AI Devices Are Coming | Qualcomm CEO Cristiano Amon

Key Concepts

• System on Chip (SoC): An integrated circuit that combines multiple components (CPU, GPU, NPU) onto a single chip.
• Agentic AI: AI systems that act as autonomous assistants, capable of performing tasks and operating applications on behalf of the user.
• Edge vs. Cloud Processing: The distribution of AI workloads between local hardware (Edge) for speed/privacy and remote servers (Cloud) for heavy computation.
• Compute Continuum: The integrated system where Edge and Cloud devices work together seamlessly.
• Model Context Protocol (MCP): An open-source system allowing AI assistants to communicate with external data and applications.
• Orchestrators: Software layers that manage AI agents, allowing them to interact with OS functions and third-party apps.
• Multimodal Models: AI models capable of processing various inputs, including audio, text, and vision.

---

1. The Evolution of AI Devices

Cristiano Amon, CEO of Qualcomm, emphasizes that we are entering a new era of "Personal AI Devices." While the smartphone remains the "sun" of our digital lives, its role is shifting.

• Dual Personas: Smartphones will continue to function as traditional devices, but will increasingly run "orchestrators" that allow AI agents to operate the phone autonomously.
• New Form Factors: Future devices will move closer to human senses. This includes smart glasses, jewelry, pendants, and pins. These devices are designed to be "always-on" and context-aware.
• The "Agentic" Shift: By 2026, Amon predicts agents will reach scale. Unlike current chatbots that require manual input, these agents will proactively manage calendars, emails, and app interactions.

2. The Changing Role of Apps

Amon argues that the traditional "App Store" model is evolving.

• From Apps to Agents: Instead of a user manually navigating a banking app, an agent will access the app’s data via credentials to provide specific answers or perform actions (e.g., "Pay this bill") upon request.
• Horizontal Ecosystems: As agents become the center of the digital experience, the industry will shift from "vertical" ecosystems (where one company makes the phone and all wearables) to "horizontal" ones, where multiple devices connect to a central AI agent.

3. Technical Architecture and Silicon Design

Qualcomm is redesigning its chip architecture to meet the demands of AI:

• NPU (Neural Processing Unit): A critical component of the SoC dedicated to running AI models locally on the device.
• CPU Renaissance: The CPU is becoming increasingly vital as the "orchestrator" that manages complex tasks, data retrieval, and app communication.
• Efficiency Constraints: Because mobile devices lack active cooling (like radiators), Qualcomm’s engineering focus is on packing maximum compute power into a highly energy-efficient, heat-managed footprint.

4. Supply Chain and Market Dynamics

• Memory Bottlenecks: A significant shortage of memory exists as capacity is currently funneled toward AI data centers. Amon expects this to ease by late 2027 or 2028 as the industry builds more capacity.
• New Entrants: Non-traditional consumer electronics companies (e.g., OpenAI) are entering the hardware space because they recognize that "winning the endpoint" (the device) is essential to accessing the massive amounts of user data needed to train future models.
• 6G and Future Connectivity: Qualcomm is already developing 6G technology, which is intended to be the wireless standard for the "Age of AI."

5. Notable Quotes

• "2026 is going to be the year of agents." — Cristiano Amon
• "The smartphone is the largest platform ever created by mankind... but the center of your digital experience is the agent that you choose." — Cristiano Amon
• "Tokens are the new currency of the AI world." — Cristiano Amon

---

Synthesis

The transition to an AI-driven world is fundamentally changing the relationship between humans and hardware. The industry is moving away from static apps toward autonomous, agentic experiences that span multiple devices. Qualcomm’s strategy centers on the "Compute Continuum," where powerful, energy-efficient SoCs enable AI to run locally on the edge while seamlessly integrating with cloud-based models. As the industry matures, the focus will shift from simple chatbot interactions to complex, agent-led workflows, necessitating a complete overhaul of current silicon architectures and consumer device form factors.