Modern II requires a large amount of computing, and only the largest companies can compete. Wouldn’t it be great if smaller outfits could dispute technological hippos with their own AI algorithms? This is what the Pluralis study thinks, and it’s one of the handfuls of startups that believe that blockchains – essentially Amazon Web – service – can be a response.
On Wednesday, Pluralis announced that he had provided $ 7.6 million to Dasei and seeds led by Venture Capa Coinfund and Union Square Ventures. Topology, option, Eden BlockAnd Bodhi Ventures also participated in the round, as well as the outstanding investor Crypto Schrinivasan and Clem Delange, co -founder of the popular AI Huggingface platform.
The increase was for capital, with a warrant for future cryptocurrency – if Pluralis decides to launch one, Alexander Long, founder and CEO, said Wealth.
Currently, Pluralis has no product, that is, investors are first and foremost betting on the Long, and the veterans team of AI researchers, which he collected. “I raised the circle:” This team is the right team that will try to solve this problem, “Long said.” No one is trying anymore.
Pluralis founder has a doctoral degree in computer science and worked in Amazon for more than three years as an engineer. He collected seven other computer scientists, all with his doctoral sciences were either doctoral students to find out if you can create powerful AI algorithms through a decentralized server network.
Currently, AI with high flight levels requires the fleet of expensive computers in stock that works together to train massive algorithms. Explorers According to estimates To obtain an earlier version of the GPT Openai or algorithm in the United States, 1300 megawatts were required, or approximately the same electricity consumed by 130 houses in the United States. It costs money, which means the only firms that can win the AI ARMS race have big pockets.
For the founders of Kryptus, whose technology relies on the ideal of decentralization, the idea that several corporations can possess such high power is why some researchers of the II started thinking about how to create or train, powerful algorithms II through a decentralized network of servers. Gensin, the main intelligence and several firms are already studying this opportunity.
The Pluralis approach is another, Long said. Most attempts to train AI through a network of decentralized computers require these computers to load the entire model. When the servers are small, it puts the ceiling in size – and the power – models. Long wants to explore whether you can train parts of the model, not the whole model on one computer. “If you can make the problem accurate enough, it often leads to immediate ways to start solving it,” he said.
Investors believe that Long, which has already started his research maybe, and they have enough faith in him and his team that they are willing to bet – even if it’s long. “If this works,” said Jake Bruchman, founder and CEO Coinfund Wealth“It will change the world.”
Originally this story was presented on Fortune.com
https://fortune.com/img-assets/wp-content/uploads/2025/03/Alex-headshot-e1742334415602.jpg?resize=1200,600
2025-03-19 14:00:00
Ben Weiss
