AI is so big, it’s now impossible for investors to avoid

Apollo’s Torsten Slok argues the traditional 60/40 stock-bond allocation is effectively replaced by a “new 60-40” framework: AI vs. non-AI. The note claims stocks, corporate bonds, and venture capital are increasingly dominated by this single theme. The takeaway for portfolios is that factor exposure is shifting from asset class safety to AI-driven winners/losers.

Analysis

The key market implication is not that AI is “big,” but that it is now acting like a quasi-macro factor: it is widening dispersion across both equity and credit, which is usually a feature of late-cycle leadership, not just early-cycle growth. That tends to favor a narrow set of balance-sheet-rich platforms and picks-and-shovels suppliers while starving long-duration, non-AI software, legacy IT services, and lower-quality venture-backed issuers of capital. In practice, the next 1-3 months should show continued multiple support for the cap-weighted large-cap complex, but broader market breadth can keep deteriorating even if headline indices hold up.

Second-order effects are most interesting in credit and private markets. As lenders and VC funds crowd into AI stories, non-AI borrowers likely pay a relative funding penalty, which can translate into slower hiring, lower marketing spend, and more M&A at distressed prices among software and internet fringe players. Over 6-18 months, the risk is that AI capex becomes a quasi-utility buildout: if monetization lags the spend wave, the market will start punishing whoever financed the build—particularly if depreciation, power costs, or working-capital intensity surprise to the downside.

The contrarian miss is that the “AI vs. non-AI” framing may be overbroad for stock selection but right for factor exposure. The trade is less about owning every AI logo and more about owning the enablers with real free-cash-flow conversion versus shorting the most crowded financing stories and weak second-derivative beneficiaries. The setup reverses if AI capex guidance rolls over, hyperscaler spending slows, or the market starts demanding proof of monetization rather than narrative.

Key Decisions for Investors

• Long QQQ / short RSP for a 1-3 month factor trade: the cap-weighted index should keep outperforming if AI concentration persists, while equal-weight is more exposed to breadth decay; exit if market breadth materially improves or megacap AI spend guidance rolls over.
• Long SMH / short IGV on a 3-6 month horizon: semis and equipment remain the clearest monetization path, while much of application software faces margin pressure from rising customer scrutiny; thesis breaks if software guides accelerating AI-driven upsell traction.
• Overweight power-grid and electrical infrastructure names (ETN, PWR, NEE, VST) as the more durable AI-enablement basket; these have better visibility than speculative AI application names and should benefit for 6-18 months if data-center buildout remains intact.
• Watchlist alert: short high-beta non-AI venture proxies or software lenders on any post-earnings bounce if they show slower net retention and higher CAC payback; the fundamental falsifier is an improvement in monetization metrics without commensurate capex growth.