AI Daily Digest — July 19, 2026 Your weekly roundup of the most consequential AI...
# AI Daily Digest — July 19, 2026
*Your weekly roundup of the most consequential AI developments.*
---
Apple escalated its legal battle against OpenAI this week, filing a trade secrets lawsuit accusing the AI company of systematically poaching Apple hardware engineers and stealing intellectual property to build its consumer AI hardware business. The lawsuit names OpenAI Chief Hardware Officer Tang Tan — formerly Apple's VP of Product Design who led iPhone and Apple Watch development — as a defendant. Apple alleges Tan instructed job candidates from Apple to bring prototype components and internal design documents to interviews.
The complaint reveals that over 400 former Apple employees now work at OpenAI. Apple's legal team sent document preservation letters to approximately 40 of those individuals, demanding they retain all communications and meet with Apple attorneys for deposition. Apple is seeking an injunction requiring OpenAI to stop using any alleged trade secrets and to redesign its forthcoming hardware products.
The legal confrontation underscores the escalating rivalry between the two companies as AI moves from software into physical devices. OpenAI is reportedly developing a portable AI hardware device with microphones and cameras — no screen — designed to serve as an always-on personal AI assistant. Apple is simultaneously developing smart glasses, camera-equipped AirPods, and a next-generation HomePod with its own AI stack.
— Apple · OpenAI · Reuters
🔗 [Apple](https://apple.com) · [OpenAI](https://openai.com) · [Reuters](https://reuters.com)
---
Meta is in preliminary negotiations to rent out its vast data center compute capacity to Anthropic, in a deal that could be worth up to $100 billion over two years. The arrangement would mark a strategic shift for Meta — which has spent tens of billions building AI infrastructure primarily for its own models — and effectively turn the social media giant into a compute supplier to a direct rival.
Under the proposed terms reported by the New York Times, Anthropic would pay Meta monthly installments over 24 months, with both sides retaining the right to terminate early. The flexible structure functions more as a capacity option than a fixed lease, reflecting the uncertainty on both sides. Anthropic proposed the deal in June, and Meta is currently evaluating it.
This follows Anthropic's earlier $450 billion, three-year compute agreement with SpaceX's Colossus 1 data center, announced in May. The spate of massive compute procurement deals signals that frontier AI labs are aggressively securing compute capacity well beyond what cloud providers alone can supply. Meta CEO Mark Zuckerberg, who had previously hinted at a cloud computing pivot, is betting that renting excess GPU capacity can help justify Meta's projected $145 billion in