Zuckerberg Pushes Back on "Compute Glut" Fears — What's Meta Really Signaling?

$Meta Platforms (META.US)$ has announced plans to invest roughly $10 billion in a new 1GW supercomputing data center in Alberta, Canada — set to become the company's largest compute hub outside the U.S. once completed. Management also reaffirmed an even bolder target: doubling total data center capacity to 14GW by 2027. The announcement lands as a direct counter to growing market concerns about AI compute oversupply. Zuckerberg made clear that Meta's appetite for compute is far from satisfied — the company intends to secure as much capacity as it possibly can. Yet in the same breath, even as he acknowledged that compute supply remains tight relative to demand, he floated a seemingly contradictory idea: whether leasing out portions of Meta's AI infrastructure to external customers could unlock additional value. "We can't get enough compute" and "should we rent some out" sitting side by side captures the tension Big Tech is navigating as AI infrastructure spending keeps escalating. Does this arms race eventually tip into oversupply, or does structural scarcity persist for years to come? Where do you see the AI infrastructure supply chain heading next? Flexibly allocate across core U.S. tech equities and diversified RWA portfolios on Bitget — achieve efficient cross-asset allocation and precision risk hedging: https://www.bitget.com/promotion/futures-rwa