AI capex boom meets Wall Street risk: are agentic bets and private credit about to reprice?

Goldman Sachs is warning investors that the market may be underestimating the scale of the AI spending cycle, with forecasts pointing to roughly $920 billion in AI-related capex in 2027 that strategists consider likely too conservative. The message is reinforced by a separate Goldman note that analysts are undercounting the AI spending boom, implying further upside for stocks tied to the theme. At the same time, the risk narrative is shifting from “AI adoption” to “AI financing and valuation,” as Goldman also frames rising capex as a source of downside risk for AI equities if expectations outrun returns. In parallel, Bank of America is stepping into the “agentic AI” trade and issuing double upgrades, signaling that large banks are still willing to underwrite momentum in AI-linked winners. Strategically, this cluster matters because AI capex is becoming a macro-financial lever: it can tighten or loosen liquidity across software, semis-adjacent supply chains, and private capital markets depending on how quickly spending converts into earnings. The power dynamics are increasingly between capital providers and capital allocators—public market analysts and banks are effectively validating the spending thesis, while private equity and private credit managers face the risk that return squeezes and liquidity constraints will force repricing. Apollo’s view that huge private equity software bets are most at risk highlights a potential divergence: public “AI winners” may keep rallying while leveraged private portfolios struggle to refinance or exit at favorable multiples. The net effect is that AI is not only a technology theme but also a stress-test for financial intermediation under higher-for-longer rates. Market and economic implications are immediate for AI-exposed equities, software private credit, and the broader credit complex. If AI capex is indeed larger than consensus, investors may bid up infrastructure and platform beneficiaries, but the “risks for AI stocks” framing suggests higher volatility and drawdown risk around earnings visibility. Private equity software is flagged as vulnerable as investors demand higher returns amid a returns squeeze, which can pressure valuations and increase discount rates applied to growth cash flows. In credit markets, Man Group’s comments that private credit has “growing pains” while higher-for-longer rates create yield opportunities point to a bifurcated environment: spreads may remain supported, yet liquidity and underwriting standards could tighten, affecting instruments such as private credit funds and related structured products. What to watch next is whether the capex narrative translates into measurable revenue acceleration and margins, or whether it triggers a valuation reset for AI equities. Key indicators include forward guidance from AI infrastructure and software firms, changes in analyst capex assumptions for 2027, and evidence of refinancing stress in private credit portfolios. For private markets, monitor liquidity metrics and any signs of slower exits or wider bid-ask spreads in software buyouts, as these would validate Apollo’s “most at risk” framing. On the policy and sentiment side, President Trump’s suggestion that Americans could share in AI company wealth adds a political overlay that could influence tax, labor, and distribution expectations—watch for follow-on statements or proposals that affect investor appetite. The escalation trigger would be a visible earnings miss combined with widening credit stress; de-escalation would be strong guidance that confirms higher capex is producing durable cash flows.

Geopolitical Implications

• 01

  AI capex is increasingly a macro-financial battleground inside the US, shaping how capital flows between public markets and private credit under higher-for-longer rates.

• 02

  If private equity and private credit face liquidity stress, it can slow AI commercialization and shift bargaining power toward large public platforms with better access to capital.

• 03

  Political narratives about sharing AI wealth may foreshadow regulatory or fiscal interventions that affect investor risk premia and corporate governance expectations.

Key Signals

• —Next-quarter earnings guidance from AI infrastructure and software firms on capex-to-revenue conversion.
• —Updates to 2027 AI capex assumptions by major sell-side houses and whether they converge upward or diverge.
• —Private credit liquidity metrics: redemption behavior, NAV volatility, and widening spreads in private credit benchmarks.
• —Evidence of refinancing stress or slower exits in private equity software deals.
• —Any follow-on US policy proposals tied to AI wealth-sharing, taxation, or worker participation.