Why Qualcomm is betting on AI data centers

Summary · 3 min read

Cristiano Amon is CEO of Qualcomm, whose chips power Android phones, Meta smart glasses and upcoming devices from OpenAI, according to the show description. Amon says he began working on 1G and has experienced each cellular generation.

Amon argues that AI agents could renew demand for smartphones by making them personal, context-aware tools that coordinate on-device and cloud computing. He says consumer AI adoption is moving quickly in China, while US discussion focuses more on frontier models and enterprise use, and he favors continued development with cybersecurity guardrails. He sees smart glasses as the leading AI wearable and says Qualcomm can bring its energy-efficient chip expertise to data centers while offering developers an open platform through Modular. He describes 6G as infrastructure for agentic devices and AI sensing, with demonstrations planned for 2028 and scale expected after 2030.

Minute by minute

1. 0:44

   China's AI adoption is device-led. Amon says consumer AI adoption is moving quickly in China, where he sees greater focus on physical AI and edge devices. He says US discussion centers more on frontier models and enterprise adoption; concerns about jobs are shared, while he sees less concern in China about data-center electricity use.

2. 4:29

   AI needs guardrails, not extremes. Amon favors continued AI development alongside guardrails, rather than stopping development or moving ahead without limits. He points to cybersecurity risks: models could help people find software vulnerabilities, so he says companies need controls and regulation. Qualcomm is building hardware controls for users' information and personal data.

3. 7:12

   Commercial links can steady US-China ties. Amon says semiconductors have become a geopolitical industry and AI is part of those discussions. He says commercial partnerships between American and Chinese enterprises create stability and shared understanding between the countries.

4. 8:14

   Memory costs suppress phone sales. Amon says the smartphone market has not recovered to its 2019 size, though about 1.2 billion phones still sell each year. He says memory prices rose 5x to 6x and the market is down 20% this year, with mid- and low-tier phones more affected. He sees pent-up demand and says growing AI workloads are lifting demand for premium phones.

5. 13:00

   Phones and cloud will share agent workloads. Amon says some AI companies want phones able to run a 100 billion-parameter model by 2028. He expects cloud and device computing to coexist: agents will need cloud capabilities, but some tasks, personal apps and fast responses will depend on the phone.

6. 18:20

   AI phones act on personal context. Amon defines an AI smartphone as a device where an agent can plan, route work between local and cloud computing, use personal context and take action. He expects third-party agents and skills to cross app and operating-system boundaries, much as third-party apps shaped the smartphone.

7. 24:55

   Smart glasses lead AI wearables. Amon describes VR as closer to a gaming-console market than an everyday device. He says generative AI made smart glasses more useful for tasks such as translation and identifying people, estimates the market at 50 to 60 million units shipped, and expects glasses to become the largest wearable category.

8. 29:22

   Wearables need trust and low friction. Amon says watches, pins, pendants, jewelry and camera-equipped earbuds may all find uses, but he cannot yet identify the next major form factor. He compares the opportunity to the Walkman or iPod and says adoption will depend on trust, clear value and low friction.

9. 30:38

   Energy constraints open data center opportunity. Amon says Qualcomm's expansion from mobile into PCs and cars built capabilities that can transfer to data centers. He sees an opportunity to make energy-efficient, high-performance CPUs and accelerators as compute demand grows and energy availability becomes a constraint.

10. 34:45

    Modular aims to open AI software. Amon says Qualcomm and Modular share a vision for an open software platform that can run AI across different hardware. He says Modular will use an Apache 2.0 license and Qualcomm will welcome deployments on competing chips, because an open platform lets Qualcomm compete on hardware.

11. 39:03

    5G made unlimited data practical. Amon says 5G made unlimited smartphone data rates economically viable and enabled high-definition video downloads. He contrasts that with 4G, which brought broadband to phones and helped make music streaming practical.

12. 41:49

    6G adds sensing and AI context. Amon says 6G is designed for AI, with high-definition video upstream for devices such as glasses. He also expects networks to use radio signals for radar-like sensing, including real-time mapping and drone detection, and says 6G infrastructure could carry tokens for agentic experiences.

13. 45:00

    6G rollout follows 2030. Amon says Qualcomm plans demonstrations at the 2028 Los Angeles Olympics. He points to prototypes and silicon in 2029 or 2030, early deployments, and broader scale following 2030.

Published Oct 8, 2026.