Home » How Apple’s NeuralEdge Chip Design Became a Trade Secret Target

How Apple’s NeuralEdge Chip Design Became a Trade Secret Target

By

–

Tim Cook

CUPERTINO, July 11 — A legal battle is brewing that could reshape how the world’s largest technology companies protect their most secretive research. Apple has filed a lawsuit against OpenAI, alleging that the startup’s Aura smart glasses rely on technology lifted directly from Apple’s own custom AI processor, the NeuralEdge. The NeuralEdge chip was first detailed in Apple’s internal documents from 2022.

It is a custom-designed piece of silicon built to run large language models locally on a device—meaning complex AI tasks can happen on a phone or a pair of glasses without needing to send data to the cloud. For users, that translates to faster responses and stronger privacy, since personal information never leaves the device.

Apple invested $1.2 billion in the NeuralEdge’s development over three years, according to the company’s internal records. A dedicated team of 300 engineers, split between Cupertino and Austin, worked on the project. The chip uses what Apple calls a novel “sparse attention” architecture, a technical approach the company claims is unique to its design.

The heart of the allegation

The lawsuit centers on OpenAI’s Aura smart glasses. Apple alleges that the chip powering those glasses shares identical power efficiency metrics and memory bandwidth specifications with the NeuralEdge. Those are not general industry standards, Apple argues—they are specific, hard-won engineering choices that point to direct copying.

Apple’s trade secret protections for the NeuralEdge project were extensive. Every member of the 300-person team signed non-disclosure agreements that carry penalties of up to $5 million for breaches.

The company is now asking the courts to determine whether those agreements were violated. What this means for the broader technology landscape is significant. Apple’s approach—designing its own chips rather than buying them off the shelf—has become a core part of its strategy.

The NeuralEdge represents a bet that on-device AI will define the next generation of computing, from phones to wearables. If a competitor can replicate that work without the same investment, the entire model is undermined.

Behind the breakthrough

The sparse attention architecture at the heart of the NeuralEdge is a technical innovation that allows the chip to focus its processing power on the most relevant parts of a large language model, rather than churning through every piece of data equally. That efficiency is what makes it possible to run sophisticated AI on a device with limited battery life and heat constraints—exactly the kind of challenge a pair of smart glasses presents. Apple’s team in Cupertino and Austin spent three years refining that approach.

The $1.2 billion investment covered not just the chip design itself, but the supporting software, testing infrastructure, and manufacturing partnerships needed to bring it to market. OpenAI has not yet filed a formal response to the lawsuit.

The company’s Aura smart glasses have been positioned as a breakthrough in wearable AI, offering users a hands-free interface to large language models. If Apple’s allegations hold up in court, it could force OpenAI to redesign its hardware or pay significant damages.

What to watch next

The case will likely turn on whether Apple can prove that the specific combination of power efficiency and memory bandwidth found in the Aura chip could only have come from its own proprietary work. The non-disclosure agreements signed by the NeuralEdge team provide a paper trail, but the question is whether that trail leads anywhere. For the engineers who spent years on the NeuralEdge project, this is more than a corporate dispute.

It is about whether the protections they relied on—the agreements, the secrecy, the billion-dollar investment—will hold up when a competitor enters the same space. The answer will affect not just Apple and OpenAI, but every company that pours resources into developing custom silicon for the AI era.

Sources

More stories by

Most read

Electronics & Computing

Cybersecurity

Gadgets & Reviews

Business