Gemini To Power The New Siri?

November 6, 2025

20 min episode · 2 min read

Sarah Friar,Mustafa Suleyman

Episode

20 min

Read time

2 min

AI-Generated Summary

Published Dec 23, 2025

Key Takeaways

✓Apple-Google Partnership: Apple pays Google approximately $1 billion annually for a 1.2 trillion parameter Gemini model to rebuild Siri, launching spring 2026 as interim solution until Apple develops comparable in-house models, with data isolated on Apple's private cloud servers.
✓OpenAI Infrastructure Strategy: OpenAI seeks federal government loan guarantees to reduce financing costs and increase debt capacity for data center deals, enabling aggressive expansion from 200 megawatts two years ago to 2 gigawatts currently, targeting country-sized computing deployments.
✓Enterprise Revenue Shift: OpenAI's enterprise sales now represent 40% of total revenue, up from 30% at year start, with customers moving from pilot programs to full production deployment in financial services and healthcare sectors, demonstrating viable path beyond consumer subscriptions.
✓Google TPU Performance: Google's seventh-generation Ironwood TPU delivers over 4x speed improvement versus previous generation, connecting up to 9,216 chips per pod, with Anthropic committing to use up to 1 million units for Claude model operations, challenging NVIDIA's GPU dominance.

What It Covers

Apple finalizes a billion-dollar annual deal with Google to power Siri with Gemini's 1.2 trillion parameter model by spring 2026, while OpenAI seeks government backing for massive data center financing.

Key Questions Answered

•Apple-Google Partnership: Apple pays Google approximately $1 billion annually for a 1.2 trillion parameter Gemini model to rebuild Siri, launching spring 2026 as interim solution until Apple develops comparable in-house models, with data isolated on Apple's private cloud servers.
•OpenAI Infrastructure Strategy: OpenAI seeks federal government loan guarantees to reduce financing costs and increase debt capacity for data center deals, enabling aggressive expansion from 200 megawatts two years ago to 2 gigawatts currently, targeting country-sized computing deployments.
•Enterprise Revenue Shift: OpenAI's enterprise sales now represent 40% of total revenue, up from 30% at year start, with customers moving from pilot programs to full production deployment in financial services and healthcare sectors, demonstrating viable path beyond consumer subscriptions.
•Google TPU Performance: Google's seventh-generation Ironwood TPU delivers over 4x speed improvement versus previous generation, connecting up to 9,216 chips per pod, with Anthropic committing to use up to 1 million units for Claude model operations, challenging NVIDIA's GPU dominance.

Notable Moment

Microsoft AI CEO criticizes treating AI systems as sentient beings with human-like feelings or rights, emphasizing chatbots should not trick users into believing they experience suffering or pain, advocating for systems aligned to human values by default.

Know someone who'd find this useful?