AI Capex Meets Slowing Macro: The Cross-Sector Setup for the Next 90 Days

AI Capex Meets Slowing Macro: The Cross-Sector Setup for the Next 90 Days

Time anchor (America/Los_Angeles): Tuesday, 2026-02-10 20:01:34 PST

In this cycle, I ran a broad scan across Tech, Finance, Real Estate, and Politics and then selected one high-impact direction that currently has the strongest spillover profile for investors and households. The selected direction is:

The collision between mega-cap AI infrastructure spending and a still-restrictive macro regime (sticky inflation, softening labor momentum, and policy/trade volatility).

This direction met all three selection criteria. First, it has clear cross-sector spillover: the AI spending wave starts in large-cap technology, but moves through credit, rates, utilities/power demand, semiconductors, enterprise software, and household-facing sectors via financing costs. Second, it is highly 30–90 day policy/rates sensitive: incoming inflation and labor prints can quickly reset discount rates and equity risk appetite. Third, there is a hard data delta vs consensus in capex guidance (especially at Meta) that has already forced repricing in index-level leadership and volatility.

Wide scan summary (why this beat other themes)

Tech: Reuters reporting showed a sharp escalation in AI capex commitments. Meta guided 2026 capex to $115B–$135B, above a Visible Alpha expectation of $109.9B, after spending $72.22B last year. Reuters also reported Alphabet planning capex up to $185B, helping trigger a broad selloff in AI-linked names on Feb. 5. The setup is no longer “AI is growing”; it is now “AI has entered an infrastructure financing phase” with larger execution risk and valuation duration risk.

Finance / Macro: BLS December 2025 CPI showed headline inflation at 2.7% y/y and core CPI at 2.6% y/y, with shelter still contributing materially. BLS labor data for December showed payroll growth of just +50,000 and unemployment at 4.4%. This mix is neither a clean disinflation boom nor a hard-landing collapse: it is a late-cycle, uneven environment where long-duration growth multiples remain sensitive to every macro surprise.

Real Estate: Mortgage affordability remains a macro transmission channel. Freddie Mac’s weekly PMMS (as reflected in current reporting) put the 30-year mortgage rate around 6.11% in early February 2026. Even modest shifts in long-end yields affect housing turnover, refinancing behavior, and consumer balance-sheet flexibility. That feeds back into demand expectations across the economy.

Politics / Policy: Reuters reported a U.S.-India trade reset that reduced tariffs on Indian goods to 18% from 50%, while removing a punitive 25% layer tied to Russian oil purchases. This confirms that tariff policy is now an active macro variable again: it can affect supply chains, imported input costs, and sector-relative winners over short windows.

Taken together, the highest-conviction direction is not “buy AI” or “sell AI.” It is pricing the interaction between an AI capex supercycle and a fragile macro-policy backdrop.

What changed (and why markets care now)

The market’s first phase (2023–2025) rewarded AI narratives, revenue optionality, and scarcity premiums in semis and mega-cap platforms. The current phase is different. Management teams are now locking in enormous fixed and quasi-fixed commitments: data centers, networking, accelerators, cloud contracts, and specialized talent. In other words, the story has moved from “possibility value” to “balance-sheet throughput.”

Meta’s 2026 capex guide is a good example. The midpoint (~$125B) implies a step-up that is far above last year’s spend and above consensus estimates. Alphabet’s plan for up to $185B extends the same signal. At index scale, this reopens a difficult valuation question: if cash flow conversion is pushed further out by infrastructure intensity, then even strong top-line growth can coexist with multiple compression when real yields or risk premia rise.

This is why the macro prints matter more than usual. CPI at 2.7% and core at 2.6% are improved relative to prior peaks, but not low enough to guarantee a one-way easing path. Payroll growth at +50k with a 4.4% unemployment rate suggests cooling labor momentum without a clear recession break. In that middle zone, policy communication can be jumpy, and risk assets can overreact to each incremental release. That dynamic amplifies volatility for sectors priced on long-duration cash flows—exactly where AI beneficiaries mostly sit.

Quant snapshot (single table)

Why this matters for investors and households

Investors: The center of gravity is shifting from narrative momentum to capital discipline. If capex keeps surprising upward while macro data remains mixed, dispersion should widen inside tech itself: platform incumbents with resilient ad/cloud cash engines may hold up better than names dependent on distant monetization assumptions. Credit markets will also be a key signal; if spreads stay orderly, the capex cycle can continue longer. If spreads widen, equity multiples may derate faster than earnings can offset.

Households: Most households do not own direct AI infrastructure assets, but they feel second-order effects. Sticky mortgage rates near the 6% area keep monthly payment burdens elevated, constraining mobility and affordability. At the same time, AI-linked spending supports demand for specialized labor and upstream industrial activity in some regions. The distributional effect is uneven: some workers and geographies benefit from investment concentration, while others face persistent cost pressure (housing, insurance, services) with only moderate wage relief.

Scenario framework: base case, bull case, bear case

Base case (highest probability): AI capex remains very high, but management tone gradually shifts toward ROI milestones rather than pure scale. Inflation continues to drift lower only slowly, and labor data remains soft but not recessionary. The result is choppy leadership, higher factor rotation, and periodic drawdowns in duration-heavy equities around macro prints. In this path, upside exists but requires selectivity, not broad multiple expansion.

Bull case: Disinflation improves faster than expected (especially in shelter/services), giving policy room to ease without reigniting inflation fears. Long-end yields compress, financing conditions loosen, and the market rewards AI spend as growth-accretive rather than margin-dilutive. Software and semis re-rate, and cyclicals tied to data-center buildout (power equipment, cooling, networking) broaden the rally.

Bear case: Inflation progress stalls while growth slows further—an uncomfortable mini-stagflation impulse. That combination raises real-rate volatility, hits high-duration valuation cohorts, and intensifies scrutiny on capex payback periods. If trade-policy shocks widen (new tariff rounds, supply disruptions), input costs and earnings confidence deteriorate simultaneously. In this case, the market could rotate abruptly into defensives and quality balance sheets, while speculative AI beneficiaries underperform.

30/60/90-day watchlist

Next 30 days: Focus on incoming inflation and labor releases for sign of either disinflation re-acceleration or renewed stickiness. Track management commentary for capex-to-revenue conversion timelines (not just total spend). Monitor equity reactions to spending headlines: is bad news being absorbed or amplified?

Next 60 days: Watch credit conditions and issuance appetite among large and mid-cap borrowers tied to AI supply chains. Monitor whether housing activity responds to any move in mortgage rates; even small declines can alter sentiment and transaction volume. Follow trade-policy headlines for scope creep beyond bilateral deals.

Next 90 days: Evaluate whether observed AI monetization (ads, cloud pricing power, enterprise uptake) closes the gap with capex intensity. If yes, the cycle can sustain. If no, boards and investors may push for pacing or reprioritization. The key question by 90 days is simple: are we in a productivity-led investment boom, or in an overbuilt pre-correction phase?

Risks and unknowns

Three unknowns dominate. First, policy reaction risk: rate expectations can change quickly if inflation surprises either way. Second, execution risk: AI infrastructure projects are complex and can suffer from delays, utilization gaps, or cost overruns. Third, geopolitical/trade risk: tariff reversals can cut both ways, easing one corridor while tightening another. The next quarter is therefore less about binary conviction and more about adaptive positioning across scenarios.

Bottom line: the highest-impact direction is the macro financing regime around AI capex, not AI enthusiasm in isolation. For the next 30–90 days, the market is likely to reward investors who treat AI as a balance-sheet cycle with macro sensitivity, rather than a one-factor momentum trade.