A new system connects devices without the internet using local networks and distributed computing across phones and hardware.
A three-month-old startup EVE NET is building an offline communication and AI platform that aims to reduce dependence on cloud infrastructure by distributing computing across existing mobile devices. Alongside the software platform, the company has also launched a hardware product called EVE device, designed for locations with limited or no internet connectivity.
Speaking exclusively with Jash Jetly, co-founder and CEO of EVE NET who revealed that the startup’s long-term goal is to make AI more affordable by distributing model storage and compute across millions of connected devices instead of relying entirely on centralized data centres.
“The biggest cost in today’s AI services is compute and storage. We’re solving that by distributing a large AI model into smaller models that run across multiple mobile phones instead of relying on centralized infrastructure. This significantly reduces the cost of running AI. The platform is still under development and has not been released publicly yet,” Jash said.
EVE device creates a localized network that enables communication and data sharing in places where internet access is unavailable or unreliable. One of its target applications is remote construction sites. Companies working in isolated locations can deploy an EVE device to establish a local communication network without relying on internet connectivity.
Another target market is educational institutions. “Universities conducting online programming exams are increasingly struggling to prevent students from using AI tools during assessments. Our device creates an offline network across a campus and includes a kill switch that blocks internet access within its coverage area, while still allowing students and faculty to communicate over the local network,” added Jash.
The startup is also targeting organizations looking for lower-cost data storage. “Our storage service is designed to reduce data transfer costs compared with conventional cloud platforms. The distributed AI platform behind it is still under development,” added Jash.
Unlike open-source mesh networking solutions that often require software configuration and hardware assembly, the EVE device is designed as a plug-and-play device that only requires a USB-C connection for deployment.
While the distributed AI platform is still under development, the startup’s approach combines offline networking with distributed computing to reduce reliance on centralized cloud infrastructure. If successful, the technology could enable communication, storage and AI services to operate in locations with limited internet connectivity while lowering the cost of running AI applications.
