Physical Address
304 North Cardinal St.
Dorchester Center, MA 02124
Briefly
- The gradient has raised $ 10 million from Panther, MultiCoin and HSG to introduce Lattice and Parallax – protocols designed to launch AI models on decentralized devices instead of centralized servers.
- The system uses unused computer power from phone, laptops and IoT devices, using salt to coordinate data and payments.
- The team says this approach reduces the costs, retains the user data to the local and pushes against AI monopolis – but critics say that latency and complexity could be the blockade of the road.
Gradient Network has closed a $ 10 million seed circle to build what he calls a decentralized AI infrastructure gathering, and the risky Pantera Capital and MultiCoin Capital companies are leading the investments along with HSG (earlier Sequoia Capital China).
The Singapore’s headquarters plan to use the means of developing two basic protocol-tlatica and parallaxes-which would allow artificial intelligence models to switch through the distributed network of the device, not in centralized data centers. The company announced that both protocols would debut this week.
“We believe that intelligence should be a public good, not corporate assets,” said Eric Yang, co-founder of Gradient Network, in the announcement divided Decipher. “This circle gives us momentum to build an infrastructure that brings decentralization to the heart of Ai.”
Time arrives as AI companies are facing strengthening criticism for privacy of data and concentration of computer power among a handful of technological giants. Gradient -O’s approach would be used in an unused processor power from smartphones, computers and other devices to create what could be basically being an equivalent to a global, full of supercomputer.
Lattica functions as a protocol of communication with data data such as bitcoin or torrent-a-thought of it as a water supply system that moves information between the device without passing through the central server. The company has announced its network “Strong nodes“It has already facilitated more than 1.6 billion bonds in more than 190 regions.
Path to Lattica
Over the past months, our Sentry Noda Community has helped to launch one of the world’s largest decentralized connections experiments with over 1.6B relationship with Peer-to-PEER connections in 190+ regions.
Today, that network develops into Lattica.
– Gradient network (@gradient_hq) June 19, 2025
Decentralization AI
Parallax deals with the problem of launching huge AI models without massive data centers. The protocol dissects large linguistic models into smaller pieces that can be performed at multiple devices at the same time. Instead of sending data to the Openi or Amazon server to processing, the parallax would allow the calculation to occur on the network of participating devices, holding the user data to the local.
To be sure, critics have caused concern that the coordination of tasks on thousands of devices brings complexity and that the delay of the network remains a challenge for decentralized systems.
But the company says its distributed approach and technology could reduce costs compared to traditional cloud computing, at the same time solving privacy problems. When AI models are running on centralized servers, user inquiries and data transmit and process these servers. The gradient system would process the data closer to the place where it is generated.
The Gradient Network operates on Solan Blockchain, selected for high transaction speeds and low costs compared to other networks. Blockchain manages the mechanisms of coordination and payment for devices that contribute to computer power network.
Startup joins the increasing area of companies trying to decentralize AI infrastructure. Competitors include a singularitynet, focusing on the creation of a market for AI services, The covenant of the superiors Network and various projects based on different blockchains. Bitensor and Gensyn followed similar distributed computer models, although with different technical approaches.
Gradient said he would publish additional protocols outside Lattica and Parallax, although he did not determine what could be included. The company also mentioned plans to publish research papers and open channels for developers that can contribute to the project.
Edited by James Rubin
Generally intelligent Bulletin
Weekly AI journey narrated by gene, generative AI model.
