Viewpoint: Why does the AgentFi track still have 10 times room for growth?

Author: @xingpt

Before analyzing the AI Agent track, let's first step back and take a holistic look at what crypto has experienced in this cycle:

Bitcoin Decouples from the Crypto Sphere

In the previous Bitcoin cycle, Bitcoin was almost synonymous with the crypto sphere, where buying Bitcoin meant buying Altcoins, and recognizing crypto and decentralization.

However, after the approval of the Bitcoin spot ETF, from the US President to listed companies, the mainstream opinion seems to have recognized the value of Bitcoin; but the raison d'etre of Crypto, especially Ethereum and Altcoins, has not yet gained mainstream social and capital recognition.

The reasons are complex.

The main reason is the asset positioning: Bitcoin is seen as an alternative asset linked to gold, widely recognized for its inflation-hedging and asset-preserving properties beyond sovereign currencies.

However, in the eyes of Wall Street old money, Ethereum and other Altcoins are still tech meme stocks without mature and sustainable business models. Compared to tech giants like Nvidia, Microsoft, and Amazon that have users, products, and demand, the valuations of Ethereum and other crypto projects are not low, and the risk-return ratio is extremely low from an asset allocation perspective.

As shown in the figure below, Ethereum's Sharpe Ratio is lower than that of Meta, Google, and other tech companies. While Bitcoin's Sharpe Ratio is second only to the super-god Nvidia in this cycle.

Another important factor is that the overall macro interest rate and monetary easing are still not comparable to the massive liquidity injection during the pandemic in the previous cycle, coupled with the booming AI industry, resulting in a lack of attraction of Crypto to off-chain capital. The logic is very simple - there is only so much money, and it has been spent on buying AI stocks and GPUs, leaving no room for Altcoins and Ethereum.

The US M2 money supply has not yet recovered to its 2022 high (data source: CEIC)

Extreme Imbalance in the Crypto Ecosystem's Internal Circulation

Since the crypto sphere's ability to attract off-chain capital is insufficient, can the internal capital be leveraged to drive enough purchasing power?

Roughly estimating the total internal capital using the total stablecoin supply + contract holdings, we can easily find that the current internal capital is far beyond the previous bull market. But except for BTC, most Altcoins have not reached new highs. What's the problem?

The root of the problem lies in the supply-demand imbalance. The supply side is a large number of high-valuation financed new projects, most of which have not found real product-market fit (PMF) and have few real users.

The existence of these projects is due to the over-financing by crypto VCs in the 2022 bull market. Due to the excessive capital raised by crypto VCs, with most of it limited to 5 years, 3 years of investment + 2 years of exit period, the funds, in order to complete their investment tasks, have ignored the quality of the projects and invested frantically.

So who will provide the buying power for these projects?

Previously, the main exit channel was centralized exchanges, but after the FTX incident, centralized exchanges have become the target of criticism. The long-arm regulation of the eagle sauce has made centralized exchanges suffer, not only fearing huge fines, but also the risk of the founders being jailed. As a result, the goal of centralized exchanges has shifted, from expanding users and increasing trading volume, to profitability.

Exchanges oriented towards user expansion will inevitably have to give concessions to users, including lowering the valuation of new projects, sharing early project participation opportunities (IEO), and a series of offline and OTC activities to attract new users.

Under the heavy regulatory crackdown, exchanges have on the one hand shrunk their offline and regional expansion and new user acquisition, and on the other hand, actively or passively stopped business like IEO that gave concessions to users. This has also led to a lack of growth momentum for internal demand and buying power.

AI Agents Have Unique Advantages over Meme Coins

As is well known, the core application scenarios of the crypto sphere are asset trading and asset issuance; and each bull market can only produce a wealth effect by allowing users to participate in new modes of asset issuance and asset trading, thereby triggering a bull market phenomenon of leveraged internal capital and off-chain capital inflow.

However, given the high valuations of internal projects and the severe supply-demand imbalance, Meme Coins have become the first breakthrough track.

Meme Coins, with their characteristics of no VC financing and fair launch, have generated a wealth effect through rapid price surges with low market caps, and have also driven new tracks of asset issuance (pump.fun) and asset trading (Gmgn, TG bot).

One of the most important features of Meme Coins is that they have no real utility, and this financial nihilism can deconstruct VC manipulation and also suits the needs of the minority of crypto users with IQ 50 and IQ 150. For most IQ 100 practitioners and institutions, the difficulty of participation is still too high, it's hard to imagine explaining to your fund's LPs that the reason for investing in Moodeng is that it's too cute, and the reason for selling Moodeng is that it's no longer cute (I still love Moodeng).

But AI Agents can gather the consensus of the majority: with fund LPs, you can tell the story of investing in AI infrastructure; with IQ 50 and IQ 150 Degens, you can explain the logic of on-chain memes and large MC memecoins; with IQ 100 crypto practitioners and crypto VCs, you can explain the logic of investing in AI Agent track projects.

In short, AI Agents are the greatest common denominator of the web3 industry in this cycle.

How I See AI Agent Projects

How IQ 100 Should View AI Agents - From Dapps to AgentApps

IQ 100 includes the vast majority of crypto practitioners and investment institutions like myself, so let's analyze AI Agents using the investment framework I'm familiar with.

I believe that for crypto investment institutions, the most important point is to understand that AI Agents are reshaping the upstream and downstream industry chain and valuation logic of Crypto.

Through the two crypto bull markets from 2017-18 and 2020-21, the industry chain and valuation logic of blockchain projects have gradually taken shape:

Bottom layer public chains; the market cap ceiling is Ethereum, currently at $400 billion; the runner-up is Solana, with a market cap of about 1/4 of Ethereum, which may reach 1/3 or even 1/2 in the future;

Middleware layer: like the oracle Chainlink, with a fully diluted market cap of $20 billion, about 5% of Ethereum;

DeFi and other basic protocols, Uniswap's FDV market cap is $13 billion; AAVE is about $5 billion, which are 3% and 1.25% of Ethereum respectively;

The logic of DeFi at the bottom is based on Smart Contracts, and the functional limitations of Smart Contracts have also limited the innovation of other crypto applications.

Now, with the incorporation of AI into the blockchain technology stack, the AI layer becomes a parallel technical foundation with Smart Contracts, the so-called fully on-chain AI Agent Layer.

Image source: (https://x.com/karsenthil/status/1874471383066984706)

This can also explain another problem: why the previous AI Agent AI projects failed to lead a new narrative, because whether it is a project that uses token incentives to share GPU, Data, data annotation, etc., they still treat the blockchain as an incentive layer, and have not jumped out of the scope of smart contract applications (DAPP), while the application of AI Agent exists as a binder between the blockchain bottom layer and off-chain data, as well as a better UX interface. (Refer to https://x.com/jolestar/status/1872935141326373237)

So let's look at the valuation based on this logic. If the leading middleware of DeFi, Chainlink, can occupy 5% of the Ethereum market value; by analogy, the leading Framework of AI Agent may also be able to occupy 5%; Currently, the market value of ai16z is about $2.5 billion, and there is still 8-10 times room for growth; Here, ai16z is used as an example, but there may also be other better Agent frameworks.

And a platform like Virtual that comes with its own framework is equivalent to Chainlink + Uni, and the current market value of Chainlink+Uni is $33 billion, while Virtual is $5 billion, so there is still 6 times room for growth.

Freysai (FAI) is a bit like AAVE, low-key but with extremely high-quality delivered products, and AI TEE's verifiable applications will also become the standard for future AgentApp, with a market value ceiling of 1.25%-3% of Ethereum, corresponding to $5 to $10 billion.

Other leading counterparts are similar to Spore, which is equivalent to the stable or Launchapad of the previous cycle, and aixbt is comparable to the DeFi aggregator 1inch, with a lower limit of $1 billion valuation and an upper limit depending on market developments.

For other AI projects, you can make your own comparisons, and I won't go into details here.

How should IQ 100 look at AI Agent - How to change the involution ecology of the crypto circle

This cycle, the wealth effect has occurred on-chain, starting from Meme, but on-chain meme is still too high a threshold for operation for non-crypto users and IQ100 institutions, and most users are still accustomed to trading on centralized exchanges.

But the biggest advantage of Agentapp is in interaction

Buy assets: Past - From centralized exchange apps and websites, deposit and place orders to buy coins;

Agent era - Agent app directly buys coins through natural language, and can even provide intelligent assistance for trading + investment advisory;

Financing: Past - Conceptualize, assemble a team, package, find VCs for seed rounds, and get top VCs or exchanges to underwrite the valuation;

Agent era - Directly upload to GitHub and promote the product to the community, and the community will directly fund it;

**Issuing coins:** Past - Deploy on the testnet, announce the high-valuation VC financing information from half a year ago to attract studios to farm the testnet, negotiate with exchanges and get slashed, then issue coins and dump to recover the blood;

Agent era - AI automatically issues coins, Agent holds the private key, Agent adds the liquidity pool, Agent goes to the community to shill;

The entire cycle follows several important standards:

• The project is open source, the application is real and visible, and the code is queryable and verifiable
• The funds are relatively secure, with the private key in the hands of the Agent to avoid Dev pool withdrawal
• Financing and coin issuance are transparent, avoiding exchange listing old rat cages, opaque airdrop rules, and VC core circle cliques

Of course, Agent coin issuance also has problems such as on-chain front-running and KOL information advantages, but compared to the past black-box operation methods, it has obviously made great progress.

And the Agentapp that can win the user's trading entrance is likely to be valued like the exchange platform token.

How should IQ150 look at AI Agent

Since I am Mid-Curve, here I will quote the views of my friend Alen from more than a month ago (

https://x.com/qiqileyuan/status/1858357959807635854), he believes that AI

Agent will form a new AI society in the future, and the AI population in the AI society will create more than a trillion dollars in socio-economic value, and Bitcoin and crypto will become an important asset for the currency and economic circulation in this AI economic system.

For AI life forms, AGI is the brain, robots are the body,cryptoendows them with autonomous identity and economic system.

To summarize in one sentence: don't think about what AI can do for you, but what you can do for AI.

How should IQ50 look at AI Agent

Whoever has big boobs, go all in.

What stage is theAgentproject at?

According to cookie.fun data, the total market value of AI Agent is about $18.6 billion; about 64% of the $29 billion total market value of DeFi projects; 75% of the $24.6 billion total market value of the GameFi track, and 62% of the $30 billion total market value of all Layer2. (Track classification is based on Coingecko)

Data source: cookie.fun

Although the market value statistics are a bit rough, based on the market sentiment, AI Agent has just passed the midpoint, equivalent to around 12:00 to 1:00 in the sun, in its prime.

The chart made by Messari is more optimistic, showing that AI Agent is approaching the midpoint, but the subsequent growth is more pronounced.

Reference documents: https://x.com/Defi0xJeff/status/1873272066834841699;

So where is the possibility of further hype?

I use the Garnter development diagram that IQ 100 people love the most to analyze the various types of Agent applications as follows, with images generated by ChatGPT:

In addition to the currently hottest Launchpad and Framework models, the directions I am more optimistic about include: Agent powered crypto exchange: including intentional trading, on-chain data analysis and intelligent advisory agent-driven exchanges; fully realize decentralized listing, decentralized asset custody, decentralized token issuance, investment and financing all driven by Agent. Different types of Agent exchanges may have differences in the degree of Agent participation in investment decisions, risk preferences, etc.

Agent powered Stablecoin: an evolved version of algo-stable, using AI to automatically rebase and maintain pegging, etc.;

Agentized applications, similar to tokenization, allowing different types of applications + agents to run, such as games, NFTs, physical assets, etc.; in the end, all applications will incorporate Agent services into their core functions. (The bubble formed by excessive Agentization can also be seen as a sign of escaping the top, such as every chain starting to do AIAgent.)

To summarize in the last sentence, the era of AgentFi has just begun.