Which business will "win" AI?

Key Shift in How AI Will Be Used

The subtitles describe a shift from traditional “prompt-and-wait” chat models—where users must actively invoke the system—to AI agents that proactively operate across everyday devices and environments.

Main Technological Concepts / Product Strategy

• AI agents everywhere (“device surfaces”) AI will increasingly be delivered through an agent embedded in common consumer and physical devices, such as:

  • Cars
  • Phones
  • Glasses / earbuds
  • Robots

• Edge computing as a requirement For an agent that can see, hear, think, and make decisions, relying only on cloud data centers is not enough. The emphasis is on on-device compute at the edge, where data is created and decisions are made.

• Edge-to-cloud infrastructure approach The focus is not limited to AI chips. Instead, it’s framed as a full platform strategy enabling AI execution across devices.

Specific Qualcomm-Related Initiatives Mentioned

• Dragonfly data center roadmap, including:
  • C1000 CPU for AI agentic workloads
  • New AI inference accelerators
  • High-bandwidth compute
• Planned acquisition of Modular
• Overall positioning: building the compute + connectivity needed to run AI across devices, rather than depending entirely on centralized cloud.

Competitive “Winner” Framing

Instead of asking “which AI app will win?”, the framing becomes:

• Who builds the infrastructure that allows AI to run efficiently everywhere, not just in the chat window.

The claim is that it may not be a single company with the best model, but rather the company that can run AI efficiently across many device types.

Reviews / Guides / Tutorials

• None indicated. This appears to be an analysis/pitch of infrastructure strategy (with a “check out more” link referenced in the description).

Main Speakers / Sources

• The video narrator/speaker (not named in the subtitles)
• Qualcomm (explicitly referenced as the partner and source of the roadmap/strategy)