the Artificial Intelligence (AI) bubble is nearing a saturation point, Joyce Chang, JPMorgan’s Chair of Global Research, offers a steadying, yet rigorously analytical perspective. Speaking on Bloomberg Television, Chang emphasized that the AI capital expenditure (capex) story remains "intact," highlighting that tech giants have little choice but to maintain their massive investment trajectories.
The Technological Arms Race and Strategic Necessity
Chang’s core thesis is built on the premise that AI is not merely a passing trend but a new "arms race" in the digital realm. The titans of technology—Microsoft, Google, Amazon, and Meta—are funneling billions into infrastructure, semiconductors, and data centers. This Capex is not optional. As JPMorgan points out, failing to invest in AI at this juncture is equivalent to a strategic retreat that could cost a company its market leadership for the next decade.
Despite concerns regarding when these investments will begin to yield tangible returns (ROI), Chang observes that retail investor flows remain remarkably resilient. "There is still a sentiment to buy the dips," she notes, suggesting that confidence in the long-term potential of the technology remains high, even amidst short-term price corrections in tech stocks.
From Hardware to Infrastructure: The Shifting Focus
While 2024 and 2025 were dominated by the meteoric rise of Nvidia and chipmakers, 2026 is emerging as the year of "physical infrastructure." Chang analyzes that Capex is gradually shifting toward energy and networking. Artificial intelligence requires staggering amounts of electricity, making utility companies and energy solution providers the new beneficiaries of the AI narrative.
- Energy Demands: The need for sustainable and stable power is driving new agreements between tech giants and nuclear or renewable energy providers.
- Network Modernization: Data transmission infrastructures must be upgraded to handle the massive throughput required by Large Language Models (LLMs).
- Geopolitical Dimensions: The concentration of chip production in Asia remains a risk factor, driving further investment into Western-based manufacturing facilities (fabs).
Market Psychology and Retail Resilience
One of the most compelling points in Chang’s analysis is the behavior of retail investors. Unlike previous tech bubbles, today’s investors appear more informed and resilient. JPMorgan observes that whenever there is a correction in the tech sector, retail flows tend to increase, supporting valuations. This phenomenon creates a safety cushion, preventing a total collapse and allowing the market to digest high valuations through organic growth and earnings expansion.
"The question is not whether Capex will continue, but how it will be allocated as we move from the model training phase to the implementation phase in the real economy," Chang states.
Conclusion: A Long-Term Commitment
JPMorgan’s analysis concludes that the AI investment story is still in its early chapters. Although growth rates may eventually normalize, the structural need for computational power and automation will continue to fuel capital expenditures. For investors, the challenge is no longer determining if AI is a real phenomenon, but identifying the players who can most efficiently manage the costs of this historic transition.
