The technology world is resorting to extraordinary systems to overcome the global hardware crisis that artificial intelligence models face in data processing processes. The daily operational costs of systems such as ChatGPT and Gemini have led technology giants to seek infrastructure that is both environmentally friendly and low-cost. In this context, Google, which is carrying out a joint project with researchers at the University of California San Diego (UCSD), is recycling old smartphones sitting in drawers with a revolutionary recycling model. Thanks to the new system developed, thousands of old portable devices have become high-performance devices. Google artificial intelligence serveris converted into a component.

Old Pixel Phones Come to Life Again in the Laboratory

During the engineering phase of the project, old smartphones that have reached the end of their useful life are put through a special sorting process in a laboratory environment. The hardware of the devices, such as the screen, camera, battery and sensor, which are not needed in data center operations and consume extra power, are completely separated from the body. The remaining bare motherboards are arranged in special server racks and connected to each other via a high-speed local network infrastructure. This new artificial intelligence cluster, programmed with a Linux-based optimized operating system, successfully undertakes the heavy computational loads in cloud computing processes.

The single-core performance power that mobile processors have reached in recent years ensures that this project runs technically flawlessly. Since smartphone processors are inherently designed to achieve maximum efficiency with minimum power, they require only a small fraction of the huge amount of electricity consumed by classical servers. In the first phase, the prototype was built with 20 old devices. Google artificial intelligence serverThe cluster managed to manage the concurrent cloud processes of more than 75 active students at the university without any delays.

A Huge Cluster of 2,000 Devices Goes into Operation in the Autumn

The goal in the next phase of the project is much bigger. Researchers at Google plan to create an industrial-scale computer cluster by connecting 2,000 old Pixel phone motherboards. This system, which is expected to become fully operational next autumn, will be directly equivalent to the total computing power of 50 standard corporate servers used in data centers.

The project, which aims to reduce the global carbon footprint of artificial intelligence operations, also stands out as one of the most concrete green technology solutions against the millions of tons of e-waste crisis.

If this pilot becomes standard practice in major data centers around the world, the hardware supply crisis in the technology sector could be greatly alleviated. In the future, the hardware of the old devices we forget in the drawers in our homes will be global devices that power the world’s most complex artificial intelligence algorithms. Google artificial intelligence serverIt can feed the internet infrastructure.