"Web3 and AI Convergence... Decentralized AI Infrastructure Emerges as New Trend in Blockchain Industry"

Dispread Research, the research division of Web3 consulting firm Dispread, announced on the 21st that decentralized artificial intelligence (AI) infrastructure will facilitate stable AI adoption and commercialization of blockchain-based projects.

With recent advances in AI technology, cases of AI integration are increasing in the blockchain industry sector. According to virtual asset data analytics firm Messari, investment in the AI-blockchain convergence sector increased by 100% year-over-year in Q4 2024, with the number of investment rounds rising by approximately 138%. This growth trend indicates that attempts to utilize AI technology are becoming more active in the blockchain industry.

The Dispread Research team stated, "While various industries and small startups are attempting to adopt AI, they are experiencing difficulties in the process," adding, "This issue is particularly prominent in the blockchain industry." They continued, "Since Web3 must maintain trustlessness (one of blockchain's core values where third-party trust is unnecessary), the development of decentralized AI infrastructure is essential for more protocols to stably adopt AI and provide services that users can trust."

AI network platforms providing decentralized AI infrastructure include Allora, Bittensor, and Gensyn.

First, Allora is a decentralized inference synthesis network that provides optimized predictions for specific situations. On Allora, AI model operators can freely execute predictions on specific topics, and the protocol (Allora) synthesizes inference values from individual models to produce final prediction values.

Allora encourages AI models to participate in the protocol through incentives and an open-source framework (Allora MDK). It assigns higher weights and incentives to models that produce more accurate inferences by comparing actual results with each model's inference values, and allows anyone to easily build and deploy AI models through Allora's open-source framework. The goal is to improve inference accuracy by attracting more models to the protocol to demonstrate a kind of 'collective intelligence'.

Previously, Allora established partnerships with Virtual Protocol, Story Protocol, and Monad ahead of its mainnet launch. The research team analyzed, "Web3 projects have high demand for decentralized inference," noting, "The ability of AI models to perform customized inference using on-chain data for each project appears to be a key factor."

Furthermore, Bittensor is a decentralized AI network that transforms the corporate-centered AI development structure into an open network. Unlike the existing industry where single companies like OpenAI and Microsoft develop their own AI models, Bittensor provides an environment where anyone can develop and participate in AI models.

Bittensor's system consists of multiple subnets. Each subnet operates according to specific functions such as natural language processing, image generation, and financial data analysis, with AI models competing within these subnets to produce the best results. High-performing AI models are evaluated based on their network contribution and receive incentives.

Meanwhile, Gensyn is a network designed to decentralize machine learning computation so that anyone can provide and utilize computing power. The network consists of solvers performing machine learning computations and validators verifying them, supporting efficient computation verification based on 'probabilistic learning proofs'. Gensyn processes AI learning in a distributed manner using idle computing resources worldwide and rewards participants who contribute to the network.

Byungjun Kim, a Dispread researcher, stated, "Blockchain technology can solve various current AI problems such as biased judgment or monopolization of the AI market by specific companies," adding, "Cases where blockchain receives attention as a core infrastructure for the AI industry will continue to increase."