It’s bubble or nothing for Google as search giant looks to plow ~$180B into datacenters this year

Steven

It’s bubble or nothing for Google as search giant looks to plow ~$180B into datacenters this year

Summary

Alphabet has raised its 2026 capital expenditure target to $175–$185 billion, roughly double last year, to support a major build‑out of datacentres and AI compute. About 60% of that spend (around $105–$111 billion) will go on fast‑depreciating kit such as servers — including Google’s TPUs and Nvidia GPUs — while the remaining ~40% (about $70–$74 billion) is earmarked for datacentre construction and networking.

The investment underpins both Google’s internal AI ambitions and Google Cloud, and also supports partners such as Apple, OpenAI and Anthropic. CFO Anat Ashkenazi said compute investment will be split evenly between internal workloads and the Google Cloud platform. Executives flagged constraints around power, land and supply chains as major operational risks.

Key Points

  • Alphabet plans $175–$185bn capex for 2026, ~double 2025 spend.
  • Approximately 60% of capex (~$105–$111bn) will buy servers and other fast‑depreciating compute; ~40% (~$70–$74bn) for datacentre facilities and networking.
  • Server purchases target AI infrastructure — Google TPUs and Nvidia GPUs — for both internal AI and Google Cloud customers.
  • Google Cloud Q4 revenue rose 47% year‑on‑year to $17.66bn, driven by enterprise AI demand.
  • Alphabet reported Q4 profit of $34.45bn and full‑year revenues above $402.8bn; ad revenue across Search, YouTube and Network grew strongly.
  • CEO Sundar Pichai highlighted scaling compute while managing power, land and supply‑chain limits as a top concern.
  • Gemini LLMs are being used to improve ad relevance and monetise more complex search intents and non‑English queries.

Context and relevance

This is a defining moment in the AI infrastructure arms race. Alphabet’s mega‑spend will ripple through chip suppliers, datacentre constructors, energy markets and cloud competitors. For vendors and investors, the move signals huge demand for GPUs, specialised accelerators and supporting supply chains. For enterprises, it means faster rollout of large‑scale AI services on cloud platforms, but also potential supply bottlenecks and geopolitical/energy implications to watch.

Author style

Punchy: Google is going all‑in. This isn’t a cautious upgrade — it’s an industrial‑scale bet on AI. If you track infrastructure, chips, cloud pricing or the economics of AI, the details matter. Read the call notes if you want to know who wins and who scrambles to keep up.

Why should I read this?

Quick and blunt: because Google spending nearly $180bn will change the market. If you work with cloud, buy servers, supply power, design datacentres or invest in AI hardware, this one spending decision affects capacity, prices and strategy for years. We’ve skimmed the numbers and pulled out the bits you actually need to know.

Source

Source: https://www.theregister.com/2026/02/05/alphabet_google_q4_2025/