628: Apple’s Air Traffic Controller Strategy, Anthropic's $30bn ARR + TPU Deal, Meta's Leaderboard, Berkshire Hathaway’s Mountain, Visa, Moody's & S&P Global, Max Life Expectancy, and Andy Weir
Summary
This intelligence brief covers several key developments across technology, finance, and science. It announces a price adjustment for "Liberty's Highlights" subscriptions, increasing monthly rates to $15 and annual rates to $150 starting May 1, 2026, citing inflation. The brief then analyzes Apple's "Air Traffic Controller" strategy in AI, suggesting Apple aims to own the platform chokepoint by integrating multiple AI models like Gemini into Siri, rather than investing heavily in compute capex. It highlights Anthropic's rapid growth, surpassing $30 billion ARR, and its strategic 3.5-gigawatt TPU deal with Google and Broadcom to address compute shortages. Meta's internal "Claudeonomics" leaderboard, which tracks employee AI token usage, is critiqued as a potential misapplication of metrics. Visa's new AI-powered dispute resolution tools, designed to reduce $ billions in annual losses from inefficient processes, are presented as a significant real-world AI deployment. The brief also touches on Berkshire Hathaway's growing cash reserves, the AI disruption risks for Moody's and S&P, and advancements in weather prediction, AI vulnerability, and cancer survival rates.
Key takeaway
For CTOs and VPs of Engineering evaluating AI infrastructure and strategy, consider Apple's "Air Traffic Controller" approach. Instead of solely focusing on massive compute investments, you should explore owning key platform chokepoints and integrating a constellation of AI models. This strategy can offer better pricing, performance, and brand control, while also mitigating risks associated with a single model provider's leadership or compute crunch. Diversifying your AI model dependencies can provide significant leverage and adaptability.
Key insights
Owning platform chokepoints and diversifying AI models can be a strategic alternative to massive compute capex.
Principles
- Network effects of standards are highly resilient to disruption.
- Rewarding quantity over outcome distorts behavior and wastes resources.
Method
Apple's "Air Traffic Controller" strategy involves integrating diverse AI models (e.g., Gemini) into its platform (Siri, iOS 27) and distilling them into custom models, leveraging powerful on-device silicon for local tasks and cloud models for complex ones.
In practice
- Consider a multi-model AI strategy for flexibility and cost optimization.
- Implement AI for high-volume, inefficient processes like dispute resolution.
- Beware of metrics that incentivize quantity over quality in AI usage.
Topics
- Apple AI Strategy
- Anthropic Growth
- AI Compute Deals
- AI Security Vulnerabilities
- Financial AI Solutions
Best for: CTO, VP of Engineering/Data, Director of AI/ML, Investor, Consultant
Related on AIssential
Editorial summary, takeaway, and curation by AIssential. Original article published by Liberty’s Highlights.