Skip to main content
TechnologyJun 15, 2026· 3 min read

Google Aims to Turn Unused Smartphones into Servers

Google and the University of California, San Diego (UC San Diego) are working on a project aimed at giving a second life to discarded smartphones, turning them into a computing platform for academic and research activities. The initiative, called "Phone Cluster Computing", involves creating a data center composed of about 2,000 Pixel smartphones that have been withdrawn from the market or are no longer used by their owners.

The idea stems from the need to tackle one of the major environmental issues in the tech industry: emissions related to hardware production, known as "embodied carbon". While the industry has been working for years to reduce energy consumption in data centers and increase reliance on renewable sources, the production of new devices continues to represent a significant share of the overall environmental impact.

According to researchers, smartphones replaced by users every few years often retain much of their processing capability. Although considered obsolete for daily use, they still incorporate processors, memory, and storage sufficient for numerous workloads. Reusing them would therefore avoid the production of new hardware and extend the lifecycle of existing components.

One of the most interesting aspects that emerged from the study concerns performance. Tests conducted with the SPEC suite show that the high-performance cores of a modern smartphone from 2023, such as the Pixel Fold, can offer single-thread performance comparable to or even superior to that of cores in some current enterprise servers.

Naturally, the overall power of a server remains significantly higher due to the much larger number of cores and the availability of large amounts of memory, but the data highlights how mobile hardware maintains relevant computational capabilities even after several years of use.

To adapt the smartphones for use in a data center environment, researchers remove all unnecessary components, including displays, batteries, cameras, speakers, and chassis. Only the motherboard is retained, which houses the System-on-Chip and is also responsible for the majority of the emissions associated with the device's production.

The Android operating system is then replaced with a generic Linux distribution, better suited for running server applications and orchestrating workloads. This modification also allows for the elimination of numerous services and mechanisms designed for consumer use that are superfluous in a cloud context.

To coordinate the simultaneous operation of dozens or hundreds of smartphones, the project uses Kubernetes and containerized applications. The devices are organized into autonomous clusters generally composed of 25-50 phones, a configuration that, according to the benchmarks conducted, can provide processing capacity comparable to that of a modern dual-socket server processor.

The main destination of the platform will be to support university activities. Many educational and research applications used in universities require relatively modest resources and are currently run on small cloud instances. Preliminary tests indicate that a cluster of about 20 smartphones can handle peak requests from courses with over 75 students, maintaining competitive response times compared to traditional cloud solutions. The ultimate goal is to create infrastructure made up of 2,000 devices, capable of simultaneously supporting about a hundred university courses.

The project will also act as a testing ground for assessing the reliability of consumer hardware subjected to continuous workloads typical of data centers. Researchers expect to complete the system's commissioning during the autumn of 2026.

The project also fits into a line of research that has already been initiated in recent years. Several academic groups have experimented with reusing discarded smartphones as mini-data centers or environmental monitoring platforms, demonstrating how devices considered outdated can continue to perform useful tasks well beyond their commercial lifespan.