投资者问答：AI抛售、高企的CPI以及投资者可以怎么做

• Good news is no longer enough. Strong AI demand remains intact, but valuations, financing costs and returns on investment are becoming more important.
• AI credit risks are in focus. CoreWeave's bond selloff highlights concerns about debt-funded infrastructure expansion, although not all AI companies face the same risks.
• Markets are caught between inflation and growth fears. December Fed tightening remains in market expectations, but slowing growth is becoming an increasingly important consideration.
• Earnings and portfolio balance matter. Investors may want to focus on cash flows, capital discipline and diversification rather than abandoning structural growth opportunities.

The AI industry has delivered some extraordinary numbers over recent weeks.

• Micron: Fiscal fourth-quarter revenue surged to $54.2 billion, from $11.3 billion a year earlier, reflecting exceptional memory demand.
• TSMC: September revenue rose 54.6% year-on-year, confirming continued strength in semiconductor demand.
• Samsung: Preliminary third-quarter operating profit increased nearly ninefold from a year earlier, although it fell short of elevated expectations.
• ASML: Raised its annual sales guidance, with customers planning equipment orders extending towards 2029 and 2030.
• Anthropic: Reported third-quarter revenue of $16.8 billion, up approximately 45% from the previous quarter.
• Oracle: Cloud infrastructure revenue increased 121% year-on-year, reflecting accelerating AI computing demand.

These are hardly signs of a collapsing AI industry. Yet technology stocks have struggled to sustain their rally.

The problem is not that demand has disappeared. Investors are increasingly questioning how much future growth is already priced into valuations, how expansion is being financed and when these investments will generate sufficient returns.

This is playing out as we anticipated in our Q4 outlook, “Raise the Bar,” where we highlighted the cost of capital as an increasingly important driver of the AI trade.

For much of the rally, accelerating AI investment was enough to lift almost everything exposed to the theme. Now markets are becoming more demanding about profitability, financing and capital discipline.

CoreWeave has become one of the clearest warning signals in the AI financing market.

Yields on some of the specialist AI cloud provider's bonds have climbed towards 15%, indicating that investors are demanding significantly greater compensation for credit risk.

The concern is that AI infrastructure requires enormous upfront investment in GPUs, data centres, networking and power systems, while the revenues needed to justify these investments may take years to materialise.

CoreWeave has relied heavily on debt to fund rapid expansion, leaving it exposed to higher financing costs and refinancing risks.

• High borrowing costs: New funding becomes more expensive, potentially reducing the returns on infrastructure investments.
• GPU depreciation: Rapid technological advances could reduce the economic value of existing equipment supporting loans.
• Customer concentration: Dependence on a few major AI customers creates risks if spending or contract terms change.
• Cash-flow timing: Infrastructure costs are incurred upfront, while contracted revenues arrive over several years.

CoreWeave has pushed back against concerns about its financial position. CEO Mike Intrator argues that markets misunderstand its business model, highlighting its substantial contracted revenue backlog of approximately $104 billion.

Long-term customer commitments provide meaningful visibility, and elevated bond yields do not necessarily indicate an imminent default.

Nevertheless, the repricing of CoreWeave's debt suggests that investors are becoming less willing to finance AI infrastructure growth at any cost.

Yes, although it is important to distinguish between companies facing genuine financing pressures and those whose main vulnerability is elevated valuations.

Several areas of the AI ecosystem deserve closer attention.

Large technology companies generally have stronger balance sheets, more diversified revenues and greater ability to fund investment internally. Some smaller infrastructure providers also have meaningful cash balances and committed financing.

However, the AI ecosystem is increasingly interconnected. Infrastructure providers depend on financing and large customer contracts, while equipment suppliers depend on their customers continuing to expand capacity.

If higher borrowing costs eventually force some infrastructure providers to moderate spending, the consequences could extend across the supply chain.

There is insufficient evidence of a systemic AI credit crisis. But the financing model behind the infrastructure boom is facing greater scrutiny.

Investors should increasingly distinguish between strong demand, sustainable financing and attractive valuations. They are not necessarily the same thing.

September US inflation came in hotter than expected, with headline CPI accelerating to 3.7% year-on-year and core inflation reaching 2.6%.

The initial reaction was predictable: Treasury yields and the dollar rose as investors increased expectations of another Fed rate hike.

But the subsequent retreat in Treasury yields, accompanied by episodes of dollar weakness and strength in the yen and Swiss franc, reflects a competing concern.

Markets are beginning to worry that the Fed may need to tighten policy just as the US economy is losing momentum.

• Growth fears are competing with inflation concerns. Higher borrowing costs could weaken consumption, employment and business investment.
• Treasuries are regaining safe-haven appeal. Equity weakness is encouraging demand for government bonds, pushing yields lower despite inflation risks.
• Currency markets are becoming more sensitive to growth expectations. Lower Treasury yields can weigh on the dollar, while the yen and Swiss franc benefit from defensive flows.

This is a potentially important change in market behaviour.

Earlier in the selloff, higher bond yields were themselves a source of pressure on equities. Now, periods of falling yields alongside weaker stocks suggest investors are also becoming concerned about growth.

However, a sustained Treasury rally is not guaranteed. Sticky inflation, fiscal risks and heavy government borrowing could keep longer-term yields elevated.

The market is caught between two risks: inflation that keeps the Fed restrictive and slowing growth that makes further tightening more dangerous.

Markets continue to anticipate another Fed rate hike in December, but the outlook is becoming increasingly dependent on incoming growth data.

The latest inflation report reinforces the case for keeping monetary policy restrictive. However, signs of slowing growth could complicate the decision, particularly if employment and consumer spending begin to weaken.

For now, rate cuts are not part of the market's near-term expectations. The debate is increasingly about the timing and extent of further tightening, rather than when easing might begin.

The key data to monitor will be employment, consumer spending, business investment and whether inflation pressures continue to broaden.

A weaker growth outlook could reduce expectations for future rate hikes, but that would not automatically be positive for equities if corporate earnings were also coming under pressure.

The first major US bank results suggest that the economy remains resilient, although higher rates and bond-market volatility are creating pressure in parts of the financial system.

• Equity trading remains strong. Goldman Sachs reported a 72% year-on-year increase in equities trading revenue, while JPMorgan and Morgan Stanley posted gains of 49% and 41%, respectively.
• Fixed-income trading has been weaker. Revenue from fixed-income, currencies and commodities declined at several major banks, reflecting challenging conditions in bond markets.
• Higher rates are a double-edged sword. Lending income has benefited some banks, but higher funding costs, bond-portfolio losses and potentially weaker loan demand could become greater challenges.

Consumer spending and lending activity have remained reasonably resilient, suggesting the economy is not yet showing broad recessionary conditions.

However, strong trading revenues may be difficult to sustain indefinitely, while higher interest rates could eventually weigh on credit demand and asset quality.

The takeaway for the broader earnings season is that headline earnings resilience may conceal growing differences in underlying business performance.

Investors should look closely at revenue quality, margins, funding costs and forward guidance rather than simply whether companies beat earnings expectations.

This could be one of the most important earnings seasons for the AI trade.

Recent results from semiconductor manufacturers suggest underlying AI demand remains strong. The challenge is whether major technology companies can demonstrate that this investment is translating into sustainable profits and cash flows.

Microsoft, Alphabet and Amazon will provide important evidence of enterprise AI adoption and demand for cloud computing.

Strong cloud growth would reinforce the structural demand story. But investors will also want to understand how quickly AI services are contributing to revenues and margins.

Higher spending commitments may support semiconductor and equipment suppliers, but could create renewed concerns about investment returns.

Markets may become particularly sensitive to another round of substantial capex increases without clearer evidence of financial payback.

Depreciation, energy costs and infrastructure spending could offset some of the benefits of strong AI revenue growth.

The distinction between companies funding expansion through operating cash flow and those increasingly dependent on external financing will remain important.

Investors will look for evidence that AI is generating revenues or productivity benefits across software, advertising, digital services and enterprise applications.

This could support a broadening of the AI opportunity beyond chips and data centres.

With expectations elevated, guidance may matter more than headline earnings beats.

Companies demonstrating improving utilisation, pricing power and disciplined investment could receive a more favourable market response than those simply announcing larger spending plans.

The most supportive outcome would be strong AI demand accompanied by improving cash flows and clearer returns on investment.

Conversely, excellent revenue results alongside sharply higher spending commitments could leave investors disappointed.

This is becoming an important risk given how concentrated the US equity rally has been.

A relatively small group of technology companies now accounts for a substantial share of major US and global equity indices. That means weakness in AI-related stocks can drag down broader markets even when earnings elsewhere remain resilient.

There are three potential spillover channels:

• Index concentration: Declines in technology heavyweights can have an outsized impact on the S&P 500 and global equity benchmarks.
• Portfolio de-risking: Investors facing losses in crowded technology positions may reduce exposure elsewhere, creating selling pressure even in fundamentally sound companies.
• Economic spillovers: If higher financing costs eventually slow AI infrastructure investment, the effects could spread to semiconductor suppliers, industrial equipment manufacturers, power infrastructure and construction.

However, an AI correction does not necessarily have to become a broader equity bear market.

If corporate earnings remain resilient and economic growth holds up, investors could rotate towards other sectors, including healthcare, selected financials, industrials and businesses with more reasonable valuations.

The key distinction is between contagion and rotation. A market where AI stocks correct while other sectors gain leadership may ultimately be healthier than one where the entire market remains dependent on a handful of companies.

What investors should watch is whether market breadth improves as technology weakens, or whether selling pressure begins spreading across sectors.

For much of the recent inflation-driven selloff, equities and bonds have struggled together. Rising yields have weighed on technology valuations while pushing bond prices lower, reducing the diversification benefits of traditional stock-and-bond portfolios.

However, the latest market moves suggest that relationship may be changing.

Despite hotter US inflation, Treasury yields have retreated as equities weakened, with investors increasingly concerned about slowing growth and the risk of excessive Fed tightening.

This raises the possibility that government bonds could regain some of their traditional safe-haven role.

The distinction depends on what is driving the selloff:

• Inflation shock: If inflation remains elevated and markets price in further tightening, both equities and longer-duration bonds could fall.
• Growth shock: If economic activity weakens sharply, high-quality government bonds could benefit from safe-haven demand and lower expected interest rates.
• Stagflation: If inflation stays high while growth deteriorates, bonds may offer less reliable protection, particularly at longer maturities.

The return of a negative stock-bond correlation would be helpful for diversified portfolios, but it cannot be taken for granted.

For investors, this argues for a more measured approach to duration rather than assuming that long-term government bonds will always offset equity losses.

We remain constructive on equities over the medium term, but the hurdle for further gains has increased.

• Corporate earnings are still relatively resilient.
• AI continues to drive strong demand across technology and infrastructure.
• Broader earnings growth beyond the largest technology companies could support a more diversified rally.

However, elevated valuations, persistent inflation, higher financing costs and concerns about economic growth leave less room for disappointment.

Three possible outcomes are worth considering.

A correction does not necessarily signal the end of the bull market.

But the next phase may depend less on further valuation expansion and more on earnings delivery, financial resilience and broader participation across sectors.

Investors should also be mindful of concentration risk. Holding US equity ETFs, global index funds and individual technology shares can create substantial overlapping exposure to the same large AI companies.

The opportunity in equities has not disappeared, but the margin for error has narrowed.

The current environment does not necessarily call for abandoning equities or AI. But with inflation and growth risks competing, portfolios may benefit from greater balance.

High-quality government bonds could benefit from safe-haven demand if economic growth weakens.

Short-duration bonds offer income with relatively lower interest-rate sensitivity, while selective intermediate-duration exposure could provide greater protection if yields decline during a growth slowdown.

However, persistent inflation, fiscal concerns and heavy Treasury issuance could limit gains, particularly in longer maturities.

The next phase of AI may reward financial strength and monetisation over aggressive capital spending.

Different business models offer exposure to AI without necessarily carrying the same infrastructure financing requirements:

• Asset-light platforms: Apple's device and services ecosystem represents a relatively less capital-intensive approach to AI, although valuation remains a consideration.
• China technology: Platforms such as Tencent offer exposure to AI adoption through existing digital businesses, while cost-efficient models provide a different growth narrative.
• Enterprise software and cybersecurity: AI adoption could support automation, productivity and digital security demand.
• AI productivity beneficiaries: Financial services, healthcare and industrial businesses may benefit from efficiency gains without undertaking the same scale of infrastructure investment.

Healthy balance sheets, sustainable cash flows and reasonable valuations remain important across these areas.

Gold may provide diversification against geopolitical and fiscal uncertainty, while copper, selected industrial metals and energy infrastructure have structural support from electrification, data centres and rising power demand.

However, these exposures carry different risks. Gold can face pressure from higher real yields, while industrial commodities remain vulnerable to a slowdown in economic growth.

Healthcare combines relatively resilient demand with longer-term structural drivers such as ageing populations, medical innovation and growing healthcare needs.

It also provides diversification away from crowded AI infrastructure trades, although regulatory, drug-pricing and company-specific risks remain relevant.

The objective is not to predict a single market outcome, but to avoid having the entire portfolio depend on it.