A command center for your financial life

• AI capital spending contributes 1 full percentage point to U.S. GDP growth, making roughly half of projected 2026 expansion dependent on a single theme.
• With 50% of the S&P 500 in tech/AI names and 50% of new corporate debt AI-related, traditional index and 60/40 portfolios have become an undiversified wager on AI.
• The Fed is expected to hold or even hike rates through year-end, as sticky inflation from tariffs and AI cost pressures keeps cuts off the table.
• Hyperscalers’ historic shift from cash-generators to capital-raisers, combined with zero-moat pricing wars and customer pushback on token costs, reveals deep structural flaws in the AI business model.
• Outside AI, the K-shaped consumer and a $500 billion private-credit bubble in software loans (with a 2028-29 maturity wall) pose serious, if contained, risks.

• September 15, 2026 — September FOMC Decision
• November 3, 2026 — US Midterm Elections
• December 15, 2026 — December FOMC Decision
• January 2027 — Tax Refund Tailwind Expires
• 2028 — Software Private Credit Maturity Wall

The U.S. economy and financial markets are now overwhelmingly driven by AI-related spending. AI capital expenditure (data centers, energy, chips) directly adds about 1 percentage point to GDP, which accounts for half of the roughly 2% growth expected in 2026. The other growth sources—reshoring/industrial renaissance (~0.3 pp) and one-time tax refunds from the “big beautiful bill” (~0.9 pp)—are either smaller or temporary. When the tax refund tailwind vanishes in 2027, GDP growth will slip meaningfully unless AI spending accelerates further.

• AI capex is interest-rate-insensitive, which explains why the economy remains resilient despite high rates hurting housing and autos.
• The tax refund boost is a one-off; without it, 2027 growth could fall below 2%.
• AI’s dominance extends beyond GDP: it now dominates the 60% equity portion of portfolios (top 10 stocks are 42% of the S&P 500), the 40% bond portion (hyperscaler debt issuance), and venture capital (87% AI).

Markets have become a massive single-factor bet. Index investing no longer provides diversification—half the S&P 500’s market cap is tech and tech-adjacent, and half of newly originated corporate debt is AI-related. A failure of AI to deliver the priced-in revenue would not only tank equities but hit investment-grade credit, causing losses across both sides of a 60/40 portfolio. The traditional safe-haven role of bonds is compromised when the same hyperscalers dominate debt markets.

• A “new” 60/40 recommendation: 60% AI, 40% non-AI hedges, because value and uncorrelated assets (consumer staples, etc.) would rocket if the AI trade stumbles.
• Even foreign buyers of Treasuries are in it partly for dollar-denominated AI exposure, tying U.S. fiscal stability to the tech theme.
• The concentrated bet means that a recession would likely be triggered not by classic cyclical forces but by an AI earnings disappointment.

The AI investment thesis is showing serious flaws. Hyperscalers that once generated massive free cash flow are now raising historic amounts of capital. Google raised $85 billion in equity, and the sector’s free cash flow is plunging toward zero. This shift from cash-rich to capital-intensive raises questions about terminal returns on the colossal capex.

• No durable competitive moats have emerged; users easily switch between models, and cheap open-source and Chinese alternatives threaten token prices.
• A pricing war is already evolving as the initial subsidized token pricing gives way to true costs, causing corporate customers to limit employee AI usage.
• The early bull case concentrated risk—Oracle’s backlog was heavily exposed to OpenAI—showing a fragile customer base.

Investors initially rotated from hyperscalers to semiconductor and power names, but that trade is also wobbling. Samsung’s stock dropped 7% despite a 1,800% profit increase on fears that hyperscaler spending growth will slow. The entire AI value chain is interconnected, so any slowdown in end-demand risks a cascading unwind.

Beneath the AI-driven aggregate data, the U.S. consumer is splitting in two. High-income households have gained roughly $1.5 trillion in savings since 2019, while bottom-20% savings are flat in nominal terms. Wage growth and inflation also diverge sharply by income, with lower-income groups hit harder by rising costs.

• The top 20% of earners account for 40% of all consumer spending, which props up headline consumption even as stress builds below.
• Auto, credit-card, and student-loan delinquency rates are rising, especially among lower-income cohorts.
• The widening K-shape is becoming a political issue, and while it hasn’t broken aggregate demand yet, it adds social and policy risk.

A pocket of real vulnerability lies in private credit, specifically software loans originating from the 2021–22 buyout boom. About $500 billion in such loans sit in BDCs, pensions, and insurance companies, carrying high leverage and deteriorating interest-coverage ratios as rates stay higher for longer.

• Software loans now trade at 12%+ yields in the secondary market, signaling significant distress.
• A maturity wall looms in 2028–29 when many of those vintage loans come due; the equity in many deals may already be wiped out.
• Unlike the GFC, this problem is not systemically levered, so the macro spillover risk is contained but still painful for exposed investors.

The chance of a Fed rate cut this year is “zero,” and markets are actually pricing two rate hikes (September and December). Inflation remains sticky due to tariffs, oil prices, and AI-related cost pressures, keeping the Fed on hold or even tightening further.

• Higher rates continue to suppress interest-rate-sensitive sectors like housing and autos, while the AI-driven portion of the economy hums along.
• The housing market is a casualty: weak supply and high mortgage rates are crushing affordability and activity.
• The Fed’s stance creates a two-speed economy where capital-intensive AI firms can raise equity, but smaller, rate-exposed businesses struggle.

One-off topics from the week: tensions with Iran and retaliatory U.S. attacks briefly impacted Circle’s stock, which fell 17.5% after a consortium including Visa, Mastercard, and BlackRock unveiled a competing stablecoin ecosystem. Nike reported an EPS beat but flat annual revenue, and cautious guidance caused the stock to drop; the host views it as a perpetual turnaround story with no near-term catalyst.

The U.S. economy and financial markets have become a concentrated, leveraged bet on AI delivering enormous returns, with half of GDP growth and portfolio performance hitched to tech spending. Cracks in the AI business model—no moats, capital intensity, and customer pushback—suggest that bet may be overpriced, while the K-shaped consumer and an approaching private-credit bust in software add layers of risk. Investors should actively seek non-AI hedges before the temporary tax tailwind fades and the AI narrative stumbles, potentially triggering a broad and painful drawdown.