A Bloomberg Technology discussion makes a simple point: as AI starts doing things instead of just answering questions, the company that controls the setup may be the one that makes the money. Enterprise leaders say control over data, infrastructure and workloads could decide who captures the value of what is being called agentic AI.
What changed
Until recently, most AI tools wrote text or answered questions. Now AI is shifting toward carrying out tasks on its own. These are often called AI agents. Instead of suggesting a reply, an agent might send it. That change raises the stakes, because the AI is no longer just advising a person. It is acting.
Why control matters
The leaders in the discussion focus on three things. Data is the information the AI reads and uses. Infrastructure is the computing systems it runs on. Workloads are the jobs it is given to do. Whoever holds these has a strong say in how the AI is used and who benefits.
Who it affects
Any business thinking about letting AI act on its behalf is touched by this. The open question in the discussion is who will capture the money that AI agents create: the businesses using them or the companies supplying them.
What is not known yet
The material available does not name the speakers or companies, and it gives no figures or dates. It also does not say which products or contracts are involved. Those details are not known yet.
What to do
- Ask who holds your data before you sign up.
- Ask where the AI work actually runs.
- Find out whether you can move your data and tasks elsewhere later.
The takeawayBefore letting AI act for your business, ask who holds your data and where the work runs.