Grayscale Highlights Top 20 AI Tokens

According to the report, it includes 20 tokens focused on decentralized AI protocols. The total market capitalization of the segment is $21 billion, compared to $4.5 billion in Q1 2023. This new sector becomes the sixth in the Grayscale Crypto Sectors system, alongside Currencies, Smart Contract Platforms, Financial Protocols, Consumer and Culture, and Utilities.

Grayscale considers artificial intelligence one of the most significant technologies of the 21st century, yet its development remains concentrated in the hands of a few corporations.
Decentralized blockchain-based solutions promise to shift this dynamic. The growing number of such projects over the past two years prompted the creation of a separate category.

How the AI Sector Is Structured

According to Grayscale, the new sector consists of three categories:

• AI Platforms: Infrastructure for developing AI applications, such as Bittensor and Near
• AI Tools and Resources: Access to data and computing, like Grass (data scraping) and Akash (decentralized computing)
• AI Applications and Agents: User-facing solutions including Virtuals and ElizaOS

The top asset by market cap in the sector is TAO from Bittensor. It follows the Bitcoin model, with a supply cap of 21 million and halving events every four years. The first halving is expected this year. Since the dTAO upgrade in February 2025, which enabled investment into subnetworks, the share of TAO allocated to them has grown to 7%.

Near has also been included in the sector. It is developing as a blockchain for AI, focusing on scalability, cross-chain transactions, and AI agent functionality. Examples include a chatbot assistant and multi-chain AI agents on the platform.

Some major crypto projects with AI potential have not yet been included in the list. For instance, Filecoin and Arweave, which offer decentralized data storage, are not part of the new sector because AI is not their primary use case.

Growth and Outlook

Since the beginning of 2023, the sector has grown more than fourfold. The best-performing token is TAO (+2%), while the worst is ElizaOS (-80%). Grayscale highlights growing interest in distributed training: Prime Intellect demonstrated the ability to train models with 30 billion parameters on distributed GPUs. Such approaches can reduce costs and resource demands for AI training.

Grass generates tens of millions of dollars annually from data sales to AI labs. Virtuals earns $30 million per year from fees related to AI agent trading. According to experts, this data indicates the maturity of some projects within the sector.

Stablecoins and AI Agents

Grayscale is also tracking the growing use of stablecoins in AI applications. Meta has announced its entry into the market, Stripe acquired a stablecoin company for $1 billion, and Coinbase introduced a payment standard for AI agents. Legislative initiatives such as the GENIUS Act are expected to further expand the application of such solutions.

This post is for informational purposes only and is not an ad or investment advice. Please do your own research making any decisions.