Alphabet in $700B AI Capex Buildout as Engineers Depart to Anthropic
GOOG•Wedbush forecasts that Alphabet will join MSFT, Meta and Nvidia in investing $700 billion in AI infrastructure this year, entering a six-to-12 month “air pocket” stage before monetization drives revenue. Alphabet lost core engineers to Anthropic in recent weeks, adding to short-term selling pressure despite long-term AI growth prospects.
1. Wedbush’s AI Infrastructure Thesis
Wedbush projects that Alphabet will join peers in spending about $700 billion on AI infrastructure in 2026, placing the company in an “air pocket” phase where data center buildouts ramp but revenue benefits lag. The firm characterizes recent selling pressure as short‐term impatience rather than a breakdown in the AI trade.
2. Engineer Exodus and Talent Risks
Alphabet has seen several core engineers depart to Anthropic in recent weeks, eroding investor confidence even as the company accelerates AI‐related hiring and development. Wedbush warns that talent losses could heighten near‐term volatility but believes Alphabet’s AI strategy remains intact.
3. Compute Costs and Long‐Term Outlook
The broker also highlights concerns over rising compute and memory costs, suggesting enterprise buildouts could slow if expenses stay elevated. With hardware price pressures expected to ease over the next year, Wedbush frames current weakness as a buying opportunity in the early years of a 10‐year AI buildout.
