2 Data Center Stocks That Could Help Make You a Fortune

Morgan Stanley projects capital spending by the five largest U.S. hyperscalers will reach $1.4 trillion by 2028, supporting a bullish outlook for AI data-center infrastructure. Iren, valued at roughly $19 billion, targets annualized revenue of $4 billion by end-2026 from $1 billion currently and has delivered the first 50MW unit under its five-year, $9.7 billion Microsoft contract; analysts project $10.7 billion of adjusted operating profit by fiscal 2029. Vertiv reported 24% year-over-year quarterly revenue growth and raised its long-term revenue-growth target to 20%-22% annually through 2030, while analysts forecast 38% annualized earnings growth.

Analysis

IREN is best viewed as a leveraged monetization of scarce energized land and contracted GPU capacity, not yet a software-like cloud business. Its upside depends on converting power inventory into signed, creditworthy take-or-pay contracts faster than equity dilution and construction financing increase enterprise value; a single large customer concentration can make reported backlog less valuable than its headline duration implies. The key 1-3 month catalyst is evidence that follow-on capacity is contracted before commissioning, while the 6-18 month risk is that hyperscalers internalize builds or negotiate lower returns once power availability improves.

VRT has the cleaner earnings-through-cycle setup because rising rack density expands content per megawatt across cooling, power distribution and service, with less direct exposure to GPU-utilization economics. However, the valuation leaves little room for ordinary execution: any deceleration in orders, backlog conversion, or gross-margin progression could drive material multiple compression even if revenue remains healthy. Eaton (ETN), nVent (NVT), and Schneider Electric (SU.PA) are plausible second-order beneficiaries, while VRT's premium rests on maintaining a faster growth and margin trajectory than these broader electrification peers.

Consensus appears to underweight the bottleneck shift from chips to grid interconnection, transformers, switchgear and liquid-cooling installation labor. This favors equipment suppliers with qualified manufacturing capacity and service networks over marginal data-center developers. Conversely, AI-capex headlines are becoming less incremental for VRT after its rerating; the more asymmetric opportunity may be selective suppliers where power-equipment backlog has not yet been capitalized into a similar multiple.

Key Decisions for Investors

• Maintain VRT as a core AI-infrastructure long only on pullbacks or after quarterly order/backlog confirmation; target a 12-18 month hold, but reduce if book-to-bill falls below 1.0x or management retreats from margin-expansion targets. Upside is continued high-density content growth; downside is substantial multiple compression from an already elevated earnings multiple.
• Express the power-bottleneck theme via a diversified long basket of ETN and NVT versus a modest short in a broad data-center-development proxy where available; hold 6-12 months. This captures electrical-distribution and thermal-content growth while reducing exposure to a single developer's financing and customer-concentration risk.
• Treat IREN as a catalyst-driven, high-volatility position rather than a fundamental compounder until contracted capacity, customer concentration, capex per megawatt, and funding sources are independently reconciled. Consider a small long only following signed capacity commitments and financing disclosure; invalidate on material equity issuance, delayed deployments, or contract repricing.
• For VRT, monitor hyperscaler capex commentary and lead times for switchgear, transformers, and liquid-cooling systems during the next two earnings cycles. A broad capex pause would likely hit developers first, but a sustained reduction in equipment lead times would be the earlier signal that VRT's scarcity premium is fading.