Every hyperscaler raised capex guidance in the same quarter
Aggregate sector capital intensity is now the dominant variable in the models of every large compute buyer. The question for each operator is not whether to spend but whether the disclosure will ever permit that spend to be evaluated.
Synchronized guidance raises across operators remove the usual explanation of company-specific timing and establish a sector-wide capital cycle. Where disclosure is absent, companies have declined to provide it, and in at least one case explicitly resisted a quarterly disclosure cadence for a revenue metric, which is itself informative.
- Alphabet capex outlook raised to approximately $185 billion, up roughly 32 percent
- Meta guidance raised to $118-126 billion with capital intensity elevated through 2028
- Microsoft quarterly capital expenditure including finance leases of $24.2 billion with outlook raised
- Alphabet disclosed that depreciation will grow materially faster than revenue for several quarters
- Microsoft disclosed approximately $88 billion in multi-year datacenter commitments
All Signals (1)
Related Themes
Cluster Debate
Three of the four largest compute buyers raised capital spending in the same quarter: Alphabet's outlook to approximately $185 billion, Meta's to $118-126 billion, and Microsoft's disclosed quarterly spend of $24.2 billion with its forecast raised. Synchronized commitments across independent buyers are the strongest demand signal the sector produces.
- Alphabet raised its capex outlook to approximately $185 billion, up roughly 32 percent year over year
- Meta raised guidance to $118-126 billion and stated capital intensity remains elevated through 2028
- Microsoft reported quarterly capital expenditure including finance leases of $24.2 billion with the outlook raised
- Alphabet disclosed that depreciation will grow materially faster than revenue for several quarters
Three of the four largest compute buyers raised spending in the same quarter, and they compete for the same power, the same contractors, and the same customers. That is not independent confirmation, that is a queue, and everyone in it is hoping the person behind them blinks first.
- Guidance raised simultaneously under identical multi-year and structural framing
- Alphabet explicitly forecast depreciation to outpace revenue for several quarters
- Meta pre-announced margin compression while stating capital intensity only through 2028
- Microsoft attributed its capex increase to capacity delivery timing rather than to demand
- Aggregate sector capital intensity is now the dominant financial variable for every operator
- The commitments extend beyond the companies' own forecasting horizons
- Capacity delivery, not demand, is the stated reason for the increases
Independent buyers with different demand profiles reaching the same conclusion in the same quarter is stronger evidence than any single raise. The alternative requires three boards to have made the same error.
The same consultants, the same three customers, and the same board dynamics produced all three raises. Synchronization is what you get when nobody wants to be the one who stopped building.
- Alphabet capex outlook approximately $185 billion, up roughly 32 percent
- Meta guidance $118-126 billion with capital intensity elevated through 2028
Disclosing that depreciation will outpace revenue removes the surprise and lets the market price the investment period properly. That is high-quality disclosure of unwelcome news.
It prepares the market for a number they cannot avoid and was paired with no monetization milestone. Pre-announced pain is still pain; the announcement only decides in advance who gets blamed.
- Depreciation stated to grow materially faster than revenue for several quarters
- Meta stated depreciation and infrastructure costs would compress operating margin
The raise is real under both readings; the dispute is whether the same decision made three times is three data points or one. The analyst is right that the scale of the commitment is only rational against visible demand, and the gossip voice is right that the shared language is at least partly a pose. The depreciation schedule would settle it: if it lands on the disclosed timeline and revenue grows through it, synchronization was conviction; if the 'through 2028' language begins to move next quarter, it was a queue.