AI-led growth constrained by higher-for-longer rates and financial-stability risks
StrengtheningAI investment can extend the global expansion and improve productivity, but the benefits are likely to be accompanied by tighter financial conditions and greater sensitivity to disappointing AI returns, long-term interest rates, refinancing needs, and geopolitical or energy shocks.
November 18, 2026 — The evidence supports a mixed but fragile scenario: AI is contributing to growth in the United States and euro area and remains a potential source of stronger global productivity, while debt-funded infrastructure, persistent inflation, higher rates, trade frictions, energy disruptions, and valuation risk could limit or reverse the upside. Banking-system capital and liquidity provide some resilience, but do not eliminate the risks.
Geography: Global
Portfolio implications
Your exposure
How this scenario reads against the holdings in your portfolio: which positions carry more exposure, and which look relatively resilient.
Log in to see how this scenario affects your portfolioDevelopments
Expected productivity gains could ease long-term rate pressure
Weakens the thesisFederal Reserve analysis finds that publicly anticipated, persistent productivity gains can lower the 10-year Treasury yield and expected real short rate five to ten years ahead, contrary to the usual expectation that stronger productivity necessarily pushes rates higher.
Why it matters: If AI-driven productivity improvements become credible and durable, they could partly offset the scenario’s higher-for-longer rate constraint by reducing longer-term borrowing costs.
Potential beneficiaries
- AI-intensive firms with long-duration investment plans
- Governments and companies refinancing at long maturities
- Long-duration asset holders
Potential losers
- Investors positioned for persistently higher long-term yields
This is a meaningful countervailing development, but it depends on productivity gains being persistent and publicly anticipated; it does not eliminate near-term financing or financial-stability risks.
U.S.-China chip policy remains selectively restrictive rather than fully relaxed
Strengthens the thesisThe administration moved to block China’s remote access to the latest semiconductors while indicating that no further restrictions were needed on sales of older Nvidia chips to China.
Why it matters: This selective approach preserves some commercial channels while maintaining limits on leading-edge access, leaving AI hardware supply chains and investment assumptions exposed to continued geopolitical and regulatory change.
Potential beneficiaries
- Suppliers of permitted older-generation chips
- Firms able to operate within the restricted export framework
Potential losers
- Chinese users seeking access to the latest semiconductors
- Companies dependent on unrestricted cross-border AI computing access
- Investors facing uncertainty over future export controls
The evidence points to continued technology segmentation and policy uncertainty, rather than either a comprehensive easing or a complete cutoff of AI-chip trade.
Potential penalties for overseas AI data centers raise infrastructure-policy risk
Strengthens the thesisThe U.S. administration said it would seek to penalize U.S. companies that build AI data centers abroad, creating an additional policy constraint on the geographic expansion of AI infrastructure.
Why it matters: Companies may face higher costs, slower deployment, or less flexibility in locating capacity, increasing the sensitivity of AI investment returns to policy decisions and cross-border operating conditions.
Potential beneficiaries
- Domestic data-center developers and suppliers
- U.S.-based infrastructure and power providers
Potential losers
- U.S. companies planning overseas AI capacity
- Foreign jurisdictions seeking U.S. AI infrastructure investment
- Investors in projects exposed to changing location rules
The announcement adds a direct policy risk to AI infrastructure expansion, although the evidence does not establish the penalties’ final design or economic scale.
AI valuation concerns trigger broader equity-market repricing
Strengthens the thesisGlobal equities have come under intensified pressure as investors question whether AI earnings and productivity gains will justify the sector’s infrastructure spending. The selloff has extended across U.S. and Asian markets, alongside expectations of more aggressive Federal Reserve rate increases and tighter U.S. chip controls on China.
Why it matters: This indicates that disappointment over AI returns is becoming a wider market risk, while monetary and geopolitical pressures are amplifying the repricing beyond AI stocks themselves.
Potential beneficiaries
- Investors holding cash or defensive assets
- Companies with lower capital intensity and less AI-related valuation exposure
Potential losers
- AI-focused equities
- Broad equity-market investors
- Companies dependent on uninterrupted access to AI-related capital
High confidence; the development is supported by simultaneous declines in U.S. and global equities and by explicitly reported concerns over AI returns, rates, and chip controls.
Higher financing costs are beginning to constrain data-center funding
Strengthens the thesisLenders are pulling back from data-center loans as funding costs rise. Data-center asset-backed-securities costs increased by more than one percentage point during 2026, while hyperscaler bond spreads widened more than those of comparable non-AI industrial companies.
Why it matters: The financing response creates a direct transmission channel from higher rates and AI-risk concerns to the infrastructure investment needed to sustain the AI buildout, increasing refinancing and execution risks.
Potential beneficiaries
- Lenders with stronger balance sheets and tighter underwriting standards
- Established infrastructure operators with reliable cash flows
Potential losers
- Highly leveraged data-center developers
- AI infrastructure projects reliant on continued cheap financing
- Hyperscalers facing wider borrowing spreads
High confidence; the evidence directly records lender retrenchment, higher securitization costs, and relative widening in hyperscaler bond spreads.
Anthropic IPO converts AI enthusiasm into a major long-term funding commitment
Strengthens the thesisStrong demand allowed Anthropic to price its IPO at $83 per share, implying a valuation of about $1.5 trillion. The offering is intended to support roughly $900 billion in long-term computing and data-center commitments, including $70 billion of planned spending over the next 12 months.
Why it matters: The transaction provides substantial financing for continued AI capacity expansion, but its scale also raises the importance of sustained demand, adequate returns and access to capital for the broader investment cycle.
Potential beneficiaries
- AI chip and cloud suppliers
- Data-center developers
- Electricity and infrastructure providers
Potential losers
- Competing firms unable to match AI capital spending
- Investors exposed to a sharp repricing if expected AI returns disappoint
The IPO is an immediate positive for AI investment momentum, while the unprecedented size of the commitments increases the system's sensitivity to valuation and execution risks.
Energy and permitting constraints become a central bottleneck for AI expansion
Strengthens the thesisThe ECB warned that weaker buffers against energy-supply disruptions and persistent trade-policy uncertainty could raise inflation, weaken growth, increase volatility and impair debt-servicing capacity. In the United States, the Senate was considering permitting changes intended to accelerate generation, transmission and data-center construction as AI-driven electricity demand rises.
Why it matters: AI capacity growth is increasingly dependent on timely power and grid investment. Faster permitting could ease supply constraints, but unresolved energy and trade risks could instead raise operating costs and financial stress.
Potential beneficiaries
- Power generators and transmission developers
- Data-center and grid-equipment firms if permitting improves
Potential losers
- Energy-intensive data centers facing power shortages or higher electricity costs
- Highly indebted firms exposed to supply disruptions and rising financing costs
The evidence points to a policy response aimed at enabling AI-related power buildout, but also shows that energy and trade shocks remain material threats to macroeconomic and financial stability.
Fed signals that AI-related disinflation may not arrive soon
Strengthens the thesisFederal Reserve minutes indicated that a vast majority of policymakers remained concerned that inflation could persist even if gas prices fell and the effects of tariffs and AI-infrastructure construction faded. Markets therefore strengthened expectations for a December rate increase and potentially additional hikes in early 2027.
Why it matters: This raises the near-term cost of capital for AI infrastructure and increases pressure on other rate-sensitive sectors, even if AI investment remains strong.
Potential beneficiaries
- Banks and other lenders
- Money-market and short-duration fixed-income investors
Potential losers
- Highly leveraged AI infrastructure developers
- Rate-sensitive equities
- Borrowers facing refinancing
A clear tightening development: policymakers appear unwilling to treat the fading of temporary AI-construction effects as sufficient evidence that inflation risks have receded.
AI investment continues to support growth and productivity
Strengthens the thesisIncreased AI investment and use are boosting US growth and productivity, while AI investment is also supporting euro-area activity. Stronger AI investment and productivity gains remain an identified upside possibility for the global economy.
Why it matters: This supports earnings, capital spending, and economic activity in technology-intensive and productivity-enhancing parts of the economy, although the evidence does not establish that the gains will be evenly distributed.
Potential beneficiaries
- AI infrastructure providers
- Businesses able to achieve productivity gains from AI
- Economies with strong AI investment capacity
Potential losers
- Businesses unable to adapt to AI-driven competition
- Areas of the economy exposed to a reversal in AI capital spending
The growth impulse is credible in the near term, based on observed investment and productivity effects, but its durability remains uncertain.
AI infrastructure becomes more dependent on credit
Weakens the thesisA rising share of rapidly expanding AI infrastructure investment is being financed through long-term debt and complex off-balance-sheet arrangements. Current spillovers into Treasury and investment-grade bond markets have been small and short-lived, but sustained issuance could change that assessment.
Why it matters: Higher leverage increases sensitivity to interest rates, refinancing conditions, and any reassessment of expected AI returns, potentially transmitting weakness from infrastructure finance into credit and capital markets.
Potential beneficiaries
- Lenders and investors able to manage AI-related credit risk effectively
- Infrastructure projects that maintain strong cash flows and financing access
Potential losers
- Highly leveraged AI infrastructure borrowers
- Credit markets exposed to a sustained increase in AI-related issuance
This is an emerging vulnerability rather than an established systemic shock; the risk would become more material if issuance remains elevated or returns disappoint.
Thesis assessment
| Component | Direction | Note |
|---|---|---|
| AI-driven growth and productivity | strengthening | AI investment is supporting growth and productivity in the US and euro area and remains a global upside possibility. |
| AI-financing and valuation risk | elevated | Debt-funded infrastructure, complex financing structures, and the possibility of disappointing AI returns increase downside sensitivity. |
| Inflation and interest-rate pressure | elevated | Persistent inflation and the risk of higher-for-longer or sharply higher long-term rates constrain financing and valuations. |
| Energy, trade, fiscal, and geopolitical backdrop | elevated | Disruptions and vulnerabilities are weighing on the outlook and could amplify a market correction. |
| Immediate banking-system resilience | partially mitigated | Capital and liquidity remain broadly strong, but identified rate, refinancing, and AI-financing exposures remain relevant. |
Bottom line
The most plausible scenario is continued AI-supported expansion with a higher risk premium: growth and productivity improve, but the cycle is increasingly exposed to higher rates, debt-funded infrastructure, valuation disappointment, and external shocks. Banking resilience lowers the probability of an immediate systemic event, yet does not remove the possibility of a broad repricing if several risks combine.
For a diversified global equity portfolio, the key sensitivity is not simply participation in AI growth but exposure to long-duration valuations, leverage, refinancing needs, energy and trade disruptions, and regional concentration. Resilience is likely greater where businesses have strong balance sheets, durable cash generation, and diversified geographic revenue.

