Meta Pledges Nearly $700B in AI Compute, Faces Monetization and Timing Crisis

Stratechery · rss · 2026-08-03

Stratechery provides an in-depth analysis of Meta's recent earnings and its AI infrastructure timing and financial dilemmas. Although Meta's revenue growth remains solid, expenses surged 55% against a 28% revenue increase, with significant capex yet to depreciate, dropping free cash flow to multi-year lows.

Infrastructure Timing Mismatch

To avoid falling behind in the AI race, Meta has had to heavily rent compute from third parties while building its own data centers, which take years. This "double-paying" for infrastructure has raised investor concerns. The company has currently committed almost $700 billion in future spending related to AI data centers and cloud computing.

Monetization & Strategic Defense

Zuckerberg hinted on the earnings call that Meta might explore a cloud computing business to sell its compute power at a premium. Furthermore, he firmly rejected relying on third-party open-source models, emphasizing the existential necessity of developing their own frontier models.

Original post →

More from Venture

Venture channel →