Google can’t stop spending on AI, even if investors want it to

I think investors are too quick to assume Google can fix its AI spending problem by simply cutting capex.

Look at the numbers. In Q3, Google Search revenue grew 12% to $63.4 billion. Google Cloud grew 103% to $30.7 billion. Cloud is growing much faster, but Search is still twice the size. And Search is the business Google cannot afford to damage.

There is a reason to worry about the economics of AI answers. A 2026 field experiment found that when AI Overviews appeared, users clicked through to outside websites 39.8% less often, and zero-click searches increased 34.5%. That doesn’t prove Google is earning less from every search, but it shows how much the experience is changing.

Meanwhile, Google spent $195 billion to $205 billion on capex this year, expects spending to rise significantly in 2027, and reported negative $9.1 billion in free cash flow for Q3.

This is the part I don’t think the usual bull or bear argument answers. If Google cuts spending, it risks falling behind. If it keeps spending, it has to prove AI answers and Cloud can make enough money to pay for it.

Cloud is showing real growth. But can it make up for even a small loss in Search profitability? I wouldn’t assume that.

Spend too much: investors complain.

Don’t spend enough: risk losing Search.

And even if Google wins the AI race, it still has to make money from it.