Qualcomm and Xiaomi on On-Device AI, Flagship Chips, and the Indian Market

At the India Mobile Congress (IMC) 2026, HT Tech interviewed Sumit Sonal (Head of Marketing, Qualcomm India) and Gautam Batra (Associate Director, Xiaomi India). The discussion focused on Qualcomm’s newly unveiled Snapdragon 8 Elite Gen 6 and Snapdragon 8 Elite Extreme Gen 6 chipsets, the future of on-device AI, and Xiaomi’s expanding flagship strategy for India.

1. What’s New Beyond Raw Performance?

While annual performance boosts are standard, Qualcomm highlighted two major user-facing upgrades in the new 2nm-based chipset family:

• Pixel-Level Subject Detection: Upgrading from traditional semantic segmentation, the image signal processor (ISP) can now detect and track individuals down to the pixel level—separating subjects (face, body, clothing) from complex backgrounds like mountains or landscape backdrops.

• On-Device AI Agents (Up to 30B Parameters): The Extreme variant can run large language models (LLMs) with up to 30 billion parameters locally. This provides three primary advantages:

  1. Lower Latency: Instant response times without cloud round-trips.

  2. Enhanced Privacy: Sensitive personal data stays on the device.

  3. Offline Reliability: AI tasks continue to function even in weak or absent network conditions.

2. Practical Examples: On-Device AI Workflows

Sumit Sonal provided an example of how localized AI handles daily tasks:

Example: If you receive an invitation email for a trip to Bengaluru, an on-device AI agent can read the context, check your personal schedule, search for travel options, and suggest flights directly within the app.

Only external data retrieval (like pulling live flight prices) requires an internet connection, while the processing and schedule management occur locally. On-device execution also reduces reliance on costly token-based cloud computing, making AI features more scalable for manufacturers and end users.

3. Dual Flagship Strategy: Why Two Chips?

Qualcomm explained that flagship smartphone buyers have split into two distinct categories:

1. Power Users & Enthusiasts: Require heavy agentic workloads, 30B-parameter models, and high-end video recording (Snapdragon 8 Elite Extreme Gen 6).

2. Everyday Premium Users: Want flagship gaming, thermal control, and camera performance without needing maximum AI computational overhead (Snapdragon 8 Elite Gen 6).

This split gives smartphone makers the flexibility to offer high-end devices at varied price points without compromising core performance or thermals.

4. Efficiency, Thermal Control, and Engineering in India

• 2nm Process Node: The transition to a 2nm architecture delivers lower power consumption and improved battery endurance alongside raw speed gains.

• India’s Contribution: India hosts Qualcomm’s largest engineering center outside San Diego (~19,000 employees). The Indian engineering team contributed directly to global platform developments, including automotive, 5G, and the 2nm tape-out process.

• Xiaomi’s Optimization: For upcoming devices like the Xiaomi 18 Pro, Xiaomi integrates hardware cooling systems with software tuning to ensure stable frame rates and thermal stability in diverse climates.

5. Xiaomi’s Strategy for AI and the Indian Flagship Market

• Focus on Real-World Reliability: While the Xiaomi 18 series features a dedicated AI button and integration with Google’s Gemini ecosystem, Gautam Batra emphasized that long-term hardware reliability, camera quality, and daily usability remain top priorities alongside AI experimentation.

• Going “All-In” on India: Addressing past gaps (such as skipping certain flagship variants in India), Xiaomi confirmed it is taking an “all-in” approach moving forward. Recognizing India’s growing appetite for premium smartphones, ecosystem products, and smart TVs, Xiaomi plans to bring its full suite of premium products to the region.

Key Chipset Feature Summary