Updated
Updated · Yahoo Finance · Sep 21
Google Launches Gemini 3.8 Voice AI Models in 97 Languages for Enterprise
Updated
Updated · Yahoo Finance · Sep 21

Google Launches Gemini 3.8 Voice AI Models in 97 Languages for Enterprise

3 articles · Updated · Yahoo Finance · Sep 21

Summary

  • Google unveiled Gemini 3.8 Live and Gemini 3.8 Live Extended Thinking, two models built to power voice-based AI agents for complex enterprise tasks.
  • The new models can make tool and API calls during conversations, process visual input nearly in real time, and in the Extended Thinking version handle multi-step reasoning.
  • Salesforce, Genspark and Lumeris are already working with Google on applications, signaling a push to embed Gemini deeper into enterprise workflows.
  • For Alphabet, whose shares have gained about 40% over 12 months, the release highlights how heavy AI spending is being aimed at faster growth in Google Cloud and other business software markets.

Insights

With Google quietly buying up corporate data to train Gemini, what secrets are hiding inside the next generation of AI voice agents?
Alphabet is burning billions on AI infrastructure in 2026, but will Gemini 3.8's enterprise push actually rescue their negative cash flow?
As Gemini 3.8 Live seamlessly handles complex API calls and real-time facial animations, are human call centers facing imminent extinction?