Big Tech’s AI spending spree is about to hit the income statement. After pouring hundreds of billions into data centers and GPUs, companies like Meta $META ( ▼ 1.55% ) , Microsoft $MSFT ( ▼ 0.13% ) , Alphabet $GOOGL ( ▼ 1.06% ) , and Oracle $ORCL ( ▲ 2.34% ) are preparing to absorb a tidal wave of depreciation costs.
Morgan Stanley estimates those firms could recognize roughly $680 billion in depreciation over the next four years. That does not even include Amazon $AMZN ( ▼ 0.41% ) , another capex heavyweight.
Trading Opex for Capex
The AI boom has flipped the old Silicon Valley model on its head. For decades, tech companies bragged about being asset light. Their biggest costs were engineers, stock based compensation, and cloud compute.
Now they are swapping operating expenses for capital expenses. Instead of hiring thousands of engineers, they are buying tens of thousands of GPUs.
As one analyst put it bluntly, GPUs do not require stock compensation.
That shift is already visible. Amazon has announced tens of thousands of layoffs. Meta has trimmed headcount in divisions like Reality Labs. Analysts say companies are prioritizing AI projects while cutting or scaling back lower return initiatives.
Productivity Up, Headcount Risk Down
Executives are framing the transition as an efficiency revolution. Meta $META has highlighted major productivity gains from AI coding tools. Alphabet $GOOGL says about half of its code is now written by AI before being reviewed by humans. Microsoft $MSFT has repeatedly emphasized efficiency gains on recent earnings calls.
Sales per employee has surged across hyperscalers in recent years. That productivity boost has supported margins so far.
The question is whether it will be enough.
The Depreciation Wave Is Coming
AI infrastructure does not just cost money upfront. Those investments are depreciated over time, reducing reported earnings quarter after quarter.
As those costs ramp, profitability will depend on how effectively companies rein in non AI spending. That could mean fewer hires, tighter budgets, or slower growth in compensation.
Some analysts warn Wall Street may be assuming a free lunch. If depreciation outpaces efficiency gains, margins could face real pressure.
The Irony of the AI Boom
Previous tech booms created high paying jobs and expanded headcount. This one may be different. Instead of disrupting other industries, AI may increasingly disrupt tech workers themselves.
In the race between jobs and GPUs, capital appears to be winning.