Author: 0xJeff
Compiled by: Tim, PANews
Since the publication of the first article on mainstream AI trends in May, there have been many changes in the market:
Market sentiment and the number of AI agent tokens for fair launches have both declined
Narrative shift
New and old participants in the ecosystem are undergoing a transition
The value proposition of the Web3 AI ecosystem is becoming clearer
Institutions and whales face more AI investment opportunities
Shifting towards real products and applications, rather than hype
Others
In this article, we will review and summarize the dynamics of the past few months, analyze current trends, outline market leaders, evaluate industry development directions, and discuss how to best prepare for future trends.
Let's get started.
Since the beginning of this year, life in the on-chain trenches has been tough.
First: We have too many tokens, which dilute liquidity and attention.
Secondly: There is too much nonsense on crypto Twitter and AI junk content, which distracts people's attention from quality content.
Third: The pump is not as strong as it used to be and will continue to decline.
Fourth: Game version iterations are very fast, and if you are not constantly paying attention or have the right information channels, it is difficult to keep up with the pace.
Looking back at Q1 and Q2, the Trump family's Meme coins, celebrity coins, and popular projects like ICM, creator coins, and Pump ICO attracted most of the active liquidity in the market, leaving little funding for the AI sector.
AI Agent Ecosystem: From Concept Hype to Product Landing.
For this reason, good opportunities have become extremely rare, and the top AI Agent project Virtuals, with the most developers, has become a highly anticipated ecosystem. The Genesis launch project performed quite well, with occasional high-return projects achieving 5-20 times returns.
However, this situation only lasted for 1-2 months because, after all, the focus of this ecosystem is on hyping new projects. When there is a lack of exciting new projects launched, the attention and hype will fade, and 'points' will become worthless.
This shift in sentiment has prompted investors and speculators to become more selective, focusing on product substance rather than market hype.
Two major AI Agent trends are emerging:
1. Investors will shift their attention from the Virtuals Genesis launch projects to selectively investing in high-quality Virtuals agency teams.
Some teams launched their tokens earlier but faced enormous selling pressure, and now they will introduce entirely new important products or features.
Some new teams are launching more refined products on Virtuals. They are not building minimal viable products in two or three months; rather, they are launching products immediately after issuing tokens.
These include (but are not limited to) the following teams:
ArAIstotle: AI Fact-Checking Platform.
PredictBase: Decentralized Prediction Market.
Mamo: Personal Finance Assistant (developed by the Moonwell team).
Billy Bets: Sports Prediction Agent, Engine, Aggregation Platform.
Backroom: AI-driven SocialFi Application.
The next trend is likely to continue, that is, high-quality projects will still be scarce, but if you closely follow the dynamics of promising AI Agent teams on Virtuals, the probability of capturing the next 10-50 times return project will significantly increase.
2. Investors will shift their attention from Virtuals to other more promising AI agent ecosystems.
CreatorBid, with its high-quality listing strategy, successfully launched AI products with clear practicality and actual use cases through the Bittensor subnet, winning mind share and favor among investors.
Additionally, the team focuses on supporting existing teams through products, partnerships, and marketing, and CreatorBid has successfully positioned itself among the leading AI agent ecosystems.
Players refer to the launch platform launched by Holoworld as a lottery-style launch platform (HoloDraw), where users can participate in the lottery by purchasing tickets (each ticket represents 0.5 SOL in purchasing power on the bonding curve). 35% of the token supply will be reserved for HoloDraw lottery, and if users do not win, they can also participate in a second lottery for consolation prizes (accounting for 5% of the supply). HOLO successfully launched on Upbit with a fully diluted valuation of $1.5 billion, sparking enthusiasm in the community.
Other projects like OpenServ focus on incubating teams, aiming to launch products, attract users, and achieve revenue first, and then issue tokens when the timing is right; meanwhile, OpenServ continues to develop its decentralized n8n product for consumers.
Next: Top agent projects on Coinbase will continue to seek product-market fit in the prediction track, while Coinbase continues to focus on enhancing the value accumulation of BID (which is their current pain point).
Other ecosystems continue to adopt a product-first strategy; their tokens may face impacts, but once they launch flagship use cases, products, or teams, the token valuations will inevitably rise.
Key ecosystems to watch: CreatorBid, OpenServ, Holoworld, Arc, Loomlay, ElizaOS.
Apart from AI agents, capital is concentrated in a few narratives and vertical fields:
Decentralized AI: From primitive intelligence to intelligent productization.
DeAI is still the favorite sector for whales and institutional investors.
Decentralized computing remains the highest revenue segment. Aethir's annual revenue reaches eight figures, and Chutes processes 50-100 billion tokens daily through its open routing platform.
Federated learning-based privacy-preserving AI technology is accelerating its landing. Flock has partnered with Web2 companies and government agencies to build privacy-preserving vertical field AI solutions while achieving product-market fit and token-market fit. Recent cooperation cases include the United Nations Development Programme, the Hong Kong government's official AI service provider, and CIMG.
Darwinian AI ecosystems are rapidly developing. Bittensor has expanded to 128 subnets, Sentient has launched the world's largest collaborative intelligence network GRID, and is building a Darwinian AI ecosystem using the 'artifact' subnet architecture.
Projects like FractionAI apply Darwinian AI to the gaming field, allowing users to deploy AI agents for competitive contests, covering various scenarios such as Polymarket predictions, agent battles, tic-tac-toe, football, and more.
Predictive AI is approaching the industrialization stage. Multiple Bittensor subnets have proven that large-scale decentralized intelligence creation can significantly enhance model performance: the SN18 Zeus network, SN44 scoring network, and SN50 synthesis network have all surpassed benchmark tests and state-of-the-art models. Among them, SN44 monetizes signals through Sire sports betting strategies (top sports hedge funds deploy $300 million using its strategies).
One of the latest fields in DeAI is data, mainly including data labeling, human feedback reinforcement learning (RLHF) data services, and evaluation (yes, this is the first and second fields with the highest investment and adoption value in Web2).
Data labeling and human feedback reinforcement learning (RLHF) is a labor-intensive task. Web3 allows participants from any region to engage in labeling, annotating, rating, feedback, and other tasks through token incentives, thus reducing related costs.
The main participants in this field include: SapienAI, FractionAI, PerleAI, PublicAI, SN52 Dojo, Synesis One.
DeFi x AI (DeFAI): From proof of concept to fully autonomous financial agents.
The development history of DeFAI: from the abstract layer and the failed landing of GPT-like interactive interfaces in Q1, to the personalized agent proof of concept that assists users in managing funds in Q2, and then to the full launch of a fully autonomous financial agent system in Q3, building a highly scalable and verifiable infrastructure.
Rapid technological advancements, with projects represented by players like Giza leading the entire track. Giza has achieved nearly $2 billion in agency trading volume and manages $20 million in agency assets around the clock.
The newly launched Swarm Finance + Pulse (Pendle agent) marks the beginning of a new era where agents optimize idle funds for users.
Almanak closely follows, adopting highly scalable and verifiable infrastructure, using agent pools and smart contracts (asynchronous tokenized treasury) to build a system architecture that does not directly manage user funds.
This can completely shield investors from illusions and AI hackers, vulnerabilities, creating an environment where institutions and whales can safely deposit large amounts of funds. Almanak's quantitative creation platform also allows anyone to launch quantitative strategies in minutes, without taking weeks of time.
Progress of DeFAI use cases:
Key players to watch: Giza, Almanak, Cod3x, Theoriq.
What are the next steps for development?
Predictive AI is one of the fields with the best asymmetric betting opportunities. The total addressable market (TAM) for prediction markets far exceeds the Web3 realm, as users from various industries log onto platforms like Polymarket and Kalshi to make predictions on topics of interest.
AI and tools built on prediction markets will be able to capture tremendous value from the growth of this field. In particular, consumer applications with user funnels and alternative prediction markets and agents will stand out. Liquidity remains the biggest obstacle, however, the emergence of AI market makers and treasury products is likely to help improve liquidity conditions across markets.
Data will be the focus in Q4 of this year. As 'useful' data becomes increasingly scarce, Web3 data service providers will face stronger demand from Web2 and Web3 AI labs. Data is the largest moat and bottleneck.
AI agents and the AI agent ecosystem will continue to attract the attention of on-chain players and the crypto Twitter community due to their simplicity and fair launch characteristics. High-return opportunities will become increasingly scarce, but high-quality projects that stand out may still achieve 10-50 times growth (the liquidity of such projects will still be quite weak, so please act with caution).
DeFAI will evolve to a stage where top players will become the key driving force behind top DeFi protocols (AAVE, Pendle, Fluid, Uniswap, Aerodrome, etc.) with massive TVL and trading volumes. Covering more DeFi application scenarios, achieving more complex strategies, providing better execution capabilities, and building safer risk barriers and infrastructure.
Conclusion.
The brutality of the battlefield will continue, many tokens will perish, attention will be diverted, and some people will leave.
However, amidst this chaos, the foundation for the next wave of Web3 AI is being laid right now.
The leaders of the next wave will be those who:
Solving real problems.
Attracting highly sticky users.
Creating products that can withstand narrative shifts.
Ironically, the biggest winners initially seem plain (boring infrastructure, privacy, data), until one day they suddenly are no longer plain.
