Vertiv Holdings agreed to acquire an AI-focused power infrastructure company for $1.45 billion in cash, with the total deal value reaching $2.6 billion if the target hits undisclosed performance milestones. The $1.15 billion contingent piece—44% of the total consideration—is one of the larger earn-out structures in recent infrastructure M&A, reflecting both the urgency of the data center buildout and the pricing volatility in AI-adjacent assets.

The target, based in Raleigh, supplies power systems designed for hyperscale data centers, a segment that has seen capacity constraints tighten as generative AI workloads push energy density beyond legacy rack designs. Vertiv, headquartered in Westerville, Ohio, already provides thermal management and power distribution gear to Microsoft, Amazon Web Services, and CoreWeave. The acquisition closes a gap in its portfolio—direct AI-optimized power delivery—while keeping half the price tied to execution risk. The deal is expected to close in the second half of this year, subject to regulatory clearance.

The structure matters. Vertiv is paying a premium for entry into a segment where order backlogs have doubled in twelve months, but it is also signaling that current valuations require proof of concept. The $1.15 billion earn-out likely ties to revenue run-rate or margin targets over a 24-to-36-month window, standard in infrastructure transactions where customer concentration and supply chain bottlenecks create forecasting risk. If the milestones are met, Vertiv will have paid roughly 8x forward EBITDA based on sector comparables; if not, it walked in at 4.5x, a defensible multiple for a bolt-on with integration synergies.

For allocators, this is a barometer for how much froth remains in AI infrastructure. Vertiv is a $33 billion market cap company with institutional ownership north of 85%, and its willingness to structure half the deal as contingent consideration suggests that even strategic acquirers are pricing in downside scenarios. The target's client base is not disclosed, but the Raleigh location and power-systems focus point to proximity to both hyperscalers and the new wave of private GPU clusters. If the earn-out pays in full, it validates the thesis that AI data center power is a sustained bottleneck, not a transient constraint. If it does not, it will mark the sector's first major valuation correction.

Watch for three follow-on signals. First, Vertiv's next earnings call—likely in late April—will clarify whether the acquisition is being financed with balance sheet cash or new debt, which will indicate confidence in near-term cash generation. Second, any disclosed customer names or contract terms in regulatory filings will confirm whether this is a hyperscaler-backed asset or a speculative play. Third, comparable deals in the next ninety days will show whether this earn-out structure becomes the new norm or whether Vertiv simply priced in more risk than the market.

The 44% contingency is the data point. It is not caution. It is the market repricing certainty in real time.