Overview

• Apple has begun early-stage discussions with Anthropic and OpenAI to train versions of their large language models for deployment on Apple’s private cloud for Siri testing.
• Internal evaluations reportedly found Anthropic’s Claude delivering stronger results than Apple’s in-house foundation models in key Siri use cases.
• Talks with Anthropic have stalled over its request for a multibillion-dollar annual licensing fee that would escalate each year.
• Apple has not taken a final decision and continues in parallel development of its internal “LLM Siri” project targeting a 2026 rollout.
• Recent leadership shifts placed Siri under Mike Rockwell with Craig Federighi overseeing broader AI strategy as Apple seeks to close the gap with rival assistants.