Planck, an infrastructure protocol focusing on the unreal intelligence growth, has launched a layer-0 blockchain geared toward supporting decentralized AI networks, the corporate mentioned Tuesday.
The blockchain is designed to function foundational infrastructure for AI functions, significantly these constructed for decentralized bodily infrastructure networks (DePINs). These networks mix {hardware}, token incentives and distributed processing to create options to centralized assets, reminiscent of cloud providers.
The transfer aligns with a rising push inside the crypto trade to convey Web3 rules of decentralization to AI growth, a sector nonetheless dominated by centralized gamers reminiscent of OpenAI and Google.
“At present, high-performance AI compute is closely centralized within the fingers of some tech giants,” a spokesperson for Planck advised Cointelegraph.
Planck is one in every of a number of blockchain tasks engaged on different approaches to AI decentralization. Bittensor, for instance, focuses on decentralized machine studying, whereas Fetch.ai facilitates the creation of AI brokers by its decentralized platform.
Planck says its blockchain will generate protocol income from transaction charges, SDK utilization, and developer tooling. In the meantime, GPU operators offering compute assets might be rewarded with the protocol’s native token primarily based on machine’s uptime (proof-of-connectivity) and precise utilization (proof-of-delivery).
Associated: AI advances audit protection for progressive multi-language digital machine
GPU rental and AI chip crunch
Most of Planck’s income comes from renting out GPUs and compute contracts. The corporate says its hourly, on-demand mannequin slashes prices by as much as 90% in comparison with conventional cloud providers.
Since February, the corporate generated $1.5 million in income, largely from renting GPU energy. The corporate additionally competes with different infrastructure suppliers like Huge.ai, CoreWeave and Lambda, all of that are capitalizing on the continuing AI chip scarcity.
That scarcity has fueled explosive progress within the GPU-as-a-service sector. In response to Priority Analysis, the market hit $4 billion in 2024 and is anticipated to develop at a 23% annual fee, reaching $32 billion by 2034.
“Centralized AI compute is prohibitively costly. By decentralizing the GPU community, Planck can provide as much as 90% value financial savings,” a Planck spokesperson mentioned.
Journal: AI Eye: AI is sweet for employment says PWC — Ignore the AI doomers
