XiaoTong Column · 2025-08-11

Chain Exploration”The Rise and Fall of Encrypted AI: What Remains After the Bubble Bursts?”

The Rise and Fall of Crypto AI: What Remains After the Bubble Bursts?

Solana AI Hackathon Lead Yash: AI Agent Infrastructure Narrative is Dead, but Crypto × AI Megatrend is Inevitable

Editor’s Note

Yash, head of SendAI (organizer of Solana AI Hackathon), believes we are approaching the end of the first phase of the crypto x AI agent hype cycle. With most AI agent tokens down over 95%, now is the perfect time to reflect and prepare for what’s next. In this article, Yash reviews lessons learned so far and predicts key trends for crypto x AI agents in the next 6-12 months and beyond.

Phase 1: Speculation, Bubble, and Experimentation

It all started with Truth Terminal ($GOAT) in October 2024. Over the next three months (November 2024 to January 2025), crypto AI reached a fever pitch, with AI agent token market caps exceeding $10 billion.

While this phase left few practical products, it offered glimpses of potential:
– Weekly new experiments (often with tokens hitting $50M market caps before crashing);
– Crypto AI frameworks and launchpads hyped by KOLs as “L1 for AI agents”;
– ChatGPT-style UIs like Griffain, Venice, and Wayfinder hitting $500M+ market caps, touted for on-chain execution (but only demos, no real progress);
– Promises of “autonomous hedge funds” (nothing but whitepapers).

Phase 2: The Crash (February-April 2025)

  • Liquidity Drain: TRUMP token issuance siphoned capital from AI, causing most tokens to drop 50-90% in market cap;
  • Development Stagnation: Founders of $1M to $100M token projects nearly stopped releasing demos;
  • Team Dissolution: Over 90% of teams quit due to lack of incentives or low token value;
  • Community Deterioration: Unrealistic expectations for token appreciation, with no intrinsic value (only derived from issuing more token derivatives);
  • Top Projects Collapse: Zerebro suspended abruptly (founders went silent), Griffain stopped demo updates, and ai16z pivoted to Memecoin launchpad (auto.fun);
  • Infrastructure Narrative Demise: Frameworks and launchpads proved useless without mass-market AI agents.

Solana AI Hackathon: Struggles and Achievements

As a catalyst of the hype cycle (via Solana AI Hackathon), Yash admits token crashes damaged the sector’s credibility, hampering progress. Yet the team still delivered:
– The first open-source Crypto Multi-Control Protocol (MCP) server for execution;
– Upgraded Solana Agent Suite Architecture (v2) for modularity;
– Expanded protocol integrations from 11 at launch to 50+ official ones.

Current Plight of Crypto × AI

Chat Frontends: All Show, No Go

Over 40-50 teams built Griffain-like UIs, but production reality is grim:
– Models lack context for crypto/Solana transactions; simple tasks like “swap 10 SOL to USDC” take 8-10 seconds—far slower than direct UI操作;
– Demos look great, but no real utility solved.

AI Agent Infrastructure Narrative is Dead

  • Open-source Frameworks Failed: Couldn’t become platforms without monetization or value accumulation;
  • Launchpads Homogenized: New “agent launch platforms” lack differentiation, reduced to meme tools;
  • Vibe Coding Immature: Tools like dev.fun and Poof (Solana-focused) show promise but face security hurdles for production use.

Solana AI Token Carnage

Tokens like ai16z ($150M), alch ($140M), goat ($100M) crashed from peaks, with most developers inactive. Over 147 AI agent tokens on Solana still have >$1M market cap, but Crypto AI 1.0 fair-launch tokens are largely dormant.

Future Hopes: 6-12 Months and Beyond

6-12 Months: From “Functional” to “Useful”

1. Smart Agent Chat Optimization

  • Model Improvements: Claude Sonnet 4, Kimi K2 enhance tool调用; ChatGPT and other base models focus on proactive operations;
  • MCP as Standard: Multi-Control Protocol (MCP) supports data reading, transactions, and prompt execution—likely the foundation for future agents;
  • Complex Task Breakthrough: Not for simple swaps, but transaction workflows could be killer apps.

2. Consumer-Grade Crypto × AI (Vibe Coding + Token Launch)

AI lowers content/app creation barriers; tokens solve attention scarcity. Examples:
– Gamified TikTok: Users build tokenized games;
– Character Interaction Platforms: Like Grok’s Ani, where users earn tokens by chatting with AI characters.

Beyond 1 Year: Deep Integration of Crypto and AI

1. Agents as Stablecoins’ “Trojan Horse”

  • Payment Protocol Integration: MCP may embed standards like Coinbase x402, enabling payment via API calls;
  • Big Tech Adoption: Stripe acquired Bridge/Privy and launched agent kits, betting on stablecoins as AI agent payment首选;
  • Business Model Innovation: Ads shift from explicit to implicit—agents earn rewards for ad/promotion engagement (pay-per-conversion).

2. Crypto Protocols Go All-In on AI

  • Contextual AI: Embedded in DeFi strategy suggestions (e.g., loop arbitrage) and asset issuance;
  • Example: Jup Studio’s magic gen uses AI to generate images and token code.

3. Crypto Economic Networks for AI

  • Verifiability Key: Crypto must solve AI training/inference verifiability;
  • Value Network Example: Bittensor (TAO ecosystem) built AI value chain financial engineering, reaching $4B market cap (though overvalued);
  • Community Capital: Crypto enables capital coordination for AI communities (e.g., cyberpunk movements).

4. Composable Personal Context Layers

  • On-chain Memories: User preferences/tone stored cryptographically (e.g., zero-knowledge proofs) for LLM access, enabling personalization;
  • Context Trading: Users monetize/authorize context (retaining ownership), potentially more valuable than IP.

5. Chat-Based Crypto Super Apps

  • UI Revolution: From web navigation to AI agent execution—all platforms/protocols as tool calls;
  • Browser 2.0: Agent browsers like Perplexity Comet, Arc Dia rise; AI-native browsers unlock design potential;
  • Example: Donut builds crypto-specific agent browsers; raycast.sendai.fun previews chat-based crypto interfaces.

Conclusion

Yash emphasizes that despite the hype cycle trough, crypto-AI convergence is inevitable. Standing at the intersection of crypto and AI—two transformative technologies—it’s time to prepare for Crypto AI’s second wave.

Read More《加密AI兴衰启示录:泡沫褪去,留下哪些火种?》

This content is AI-generated and does not constitute investment advice. Please exercise your own rational judgment.

创业时评《从“向上”转向“守住”,年轻人的成长变成伪命题了吗?》

ZhiXing-AIx
Chatbot