Qualcomm’s AI Agents: The Post‑App Future of Phones, Wearables, and Smart Glasses

“Agents Are Going to Be the New App”

In a recent CNBC interview, Amon laid out a vision where AI agents don’t just assist — they replace the interface layer entirely. Instead of opening your banking app, drilling through menus, and hunting for a transaction, you ask. The agent retrieves it. Done.

“Apps are not dead,” Amon said, “but apps are going to change.”

That’s a carefully worded statement from someone whose chips power a significant chunk of the world’s smartphones. It’s not a eulogy for apps. It’s a warning that the front door to software is being redesigned.

40+ Device Designs and Counting

Qualcomm isn’t just theorizing. The company is actively working on over 40 new AI device designs — spanning jewelry, camera-equipped earbuds, pins, and watches.

The common thread isn’t the form factor. It’s the function: something you wear, something always present, something that can see the world around you and connect you to an agent in real time.

That’s a meaningful design constraint. It shifts the question from “what screen size?” to “what context can this device capture?”

Smart Glasses: The Sleeper Category Waking Up

Amon is particularly bullish on smart glasses — and the numbers give him reason to be. Shipments are already in the “tens of millions” annually, and he projects that figure could hit hundreds of millions within a couple of years, potentially rivaling smartphone scale.

For context: 1.26 billion smartphones shipped in 2025. That’s the ceiling smart glasses are being measured against.

Meta, Samsung, and a growing list of players are already in the race. The glasses aren’t just a fashion experiment anymore — they’re becoming a serious endpoint for ambient AI.

The Real Gravity Shift

Here’s the structural insight buried in Amon’s comments: the agent becomes the center of gravity, not the device.

“The phone is around the agent. The new classes of devices are going to be around the agent as well.”

This reframes the entire competitive landscape. Apple and Samsung have spent decades competing on hardware specs and software ecosystems. If agents become the primary interface, the moat shifts toward whoever controls the most capable, most contextually aware AI — not necessarily the best screen or the thinnest chassis.

That’s a genuinely uncomfortable position for traditional hardware incumbents.

Why AI Companies Are Building Hardware Now

OpenAI’s acquisition of Jony Ive’s hardware startup wasn’t a vanity project. Amon’s framing explains the logic clearly: AI companies need to own the endpoints where agents live.

There’s a second, less-discussed motivation — data. Wearable devices that see, hear, and move with you will generate data at a scale “exponentially larger” than what trained today’s models. Whoever owns those endpoints owns the next training pipeline.

That’s not a feature. That’s a strategic land grab.

What This Means for the Chip Layer

None of this works without radically more efficient silicon. Smaller form factors demand chips that are simultaneously more powerful and dramatically more energy-efficient.

Amon didn’t mince words: “Our entire roadmap is in a process of upgrade right now. None of the devices we have today are prepared for the future.”

That’s a remarkable admission — and a useful signal. If Qualcomm is rebuilding its entire chip roadmap around agent-first devices, the hardware transition isn’t a distant scenario. It’s already in motion.

What to Watch as an AI Adopter

• Agent-first UX is coming to mobile. If your product lives inside an app, start thinking about how it exposes functionality to agents — not just users.
• Smart glasses are a real platform bet. The category is moving from novelty to scale. Distribution and context-capture capabilities will matter more than aesthetics.
• Data ownership is the new hardware moat. Companies entering consumer hardware aren’t chasing margins — they’re chasing training data and user context at ambient scale.
• On-device AI will accelerate. As chips evolve for smaller, always-on devices, expect a wave of tools optimized for edge inference rather than cloud dependency.

The Takeaway

The post-app future isn’t about apps disappearing. It’s about apps becoming invisible — absorbed into agents that understand intent and act on it without requiring you to navigate anything.

Qualcomm is betting its entire roadmap on that shift. Forty-plus device designs, a bullish call on smart glasses, and a chip architecture overhaul don’t happen because someone thinks this might be interesting.

The interface layer is being rebuilt. The question isn’t whether to pay attention — it’s whether you’re positioned for what comes after the icon grid.