DataBank’s Incoming CEO on AI Buildout, Nvidia Earnings

DataBank’s incoming CEO Kevin Ooley said the firm operates 70+ data centers across 25+ markets, with expansion constrained primarily by power availability and broader infrastructure limits. The discussion frames near-term capacity risk tied to utilities/energy constraints rather than a specific earnings or policy trigger. Overall, it’s a cautious operational outlook but without quantified financial impact in the article.

Analysis

The economic center of gravity is shifting from demand generation to power interconnectivity, which tends to re-rate the boring parts of the stack: switchgear, thermal management, EPC, and grid equipment. That is a better setup for names like VRT, ETN, and PWR than for pure-play capacity owners whose growth is increasingly gated by permitting and utility timelines. Second-order, the scarce asset is not the building but the energized shell, so operators with already-secured megawatts should take share from smaller entrants that can’t fund or permit the same footprint.

Near term, the biggest market risk is that capex gets booked before revenue shows up. If power delivery slips by quarters, lease-up assumptions and backlog conversion for data-center landlords and hyperscale suppliers can disappoint even if end demand stays strong, which argues for lower tolerance on long-duration multiples over the next 1-3 months. The bottleneck also raises working-capital intensity, so balance sheets with high leverage or heavy development pipelines are the most exposed to a funding squeeze if rates stay high.

Contrarian view: consensus may be overestimating how quickly AI demand translates into monetizable capacity. Power scarcity can actually be bullish for incumbents with inventory, but it is bearish for the pace of industry revenue recognition; the winner is the one with the cleanest queue position, not the loudest growth narrative. The thesis breaks if utilities accelerate interconnects, modular generation comes online faster than expected, or if data-center orders soften enough to relieve pressure on equipment backlogs.

Key Decisions for Investors

• Long VRT on pullbacks for a 1-3 month trade; the catalyst is backlog re-acceleration and margin leverage from higher power density. Risk/reward is attractive if the market is still pricing in linear AI capex growth, but the thesis is invalidated by any order slowdown or guide-down next quarter.
• Long ETN as a six- to twelve-month structural winner from the grid and data-center power buildout. Use it as the cleaner expression of the bottleneck theme; exit if backlog growth decelerates faster than revenue or if distributor inventories normalize sharply.
• Long PWR as a second-order beneficiary of interconnect, substations, and utility buildout. This is a slower-burn idea, better for a basket than a catalyst trade, and it fails if permitting or project awards slip materially over the next two earnings cycles.
• Do not short DLR or EQIX blindly; the power constraint can be supportive for landlords that already control energized inventory. Treat any weakness there as a watch item, and only press a bearish view if reported power-secured capacity or leasing spreads roll over.