Recently, the buzz around AI cryptocurrencies has been really hot. The market cap has reached $30.6 billion (about 4.7 trillion yen), and it feels like the speculative phase is over. Major asset management firms like Grayscale are creating AI-focused funds, so institutional investors have already begun to seriously enter the space.

The reasons why AI × Blockchain is gaining attention can ultimately be boiled down to three points. First, decentralized GPU computing. Currently, AI development is concentrated in large companies like NVIDIA and Amazon, right? Using blockchain, we can utilize idle GPUs worldwide. Second, data transparency. We can build systems that properly protect creators' rights. And third, censorship resistance through decentralization. When these elements come together, AI democratization can accelerate rapidly.

Looking at the entire sector, AI agent-related projects amount to about $3.18 billion, and AI framework-related projects are around $1.78 billion, forming markets in the tens of billions of dollars in each category. The evolution of generative AI and the surge in GPU demand are fueling this growth.

As for specific projects attracting attention: Render Token (RENDER) is central to decentralized GPU rendering, currently at $1.80. NEAR Protocol (NEAR) is an AI-compatible Layer 1 blockchain, at $1.22. Virtuals Protocol (VIRTUAL) is rapidly growing as a platform for issuing AI agents, at $0.67. SingularityNET (AGIX) is a decentralized AI learning network, holding the largest share in Grayscale’s fund, currently at $315.30. Other projects like Filecoin (FIL) for decentralized storage, The Graph (GRT) for data indexing, the FET AI platform, Akash Network (AKT) for cloud computing, Worldcoin (WLD) combining biometric authentication and AI, and ChainOpeRAI (COAI) for on-chain infrastructure each have their own unique value.

The AI agent topic is especially interesting. Traditional AI is passive, just answering questions. But AI agents are autonomous, active entities that execute tasks on their own. As a CEO of a major exchange announced, the era of AI bots trading automatically via cryptocurrency wallets is already here. Major venture capital firms also see this as a key trend by 2026.

However, looking at actual incidents, there are many challenges. In February 2026, an autonomous AI agent mistakenly sent $250,000 worth of meme coins—intending to send $4 but ended up sending all assets. This teaches us that safety mechanisms and monitoring systems are essential for AI agent wallets. There have also been incidents like $1.78 million being drained from DeFi protocols, malware infiltration cases, and other security risks becoming apparent.

Regarding risks, the DeepSeek shock is notable. In January 2025, a Chinese AI company released a low-cost, high-performance model, causing AI-related cryptocurrencies to plummet. This shows that AI cryptocurrencies are highly sensitive to shifts in the AI industry landscape. Additionally, some projects tout AI in their whitepapers but haven't reached practical implementation. Simply listing AI terms without real technological integration is common.

When making investment decisions, it’s crucial to verify the development team’s real names and backgrounds, the status of source code on GitHub, whether they are listed on reputable institutional investor lists, and whether the actual product is operational.

For those in Japan looking to buy AI-related tokens, Render Token, NEAR Protocol, Filecoin, The Graph, and Worldcoin are available on Bitbank, SBI VC Trade, GMO Coin, and OKCoin Japan. Since you can purchase directly in Japanese Yen, they are beginner-friendly. Tokens like Virtuals Protocol and SingularityNET, which are not listed domestically, require using overseas exchanges or DEXs—buy Bitcoin or Ethereum, transfer, and then purchase the target tokens from there.

The AI cryptocurrency sector is definitely expanding, but unexpected fluctuations like DeepSeek can occur, and security risks are real. It’s safest to diversify investments across multiple tokens within your risk tolerance. Conduct thorough research on the technical details before making decisions.