Why Meta’s AI Expansion May Actually Strengthen Neoclouds

The idea that $META is an immediate threat to established neoclouds may be missing the bigger picture.
Morgan Stanley estimates that leasing just 250 MW of AI capacity to third parties could add roughly $3 to $META’s EPS by 2028 , an earnings boost of around 8% .
At 1,000 MW , that potential upside could rise to approximately $12 per share .
But the real advantage is timing.
The estimated cost of building 1 GW of AI compute capacity reportedly doubled from roughly $50 billion to $100 billion within only a few months. $META secured much of its capacity before that surge.
For companies entering the market today, the capital required to build comparable infrastructure has become enormous.
That creates a growing barrier to entry and benefits operators that already have capacity in place, including $NBIS and $CRWV.
The higher replacement cost becomes, the more valuable existing AI infrastructure may be. Bullish on established neoclouds.