Google CEO Pichai says Gemini's next leap depends on building "much larger base models"

2026-07-23 · Source: The Decoder · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Cloud Computing & IT Infrastructure, Corporate Strategy & Leadership · Depth: Intermediate, extended

Summary

Google reported strong Q2 2026 financial results, with revenue increasing 24% year-over-year to \$119.8 billion, exceeding analyst expectations. Google Cloud revenue grew 82% to \$24.8 billion, and the Gemini app reached 950 million monthly active users. CEO Sundar Pichai stated that the next leap for Gemini depends on building "much larger base models" for Gemini 4, with the most ambitious pre-training run yet underway. The company also noted efficiency gains in AI response costs and strong demand for its cost-effective Gemini Flash series. Alphabet raised its 2026 investment forecast to between \$195 billion and \$205 billion, reflecting continued high demand for compute capacity.

Key takeaway

For Directors of AI/ML evaluating future model investments, Google's strategy signals that foundational "much larger base models" are critical for achieving frontier AI capabilities. Your teams should prioritize investments in scalable compute infrastructure and full-stack AI solutions, recognizing that efficiency gains in serving models like Gemini Flash can significantly impact operational margins. Be prepared for continuous, rapid model iterations and focus on improving agentic coding to address current AI limitations.

Key insights

Future frontier AI capabilities hinge on developing "much larger base models" and a full-stack approach to AI solutions.

Principles

• Prioritize AGI development for long-term AI leadership.
• Optimize AI models for both performance and cost efficiency.
• A full-stack approach enhances integrated AI solutions.

Method

Google is undertaking its "most ambitious pre-training run yet" for Gemini 4, focusing on larger base models and continuous iteration, with monthly model releases planned.

In practice

• Utilize Gemini Flash models for cost-effective, high-performance AI serving.
• Explore agentic coding solutions to improve development workflows.
• Consider Google Cloud's integrated AI portfolio for enterprise solutions.

Topics

• Gemini 4
• Large Language Models
• AI Infrastructure
• Google Cloud
• AI Monetization
• Compute Capacity

Best for: CTO, VP of Engineering/Data, Executive, Investor, Director of AI/ML, Tech Journalist

Related on AIssential

• Why Google is Limiting Meta’s Gemini AI Use as Demand Soars — Global AI expansion faces severe compute infrastructure bottlenecks, impacting even major tech companies' growth and operations.
• Sundar Pichai admits Google trails in AI coding — Google acknowledges trailing in AI coding, stressing larger base models and improved agentic capabilities are vital for future competitiveness.
• Google's take on openclaw — The AI landscape is rapidly evolving with major model releases, strategic talent acquisitions, and significant financial shifts among key players.
• The Model Eats the Scaffolding: DeepMind's Logan Kilpatrick & Tulsee Doshi on 3.5 Flash, Omni & More — Google prioritizes cost-effective, fast AI models and integrated agentic harnesses for broad product integration and user experience.
• 😺 You Need 3 Geminis Now — AI model specialization for specific tasks is replacing the "one model does everything" paradigm.
• Google Delays Gemini 3.5 Pro to July 17: The Strategic Play Behind the Scrapped Base Model — Frontier AI competition demands generational leaps, not incremental model iterations.

See Counsel's argued verdicts on the open AI decisions leaders are weighing →

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by The Decoder.