2 weeks ago
AI Boom Spending Hits $165 Billion as Cash Flow Shrinks
The biggest technology companies are spending a huge amount of money to build artificial intelligence, or AI.
Together they have spent about $165 billion on computers, chips, and giant data centres.
This spending uses up so much cash that the money left over, called free cash flow, has dropped a lot.
Last year the leftover money was $60 billion, and now it is only $7 billion.
A company called Jefferies thinks that number could go below zero soon, meaning the companies would spend more money than they bring in.
The computers they buy also lose value over time, which is called depreciation, and that cost is rising too.
But there is good news as well.
Their cloud business, where other companies rent their computers, grew by 38% to $126 billion.
The companies that make the chips and machines are making money right now, while the big tech companies have to wait longer for their payoff.
In the end, everyone is watching to see if all this AI spending will be worth it.
Big Tech's spending on AI infrastructure reached $165 billion, while aggregate free cash flow fell to $7 billion in Q2 2026 from $60 billion in Q4 2025.
Jefferies forecasts aggregate free cash flow will turn negative, at -$12 billion, in Q3 2026.
Combined depreciation expenses for Microsoft, Amazon, Alphabet and Meta reached $44.5 billion in Q2 2026, up 24% year-on-year.
Cloud revenues for Microsoft, Google and Amazon rose 38% year-on-year to $126 billion in Q2 2026.
Jefferies says the AI capex race remains earnings-accretive for infrastructure suppliers and expects 48% annualised EPS growth for the S&P 500 AI basket during 2026-27.
- Who
- Big Tech companies Microsoft, Amazon, Alphabet and Meta, along with analysts at Jefferies
- What
- Record capital spending on AI infrastructure is shrinking Big Tech's free cash flow and raising depreciation costs, even as cloud revenues grow
- Where
- Not explicitly stated in the articles
- When
- Q2 2026, with Jefferies forecasting a further decline in Q3 2026
- Why
- Because building the data centres and computing equipment needed to power the AI economy consumes enormous amounts of cash
Skeptical View
Optimistic View
Will AI infrastructure spending pay off?
Skeptical View
Enormous capital spending has shrunk free cash flow to $7 billion and is forecast to go negative at -$12 billion in Q3 2026, and with depreciation rising, the eventual revenue will need to be substantial to justify the investment.
Optimistic View
Cloud revenues for Microsoft, Google and Amazon grew 38% year-on-year to $126 billion, and Jefferies says the AI capex race remains strongly earnings-accretive, with the S&P 500 AI basket poised for 48% annualised EPS growth in 2026-27.
Is shrinking free cash flow a warning sign?
Skeptical View
Free cash flow falling from $60 billion to $7 billion, with a negative quarter expected next, signals the spending is becoming harder for Big Tech to sustain.
Optimistic View
The cash-flow decline is not necessarily a sign the AI boom is failing; it reflects heavy strategic investment, and companies supplying the infrastructure are already booking revenue and profits.
Key facts
- Big Tech capital spending
- $165 billion
- Aggregate free cash flow, Q2 2026
- $7 billion (down from $60 billion in Q4 2025)
- Forecast free cash flow, Q3 2026
- -$12 billion (Jefferies estimate)
- Combined depreciation expenses, Q2 2026
- $44.5 billion, up 24% year-on-year
- Cloud revenues, Q2 2026
- $126 billion for Microsoft, Google and Amazon, up 38% year-on-year
- S&P 500 AI basket EPS growth forecast
- 48% annualised in 2026-27, vs 23% for the broader S&P 500 and 12% excluding AI
- Companies covered
- Microsoft, Amazon, Alphabet (Google) and Meta










