Last year was a spending spree for the ages. Amazon $AMZN ( ▼ 1.56% ) , Alphabet $GOOGL ( ▼ 2.08% ) ,Microsoft $MSFT ( ▲ 1.42% ) , and Meta $META ( ▲ 1.46% ) collectively poured nearly $400 billion into capital expenditures, excluding leases. Most of that cash went straight into building the AI infrastructure they believe will define the next era of tech.
And somehow, that was just the warm-up.
The AI Arms Race Gets Expensive
If you thought 2025 looked wild, 2026 is shaping up to be even bigger. The same four tech giants are expected to boost their combined capex by roughly 50% next year, pushing total spending toward a staggering $600 billion.
Amazon alone says it is on track to shell out about $200 billion. Alphabet is guiding for somewhere between $175 billion and $185 billion. Meta is projecting a range of $115 billion to $135 billion. Microsoft did not offer formal guidance, but analysts estimate its calendar 2026 capex will land near $114 billion.
In other words, each company is now operating on infrastructure budgets that rival the GDP of small countries.
Why They’re Spending Like This
This is not about fancy office buildings or new headquarters. The bulk of this money is being funneled into data centers, advanced chips, networking gear, and the power capacity needed to run increasingly massive AI models.
Training and running AI systems at scale requires enormous computing clusters, specialized semiconductors, and energy-hungry facilities. These companies are racing to make sure they have enough digital horsepower to support everything from AI-powered search and copilots to ad systems and cloud services.
Fall behind on infrastructure, and you risk falling behind in AI. No one in this group wants to be second.
Wall Street Keeps Underestimating the Bill
One interesting twist: analyst forecasts for 2026 spending have already proven too conservative for most of these companies. Earlier projections for Amazon $AMZN ( ▼ 1.56% ) , Alphabet $GOOGL ( ▼ 2.08% ) , and Meta $META ( ▲ 1.46% ) were significantly lower than what management teams are now signaling.
That suggests the AI buildout is accelerating faster than even Wall Street expected. If these trends continue, capex estimates may keep getting revised higher as the year unfolds.
Big Tech is not just dabbling in AI. It is committing hundreds of billions of dollars to make sure it owns the pipes, the chips, and the power behind the future of computing.