A group of American researchers turned old smartphones into servers

According to Google Research, the amount of smartphones discarded after a few years of use is one of the biggest causes of the increase in electronic waste. To reuse this source, the research team at UCSD removed all unnecessary components of the phone such as screen, battery, camera, speaker and case, keeping only the motherboard containing the processor chip.

The original Android operating system was also replaced with Linux specialized for data centers to eliminate junk software and allow the installation of management software such as Kubernetes.

The results show that smartphones that have been released for three years still achieve higher single-core performance when compared to dedicated servers in some large data centers, such as the Asus RS720A-E11 model, which is equipped with two AMD EPYC processors and Nvidia H200 or RTX Pro 6000 graphics chips. Testing with the SPEC scoring tool shows that a cluster of 25-50 old phones can provide the same computing power as a high-end processor. two-socket server.

Theo Tom’s Hardware, The research team said a cluster of 20 old phones could run the application for a classroom of more than 75 students. Instead of having to rent expensive cloud services, schools can run their own applications on this homemade server system at a fraction of the cost of building a data center from new components, especially in the context of increasing prices of memory chips and hard drives.

UCSD is planning to collect 2,000 old phones to create an internal data center capable of supporting 100 classes at the same time. The system is expected to be put into trial operation later this year to evaluate the durability of consumer electronic components when subjected to continuous high-intensity operation.

Previously, another group of scientists connected old mobile devices into a micro data center for underwater monitoring in June 2025. In February, the US Space Agency (NASA) also reused the Qualcomm 801 chip launched in 2014 to integrate into the Ingenuity helicopter on Mars, supporting the positioning of the Perseverance self-propelled vehicle.

Experts say that artificial intelligence technology giants will not be interested in this because they need specialized hardware with high reliability and synchronization. However, this is a feasible and economical solution for universities, educational institutions or small businesses that do not have the financial potential to compete for new component supplies with large technology corporations.