Meta’s AI data center cost went from $10 billion to $50 billion in under 2 years—and split the town in two

Meta will expand its Hyperion AI data center in Richland Parish to a 5-gigawatt facility costing $50B+ (from a $10B start in 2024), with power provided via Entergy’s 10 new plants and 240 miles of transmission lines funded by Meta. While Louisiana secured $150B+ in new investment and Meta reports $1.6B+ already contracted with local companies, consumer groups are pressing regulators over financing/utility-rate impacts, arguing the deal could increase costs for ratepayers. Sentiment is split: local business growth and teacher bonus payouts (school bonuses topping $50k) are tempered by concerns over evictions, rent/traffic strain, and affordability of utility bills.

Analysis

The real market signal is that AI compute is moving from a software-style growth story to an industrial capital-allocation arms race. That is negative for META’s near-term free cash flow and multiple if investors start discounting a longer payback period for incremental AI capex, and it raises the bar for MSFT, GOOGL, and AMZN to keep pace or risk strategic slippage. Over the next 3-12 months, the first derivative is not model quality but who can source power, land, and financing cheapest; that favors vertically integrated or utility-backed ecosystems over pure software narratives.

ETR is a mixed winner: a larger rate base and transmission buildout can be accretive, but the political optics are deteriorating fast. The more the project becomes a public-policy issue, the greater the risk that cost recovery gets delayed, capped, or reallocated, which is a classic utility multiple overhang. OWL benefits from the growing appetite for off-balance-sheet infrastructure finance, but that same structure invites scrutiny if regulators conclude hyperscalers are externalizing too much grid risk; that is a 6-18 month reputational and policy risk, not an immediate earnings issue.

Contrarian view: the market may be underestimating how replicable this financing model becomes if the structure works cleanly for Meta, which would be bullish for asset-light capital providers and bearish for traditional balance-sheet constraints. The bigger overdone move may be in assuming local tax concessions materially change project economics; the economic fight is really over power pricing and depreciation of infrastructure, not sales tax. What would falsify the bearish read on META is evidence that capex intensity stabilizes and AI monetization ramps faster than depreciation; what would falsify the bullish read on ETR is a regulatory decision that fully shields ratepayers and validates recovery.

Key Decisions for Investors

• Fade META on strength with a 3-6 month bearish put spread; thesis is continued capex inflation compressing FCF conversion and keeping the stock on the defensive until the next earnings cycle.
• Avoid chasing ETR until Louisiana rulemaking and utility filings clarify cost recovery; if the stock rallies on the buildout, consider a tactical short vs. XLU as a political/regulatory overhang trade.
• Use OWL as the cleaner expression of infrastructure-finance growth, but size modestly; buy pullbacks only if regulators continue to tolerate asset sell-down structures. Falsifier: a formal probe into hyperscaler financing structures.
• Relative-value: overweight MSFT/GOOGL over META for 1-3 months if you want AI exposure, on the view that diversified cloud businesses absorb capex pressure better than a single mega-project headline.