Google Finance adds prediction markets and AI features for research, earnings (blog.google)

Google has upgraded Google Finance with AI-driven research and market features: Deep Search, prediction-market feeds, and a richer earnings experience, plus an initial expansion to India (English and Hindi). Deep Search uses Google’s Gemini models to run up to hundreds of concurrent searches and reason across disparate sources, returning fully cited, multi-step research answers with a visible research plan and follow-up querying. Prediction-market data from Kalshi and Polymarket is now integrated so users can query crowd-implied probabilities (and view their time series) for future events like GDP or election outcomes. Earnings tooling adds an “Upcoming earnings” calendar, an Earnings tab with live audio and real-time transcripts, AI-generated “At a glance” insights before/during/after calls, and quick comparisons to historical financials and filings. For the AI/ML and finance communities this tightens the bridge between large language models, structured market data, and decision workflows: researchers and traders get faster, more transparent synthesized research and crowd-sourced probability signals in the same interface. Technical implications include heavy multi-query orchestration by Gemini, provenance/citation surfaces for explainability, and tiered capacity (higher Deep Search limits for AI Pro/Ultra subscribers). The rollout via Labs and staged US-first feature availability signals cautious testing — but also raises questions about model accuracy, data sourcing, and how practitioners should audit AI-driven investment insights.