Google's $175 Billion AI Bet: What Happens to the Humans in the Way

The Numbers

Google's planned AI capital expenditure for 2026: $175-185 billion. For context, that's more than the GDP of 130 countries. It's roughly the entire annual revenue of Alphabet being reinvested into AI infrastructure.

Meanwhile, CBO Philipp Schindler sent a memo offering voluntary exit packages to GBO staff who aren't 'all in' on AI.

Connect the Dots

When a company announces it will spend $175 billion on machines, the budget has to come from somewhere. Google's operating margins are healthy but not infinite. The equation is brutal: AI infrastructure up, human headcount down.

The Voluntary Playbook

Google has perfected the 'voluntary' exit strategy:

• June 2025: Knowledge and Information teams • October 2025: YouTube reorganisation • February 2026: GBO — sales, solutions, corporate development

Each round targets different teams, but the pattern is identical. Offer packages, wait 30 days, then restructure whoever remains. It avoids WARN Act notices, reduces lawsuit risk, and looks better in the press than 'mass layoffs.'

Who's Next?

If you work in a role that: - Involves repetitive decision-making - Relies on information Google's AI can now summarise - Sits between a customer and a product (sales, support, account management) - Doesn't directly write code or build AI systems

...you should be paying attention.

The Industry Ripple

Google sets the template for Big Tech. Meta's 'year of efficiency' in 2023 was copied by every major tech company within 12 months. Google's 'voluntary AI alignment' exit programme will follow the same pattern. Expect Microsoft, Amazon, and Meta to announce similar programmes by Q3 2026.

What $175 Billion Buys

Data centres. GPUs. Custom TPUs. Cooling infrastructure. Power contracts. Model training compute. And a very clear signal that Google's future workforce is silicon, not carbon.

The Bottom Line

Google isn't laying people off because it's struggling. It's laying people off because it's transforming. And in transformation, the humans who built the old thing rarely get to build the new one.