AI and XR take center stage at Google I/O 2025

Google’s I/O event this year showcased a strategic pivot toward artificial intelligence with extended reality (XR) as a complementary focus, reflecting the company’s evolving priorities beyond its Android ecosystem. The event revealed an impressive acceleration in Google’s AI initiatives, particularly with Gemini models seeing 50x growth in token usage over the past year, while simultaneously advancing multiple XR hardware and software partnerships that position Google at the intersection of these transformative technologies.

The big picture: Google is aggressively expanding its AI capabilities while strategically re-entering the XR market through partnerships rather than solo hardware ventures.

• Gemini AI usage has increased 50x in monthly token consumption compared to last year, marking substantial adoption.
• The company is replacing Google Assistant with Gemini across multiple platforms including Google TV, Android Auto, and WearOS.

Key developments: Gemini 2.5, released in March 2025, represents a major inflection point in Google’s AI strategy.

• The model delivers significant performance improvements that have accelerated adoption across Google’s ecosystem.
• This expansion demonstrates Google’s commitment to integrating AI capabilities throughout its product portfolio.

Hardware partnerships: Google has announced multiple XR collaborations instead of pursuing its own standalone devices.

• Samsung‘s Project Moohan, Warby Parker and Gentle Monster for smart glasses, and Xreal’s Project Aura AR headset represent Google’s partner-focused approach.
• These partnerships leverage Google’s software expertise while allowing hardware companies to handle device manufacturing.

Enterprise focus: Google Beam, the evolution of Project Starline, shows Google’s commitment to enterprise XR applications.

• Developed with HP, the XR conferencing platform features near-real-time translation capabilities.
• Initial customers include major corporations like Deloitte, Salesforce, and Citadel, signaling enterprise market potential.

On-device intelligence: Google is enhancing its edge AI capabilities to complement cloud processing.

• Improvements to the LiteRT runtime and the introduction of the AI Edge Portal demonstrate a commitment to hybrid AI processing.
• This approach balances the benefits of on-device processing with cloud-based capabilities for optimal performance.