Two of the world’s largest technology companies went to their investors in the same week and said, essentially: we are spending whatever it takes.
Meta has guided for $115B–$135B in 2026 capital expenditure, up from $72.2B in 2025, according to reporting from Investing.com. Alphabet has guided for approximately $175B or more in 2026 capital expenditure, a significant increase over 2025 levels, corroborated by multiple T3 sources. The combined midpoint of these two guidance ranges exceeds $300B. Neither figure is a reported actual, both are forward-looking guidance disclosures made during earnings calls, and both should be read as the companies’ stated intentions, not confirmed spending.
The market’s response to Alphabet’s announcement was unambiguous. Alphabet shares declined approximately 7% following its earnings disclosure, a reaction that captures a growing investor unease: the spending is accelerating, but the revenue case for returns at this scale remains contested. Meta’s share reaction following its own earnings disclosure was reported positively by some sources but could not be independently confirmed in available source materials and is not stated here as fact.
Why this matters for markets. The Alphabet share decline is the more informative signal. Investors have watched hyperscalers expand AI infrastructure spending for three consecutive years. The thesis, that frontier AI capability translates to monetizable cloud services, advertising intelligence, and enterprise revenue, has not yet produced returns proportionate to the investment cycle. A 7% decline on an infrastructure announcement from a company the size of Alphabet reflects genuine skepticism about when, and whether, the revenue inflection arrives.
That skepticism doesn’t mean the spending stops. Meta’s guidance represents roughly a 60% increase from 2025 levels at the midpoint of its range. Companies at this scale don’t announce capex guidance of this magnitude unless the internal investment committee has already approved it. These are commitments, not aspirations.
Context. This isn’t an isolated event. Hyperscalers have been building toward this spending posture for several years, absorbing compute demand from AI model training and inference workloads that have grown faster than most forecasts anticipated. The infrastructure being built now, data centers, custom silicon, networking, has a 3–5 year useful life. Hyperscalers are locking in physical capacity for AI workloads that don’t fully exist yet at commercial scale.
What to watch. The key question over the next two quarters is whether Alphabet’s revenue guidance catches up to its capex guidance. If it does, the 7% sell-off looks like an overreaction and signals a buying opportunity. If it doesn’t, other hyperscalers face the same investor pressure. Watch Alphabet’s Q2 2026 earnings for cloud segment revenue and AI-attributed advertising growth. Watch Meta’s 2026 actuals against the $115B–$135B range, any downward revision signals demand softness in AI services that the guidance assumed.
TJS synthesis. When two companies simultaneously announce capital commitments of this scale, the story isn’t the numbers. It’s the implied assumption behind them: that AI infrastructure demand will grow fast enough to justify returns on $300B+ in a single calendar year. The Alphabet market reaction suggests investors aren’t yet ready to price that assumption as certain. Companies building on hyperscaler infrastructure, or competing against it, should treat that investor skepticism as a leading indicator worth monitoring.