Original Article Title: AI Trends in the Trenches
Original Article Author: @Defi0xJeff
Original Article Translation: zhouzhou, BlockBeats

Editor's Note: The AI Agent project attracted short-term attention due to a Genesis pump, but institutions are more focused on decentralized AI infrastructure such as $TAO, $GRASS, etc. Market trends indicate that teams combining narrative and technology have more long-term potential. The strategy is to quickly flip Agent projects for profit and roll that into DeAI infrastructure, betting on teams that can tell a good story and deliver on the technology.

Below is the original content (slightly reorganized for better readability):

From the bottom of the AI agent race track market (with a market cap of about $4 billion) to now, it has been approximately two and a half weeks, and the market is currently in a full-blown bull run led by @virtuals_io.

That's right... only Virtuals are currently surging. This rally is very similar to last October to November when Virtuals just launched their agent tokenization platform, establishing themselves as the pioneer of AI project "tokenized issuance" and the first legitimate player to offer a top-tier issuance network (provided you were willing to have a "fair launch").

What's different this time is that they have just launched a new feature: "Genesis Launch" — a more equitable distribution method that also rewards early supporters. This leads us to the first trend we want to discuss:

Fair Launch / Gamified Launchpad

From the days when degenerate players would randomly rush into some nonsense tokens on pumpfun to now almost guaranteed 5-10x returns on brand-new projects.

Behind this is the introduction of a "point" mechanism that better aligns the interests of all parties, along with a fixed market value, fixed supply emission model — where each project starts with 112,000 $VIRTUAL tokens (equivalent to approximately a $200k FDV). Each participant can receive up to a 0.5% token share based on the points they earn. Ways to earn points include: holding/staking $VIRTUAL, holding mainstream agent tokens (they have an official list), and even being active in discussing Virtuals on social platforms.

Recently, Virtuals has also introduced a "Cooling-off Period" mechanism to restrict rapid sell-off behavior, further enhancing the attractiveness of the Genesis Launchpad—making you think twice before selling.

Currently, this "Genesis Launch" platform has already achieved significant success, with @BasisOS being the most successful example so far, where participants have received returns of up to 200x. Subsequently, new projects launched on Virtuals have seen returns ranging from 5 to 40 times.

It is precisely because of the explosion of Genesis Launch that funds and attention have once again been concentrated in the Virtuals ecosystem, raising the valuation floors of almost all projects on the platform.

However, despite the reignition of market enthusiasm, the "scarcity of high-quality projects" remains one of the biggest challenges in the Virtuals ecosystem—leading to the emergence of the second trend:

2. Hype and Trading Structure Trumping Project Quality

As a speculator/trader, as long as you understand the structure of Genesis Launch, you can make money even from mediocre projects with average teams.

Because projects starting from a $200,000 valuation are very likely to pump first and then dump, especially when you know the project doesn't have any "skeletons in the closet" and hasn't conducted a private fundraising round.

As a project team, if you have a relatively novel idea, you can simply launch the token first and figure out the rest later.

No need to develop the product first, no need to understand who the target users are, no need to validate if there is a genuine market demand, let alone revenue and user growth. Just create maximum hype and go full steam ahead (having a demo is a bonus, but not necessary, lol).

As long as the project meets the basic requirements (a decent whitepaper, a coherent product concept, a team that doesn't look too shabby), it can pretty much take off smoothly on Virtuals' Genesis Launch.

For investors, the most important lesson here is: consider these launches as short-term speculative opportunities, not as medium- to long-term value investments.

Because most of these so-called "AI projects" are fundamentally just packaged shams pretending to be AI. I even provided a very typical example on my own Substack, so feel free to check it out if you're interested.

But precisely because there are now too many shitters, the opportunities for truly high-quality projects have emerged—whether in the AI track or in directions outside of AI. This also leads to the third trend:

3. The Potential Opportunity of DeFi Projects on Virtuals

Two months ago, I chatted with the @logarithm_fi / @BasisOS team. Their product is somewhat like Ethena, doing something similar to a "Delta Neutral Strategy," but without the stablecoin part.

I got to know this team when they were working on Logarithm, back when they were developing an LPDeFi narrative (a product using Uniswap v3 liquidity for strategies). I quite liked this team, so I gave them some tokenomics and listing plan advice.

To be honest, I didn't originally think their token would increase by this much because while their DeFi product had fundamentals, the so-called "AI" part was actually very early-stage. But the results proved that none of this really mattered—this project still wiped out the entire market.

From Basis' success, it is clear that the Virtuals platform has indeed provided a significant opportunity window for DeFi projects.

Even if you don't have the traditional DeFi project's "token incentive TVL" mechanism, as long as your product is solid and the logic is strong, coupled with Virtuals' traffic and sentiment, the TVL naturally increases.

So if you have a mature DeFi product and want to enhance user acquisition efficiency through AI, or optimize the product's structure with AI technology, feel free to reach out to me for a chat. I am very willing to brainstorm together.

Aside from the "getting market attention" layer, a new experimental gameplay is now emerging on Virtuals, much like the 2023-24 Ethereum shitter token craze—this also leads to the fourth trend:

4. Using Protocol Revenue Generated by "Trading Volume" as a Growth Engine

There was a time when many small-scale Ponzi DeFi projects emerged, crazily experimenting with various token models: for example, charging a 1-3% fee on each transaction, using this revenue to grow the project's treasury, supporting their Ponzi DeFi strategy, and then returning profits/dividends to token holders.

Back then, there weren't many projects on Ethereum, and users were also "rich and daring." Using this playbook, project teams could earn a six- or even seven-figure income in just one or two weeks.

Now, we are starting to see similar things happening in the AI agent circle.

Every transaction on Virtuals incurs a default 1% transaction fee, with 70% of it being returned to the project team. Some other Launchpads also have a 1-2% fee mechanism, with return ratios ranging from 70% to 100%.

For example, recently, on the @autodotfun platform, the @Squidllora project (backed by @AlloraNetwork's smart support) used this creator fee portion to expand its treasury, then used this money to trade mainstream coins, with a strategy relying on the reasoning model provided by Allora (if you're not familiar with Allora, its positioning is somewhat like a "purely financial version of Bittensor"—where scientists compete to create the strongest encryption prediction model covering various time dimensions).

A portion of the profits earned from transactions is used to buy back the $SQUID token.

The advantages of this model are very obvious, especially suitable for well-funded projects that do not rely on fees to sustain the team.

The project team can treat the AI agent token as a marketing tool, a user diversion tool, to attract attention, accumulate fee income, and thereby initiate a whole new AI experiment.

However, looking at the entire battlefield of AI agents, apart from the Virtuals ecosystem, other platforms are basically unable to compete. Many teams, even if they develop openly and continuously deliver, still do not see a price increase.
This also brings us to the fifth trend:

5. Virtuals Dominates Alone, While Other Platforms See Little to No Growth

Due to the push from the Genesis Launch, the valuation floors of Agent projects on Virtuals continue to rise. However, it is important to note that the emergence of these new projects does not necessarily indicate a significant improvement in fundamentals or technology. The true catalyst is the massive optimization of the transaction structure—in other words, more people are now willing to play Virtuals games because they know they can make money.

This trend may continue until the price of existing projects reaches a local top or ceiling.

Once the upward movement stalls, some attention will naturally begin to overflow and shift to other ecosystems, such as @CreatorBid, @arcdotfun, @autodotfun, especially projects with lower market capitalization but significant fundamental improvements (new features, products, partnerships).

Currently, CreatorBid (CB) and Arc are the two most noteworthy ecosystems, with several undervalued projects that have yet to start moving: such as the 3-4 targets on CB that are integrated with subnets or focus on Bittensor product development;

or some handshake projects on Arc directly related to Ryzome, which are also worth keeping an eye on.

The best way to position oneself is to take advantage of the low-key layout of these tokens before everyone else notices their value.

6. For institutions, AI investment opportunities are still very limited

Despite the significant price increase of main AI Agent ecosystem tokens such as $VIRTUAL and $AI16Z, many institutions are still observing from the sidelines for a simple reason: these tokens that have surged are only suitable for retail traders/degens, with scarce liquidity and a fragile LP structure (especially on Virtuals).

It is for this reason that—due to a lack of mature liquidity infrastructure + a rising interest in decentralized AI—institutional funds are starting to turn toward investing in AI infrastructure, Agent L1, and chain-level AI labs, rather than the current batch of short-term speculative Agent project tokens.

You may wonder, what are these institutions buying?

$GRASS

$TAO (including subnets)

$VANA

$FLOCK

$PROMPT (perhaps)

There is also a group of projects that are not yet launched but have a clear moat, such as @NousResearch, @PluralisHQ, @PrimeIntellect

Most of these projects are working on the Web3 transformation of truly high-performance models, decentralization of model ownership, not the kind of "shell" decentralized GPT, but truly complex infra + model layers, with a high barrier to entry, incomprehensible, and no one knows how to invest.

How to Better Position Oneself Based on These Trends?

My own strategy is simple: take the profits earned from short-term AI Agent projects (especially those with a strong Genesis but a weak team) and gradually roll them into DeAI Infrastructure projects.

It's important to note that these infrastructure projects will not experience a short-term explosion because most of them do not have consumer products and are slowly building at the Infra layer.

However, just like how OpenAI, Grok, Anthropic suddenly evolved in tasks such as day-to-day operations, real-time retrieval, and code generation, Web3 models will also experience a qualitative leap at some point—a true capability to handle native tasks of Web2/Web3.

So, Should We Overweight in DeAI Infra?

Not exactly. Crypto is essentially a narrative- and distribution-driven market.

Distribution ≫ Technology, this is a "90% storytelling + 10% model" game.

A project that can tell a story, has good UI/UX, knows how to go live, how to design tokenomics, and how to attract and retain users, can thrive.

Therefore, my core investment logic remains: invest in teams that understand the dual-drive of "narrative + technology."

It's like how Web2 VCs would invest in some vertical SaaS domains, even though they might be using open-source models in the background, relying on their data and operations to maintain strong product barriers.

I believe this logic still holds in Web3 in the short term, especially now when hype and community are the main driving forces, where easy-to-understand concepts are easy to sell.

The market is also moving in this direction:

· Genesis Launch is becoming an "industry default"

· More and more teams are starting to collaborate with Infra projects, not just for speculation but also to actually use each other's technology

· While working on distribution, they also work on real AI technology, blending storytelling with action

Quick Summary

Virtuals' Genesis Launch model is currently dominating market attention and profits;

Most projects prioritize hype over technology, focusing on short-term trading at their core;

DeFi projects with strong fundamentals have found incremental opportunities in the Virtuals ecosystem;

Trading volume drives creator revenue, triggering a new wave of experiments;

Virtuals currently lead the way, but other ecosystems like CreatorBid and Arc may see a surge soon;

Institutions are still off-chain, with more focus on decentralized AI infrastructure bets;

The optimal strategy is: roll short-term gains from Agent projects into DeAI Infrastructure and bet on teams that understand both narrative and technology.

Original tweet: 「Tweet Link」

Welcome to join the official BlockBeats community:

Telegram Subscription Group: https://t.me/theblockbeats

Telegram Discussion Group: https://t.me/BlockBeats_App

Official Twitter Account: https://twitter.com/BlockBeatsAsia