Ex-Google engineers charged with stealing Pixel chip secrets

Two former Google engineers and Samaneh Ghandali’s husband were indicted in the US on 14 felony counts, including conspiracy, theft of trade secrets, and evidence destruction, related to Google’s Tensor processor for Pixel phones.

The accused, all Iranian nationals, include Samaneh Ghandali, a former hardware engineer at Google, her sister Soroor Ghandali, an intern, and Mohammadjavad Khosravi, who applied for a Google position but was not hired.

They were arrested in San Jose, California, and appeared in federal court on February 20, 2026.

If convicted on the most serious charges, they face a minimum of 20 years in prison.

Google has strengthened safeguards to protect confidential information and cooperated with law enforcement following the incident.

🔗 Source: Bloomberg

🧠 Food for thought

Implications, context, and why it matters.

The stolen secrets tie to Google’s long-term push for custom AI hardware

Google’s Tensor chips are custom smartphone systems-on-a-chip (SoCs) built for on-device AI and machine learning, which Google says unlocks Pixel features that off-the-shelf chips did not support 1.
That design supports Pixel tools such as live language translation and computational photography modes like Motion Mode, Action Pan, and Long Exposure 1.
The effort echoes Apple’s move to custom silicon, which took years and tightened the fit between hardware and software 1.
The Tensor name comes from Google’s data center Tensor Processing Units (TPUs) used in Google Cloud infrastructure 1. Google said TPUs contributed to its invention of the Transformer architecture 2.

The incident raises the stakes in the AI hardware race

An indictment tied to alleged Tensor trade secret theft puts custom chip designs on the same battleground as AI software 2.
U.S. export controls on advanced AI chips can raise the sensitivity and value of chip-related intellectual property 2.
The situation may drive tougher security checks for engineers at big tech firms, which could slow hiring and collaboration for people with ties to sanctioned countries.
It adds weight to in-house chip efforts at cloud giants such as Google, Amazon Web Services, and Microsoft, which build custom application-specific integrated circuits (ASICs) while still relying on Nvidia graphics processing units (GPUs) 2.