Hyperscalers: AI Investment Will Outpace Near-Term Earnings

Investments in artificial intelligence will far outpace earnings growth through 2028 at four of the largest technology companies. This prospect exposes the so-called hyperscalers—Amazon AMZN, Meta Platforms META, Google parent company Alphabet GOOG/GOOGL, and Microsoft MSFT—to the risk that, in a downside scenario where returns on AI spending lag expectations, their leverage and cash generation could weaken. Still, the very large size and strong starting position of the hyperscalers leaves their credit profiles less vulnerable than those of smaller, more concentrated participants in the AI buildout.

Hyperscalers, Hyper-Capex

PitchBook’s consensus median forecasts that hyperscaler capex will continue to rise to fund AI infrastructure spending. While these investments contribute to EBITDA growth, PitchBook’s estimates suggest near-term earnings realization continues to lag capital deployment.

Our analysis of those estimates shows capex reaching 80%-100% or more of EBITDA through 2028, while incremental EBITDA relative to prior-year capex will generally remain in the 20%-40% range. This suggests hyperscalers will remain in an aggressive buildout phase, deploying capital ahead of the full realization of earnings from those investments.

The capital expenditures of the four hyperscalers have grown substantially, as discussed in our commentary Hyperscalers: AI Capex Weakens Credit Metrics, but Downside Risk is Limited, published Oct. 6.

Incremental EBITDA as a Gauge

This data suggests the unprecedented scale of current investment exceeds the near-term earnings contribution from those investments. We define incremental EBITDA yield as the year-over-year increase in EBITDA divided by the prior year’s capex. Thus, incremental EBITDA indicates how quickly incremental earnings are emerging relative to the scale of recent investment. Using this measure, a yield of 20%-40% means that for every dollar of prior-year capex, the increase in EBITDA in the following year is equivalent to approximately $0.20-$0.40. This suggests a meaningful portion of the investment case depends on earnings growth beyond the first year.

Given the scale of AI-related capex, the credit case for the AI ecosystem depends on continued growth in use, monetization, enterprise, and consumer adoption.

Effect on Credit Profiles

In our view, the key credit risk is a scenario wherein EBITDA growth slows or becomes negative while investment remains elevated. Since a portion of capex is being financed through debt issuance, weaker-than-expected earnings growth could translate into higher leverage, weaker free cash flow, and reduced financial flexibility. Under such a scenario, credit quality in the sector is likely to bifurcate. The large hyperscalers appear better positioned to absorb the pressure given their scale, cash flow generation, and financial flexibility, whereas smaller participants in the AI ecosystem may experience greater pressure on their leverage, free cash flow, and overall credit quality. (As we noted in our commentary AI Investment: How a Slowdown May Affect Tech Subsectors, published Oct. 16.)