The AI race has shifted into top gear. Q1 results from the four hyperscalers laid bare just how aggressive the build-out has become. The wave is funded by frontier-model-scale data centres designed for LLM training and inference at unprecedented scale.
The Engine Behind the Spend: Data Centres
Most of the capex flows into AI data centres. Training and serving large language models demands extraordinary compute. Meta CEO Mark Zuckerberg: "Every signal we see, internally and across the industry, is increasing our confidence in this investment."
Meta lifted its 2026 spend guidance from $115–135B to $125–145B. Google now expects to spend at least $180B. Combined, the four companies are on track to spend roughly $700B in 2026.
What's Next — and the Open Question
The astronomical spend marks AI's move out of the experimental sandbox and into the centre of corporate identity. Investors are still waiting on the critical question: will all of this actually pay off?
Cite this: AI Mevzuları · May 4, 2026 · aimevzulari.com/haberler/yapay-zeka-harcamalari-rekor-130-milyar-dolar