← Back to timeline
AMDEarnings2026-08-05
Su describes how the shift of corporate AI projects from trials into everyday operation is changing what kind of computing hardware customers ask for.
Loading moment…
Lisa Su · CNBC Television · AMD CEO Lisa Su on Q2 results: We're in an extraordinary time for AI compute right now
“SO WHAT WE'RE FINDING IS AS AI IS BECOMING MORE USEFUL FOR PEOPLE, AS WE'RE USING MORE AGENTS, AS ENTERPRISES ARE USING MORE OF THEIR WORKLOADS FROM, LET'S CALL IT, WHAT WAS EXPERIMENTATION INTO NOW PRODUCTION WORKLOADS, PEOPLE NEED A LOT MORE CPU COMPUTE AS WELL AS GPU COMPUTE”
Key terms
- agents
- — AI software that carries out multi-step tasks on a user's behalf rather than just answering a single question.
- production workloads
- — Software running as part of a company's day-to-day operations, as opposed to a test or pilot project.
- CPU compute
- — Processing done by general-purpose central processors, the main chips that run most software.
- GPU compute
- — Processing done by graphics-style chips, widely used to train and run AI models.
Educational summary of a public statement. Not investment advice.