1. News
  2. AI models & agents
  3. AI · Business costs
AI · Business costs2 min read

AI Giants Face a Big Question: How Will They Pay for Their Data Centers?

A Bloomberg columnist says AI leaders have weak answers on how they will pay for giant data centers, which could affect the AI tools your business relies on.

By the Dashline deskSource: Bloomberg Technology
AI Giants Face a Big Question: How Will They Pay for Their Data Centers?
Photo: Victorgrigas, CC BY-SA 3.0

A Bloomberg Opinion columnist says the companies behind popular AI tools have not given a strong answer to a simple question: how will they pay for it all? Parmy Olson says the explanations she hears from Silicon Valley executives are weak.

What is going on

Anthropic and OpenAI are building huge computer warehouses, often called data centers. These buildings hold the machines that run AI tools, and they cost a great deal to operate. Olson points out that the productivity gains from AI have so far been small. That means the money businesses save or earn from using AI has not yet matched the size of the spending.

Who it affects

Anyone who uses AI tools for work could be touched by this. That includes small businesses that use AI for writing, customer service or research. If AI firms struggle to earn enough, prices or terms for those tools could change.

What is not known yet

The material available is a short summary of Olson's views, not the full column. It does not say how much these companies are spending, how much they earn, or what the executives actually told her. It also does not say that any price change is planned. This is one columnist's opinion, and the questions she raises remain open.

What to do

  • Avoid locking in too deeply to one AI provider.
  • Keep your data and key work in formats you can move.
  • Review what you pay for AI tools and what they save you.
The takeaway

AI firms have big bills and unclear answers, so avoid tying your business too tightly to any one AI tool.

More articles

You should really read these.

Next step

Let’s talk