Principal Global Investors Strategist Weighs In on the AI Capex Cycle and Market Sentiment

Seema Shah, Chief Global Strategist at Principal Global Investors, discussed the AI capital expenditure cycle and its effect on investor sentiment. Bloomberg recorded the conversation with Avril Hong on the sidelines of the Milken Institute Asia Summit in Singapore. No full transcript is publicly available, so we do not retell Shah's specific assessments. But the topic itself, raised at an Asian financial forum, shows where the market's central question lies right now: will the hundreds of billions of dollars that the largest cloud companies are pouring into data centers, chips and energy pay off?

What the announcement tells us

The material is short: it is an announcement of an interview. It contains few facts:

• Speaker: Seema Shah, Chief Global Strategist, Principal Global Investors.
• Interviewer: Avril Hong, Bloomberg.
• Venue: Milken Institute Asia Summit, Singapore.
• Topic: the AI capex cycle and investor sentiment.
• The announcement text contains no estimates, forecasts or quotes.

Any conclusions about Shah's actual position would be speculation for now. But we can look at why such a conversation is at the top of the agenda today.

Why capex has become the key indicator

Analysts estimate that the combined capital expenditure of Microsoft, Alphabet, Amazon and Meta in 2025 approached $350–400 billion. A significant share of that money went into AI infrastructure. For the market, this is not an abstract number. Every dollar of hyperscaler capex becomes revenue for Nvidia, TSMC, Broadcom, SK hynix, and for makers of servers, cooling systems and electrical equipment.

Sentiment around AI spending therefore hits the entire chain at once. When investors believe demand for compute is growing faster than supply, both chip designers and energy companies get more expensive. Once doubt about returns creeps in, valuations of every participant get revisited. A global strategist who looks at stocks, bonds and sectors at the same time has to answer two questions. Can the earnings of the tech giants withstand this level of spending? And how does this cycle affect other markets, from credit to commodities?

Why the conversation took place in Singapore

Asia is not just a backdrop here. Leading-edge chips are made in Taiwan, memory is made in South Korea, and a significant share of packaging and assembly is concentrated in the region. Singapore remains one of the key data center hubs in Southeast Asia, and constraints on energy and land there have long been a regulatory issue. Asian investors see the AI cycle from two sides: as equipment suppliers and as a market where capacity needs to be built.

Which signals to watch

From public data, you can assemble a set of metrics that influence sentiment the most:

• quarterly hyperscaler capex guidance and revisions to it;
• revenue from cloud AI services relative to investment;
• orders and margins at GPU and memory suppliers;
• timelines for connecting new data centers to power grids;
• the cost of financing AI infrastructure, including debt issuance.

How to use this information

Read the announcement as a marker of the topic, not as a source of data. If you follow the AI sector as an investor or a supplier, wait for the full recording of the interview and compare Shah's position with the hyperscalers' earnings reports for the coming quarter. For companies that buy cloud capacity, something else is more useful: providers' capex directly affects GPU availability and inference prices. Ignore headlines about a "bubble" or a "boom" and watch three numbers: the capex plan, cloud revenue growth and capacity utilization. They will show which way sentiment shifts.