`There Is No AI Bubble,' Says BI's Rob Schiffman

Bloomberg Intelligence’s Robert Schiffman said the AI buildout borrowing spree is coming from the "Mount Rushmore" of credit and argued there is "no AI bubble." The comments, made at the Bloomberg Global Credit Forum on June 3, suggest continued confidence in AI-related financing and credit market support. The article is commentary rather than a data-driven market event, so near-term price impact is likely limited.

Analysis

The key signal is not the headline view that AI funding is easy; it is that the marginal borrower is still viewed as top-tier credit. That matters because it extends the runway for hyperscale capex without an immediate funding penalty, which keeps the AI spending cycle self-reinforcing for longer than skeptics expect. In the near term, that supports a broad risk-on bid in the AI supply chain, especially where revenue visibility is tied to backlog rather than spot demand.

Second-order, the credit channel changes the competitive map. If the best-rated issuers can lever up at relatively benign spreads, they can front-load infrastructure and lock in scarce inputs, which pressures smaller cloud, networking, and power-adjacent competitors that need to finance capacity at worse terms. The real bottleneck risk shifts from funding to execution: grid interconnects, data-center build times, and equipment lead times become the constraints that can turn into margin compression for vendors if spend outruns monetizable deployment.

The contrarian read is that “no bubble” is a credit observation, not an equity valuation call. Credit can remain rational well after equity expectations become stretched, so the mispricing—if any—is likely in the duration of the AI payoff, not the ability to borrow today. The catalyst that would reverse this setup is a 2-3 quarter lag between capex and monetization, which would first show up as tighter issuer guidance, then wider spreads in lower-tier AI-adjacent names, and finally selective capex pullbacks.

For timing, the setup is bullish over the next 3-6 months, but the risk window widens into 2026 if financing costs stay low while returns on deployed AI infrastructure stay opaque. That creates a classic late-cycle dispersion trade: the strongest credits keep funding, while weaker vendors and secondary beneficiaries face a cash-conversion test.

Key Decisions for Investors

• Long high-quality AI infra beneficiaries vs. weaker adjacencies: buy NVDA / MSFT / AMZN on pullbacks and avoid lower-quality AI hardware names with funding dependence; hold 3-6 months with upside from continued capex front-loading.
• Pair trade: long HYMB or LQD-quality credit exposure to top-rated AI issuers vs. short a basket of lower-rated tech/media credit proxies; thesis is spread compression at the top and widening at the margins over the next 6-9 months.
• Buy call spreads on AI power/grid beneficiaries such as VRT or ETN into weakness; risk/reward improves if capex shifts from compute to infrastructure bottlenecks over the next 1-2 quarters.
• Fade the most levered AI-adjacent equities on rallies if they trade on 2026 revenue assumptions; use out-of-the-money put spreads as 6-9 month downside protection against monetization slippage.
• Monitor new issue spreads for large-cap tech in the next 30-60 days: if primary prints stay tight, stay long the basket; if concession widens by >25-40 bps, reduce exposure to second-tier AI suppliers first.