The AI arms race indicates a troubling concentration of power among a few technology giants, which control the most advanced AI models, the largest GPU resources, and critical data pipelines. Jake Brukhman, the founder of CoinFund, posits that decentralized networks leveraging distributed GPU resources can act as a substantial counterbalance to this concentration.
CoinFund has successfully raised $158 million to invest in a variety of crypto and AI startups, developing a portfolio that features numerous prominent decentralized computing projects.
#How can distributed GPUs change the landscape?
Brukhman emphasizes that numerous consumer and data-center GPUs, which often remain underutilized worldwide, can be combined into networks robust enough to collaboratively train sophisticated AI models. At the Theta Capital Legends4Legends conference, Brukhman elaborated on his vision, highlighting technological advancements that have transitioned decentralized model training from a theoretical proposition to a tangible reality. He believes a robust competition in decentralized AI training is on the horizon and will arrive sooner than anticipated.
#Where should investors focus their attention?
One of CoinFund's notable investments is in Prime Intellect, a project aimed at constructing an open-source decentralized AI framework. This initiative successfully attracted $5.5 million in seed funding co-led by CoinFund in April 2024, followed by an additional $15 million in later funding. Another example is Pluralis Research, which secured $7.6 million in a seed funding round co-led by CoinFund and Union Square Ventures in 2025. Furthermore, Gensyn operates a decentralized training network that utilizes an ERC-20 token called $AI. This token has several functions within its network, including verification, staking, payments, and governance, with a total supply capped at 10 billion tokens.
#What implications does this have for crypto investors?
For investors in the cryptocurrency market, decentralized AI initiatives introduce a fascinating new realm of token utility. Unlike tokens that primarily gain value through speculation or decentralized finance mechanics, AI-centric tokens are linked directly to computational workloads. A token that facilitates payment for GPU time within a decentralized network derives its demand from genuine resource consumption rather than mere trading activity.
CoinFund, established by Brukhman in 2015, began as a Slack group for early advocates of decentralization. Investors observing this sector must focus on compute utilization metrics, rather than simply token prices. If projects such as Prime Intellect and Gensyn can prove that decentralized GPU clusters effectively train models at scale, the rationale for investment shifts from speculative to foundational in nature.