🥛 PRO | AI agents: The next frontier or overhyped bots? 🤖

GM, this is Milk Road PRO. The newsletter that teaches you how to profit from our future robot overlords! 

Remember when everyone was talking about AI agents last year?

The hype was off the charts, the gains were massive, and FOMO was at an all-time high. 

So, it’s not surprising that AI dominated the end of 2024.

The 10 largest AI tokens exploded by a mind-blowing 2,940% (that's almost 30x!), leaving even the memecoin madness in the dust with their "measly" 2,185% returns. 

Sure, nobody times the bottom and top perfectly, but these numbers show how powerful crypto narratives can be when they catch fire. 🔥

That insane AI performance was largely driven by the wild price action in the AI agent sector. And we're not just talking about the agents themselves, but the entire ecosystem, including:

• AI-related memecoins

• Launchpads for AI Agent tokens

• Frameworks to build or customize AI Agents

There's a whole universe of AI crypto projects out there, sorted into several key categories:

AI-Native L1s: Networks like Bittensor and NEAR are building blockchains specifically optimized for AI operations and tokenized machine learning.

Decentralized Computing: Platforms like ICP, Golem, and Akash creating marketplaces where users can rent out and utilize distributed computing power for AI training and inference.

Data Management: Projects like The Graph and Filecoin organizing, indexing, and making blockchain data accessible for AI applications.

AI Services: SingularityNET and Fetch.ai offering decentralized marketplaces for AI algorithms and services that anyone can tap into.

Specialized Infrastructure: Render for GPU rendering, Oasis for privacy-preserving computation, and AIOZ and Theta for content distribution networks.

Everyone was scrambling to claim the "AI" label. And while some deserved it more than others, we're zooming in on one particular category—one that truly earns the title, maybe more than any other: AI agents.

✍ AI agents are like digital workers that follow instructions—but with a twist. They have their own "brain", meaning they can adapt, learn, and make independent decisions when needed.

Unlike regular AI models that just respond to prompts, these agents actually DO stuff - searching the web, booking appointments, trading assets, or creating content independently.

👉 We've only just scratched the surface of what AI agents can do—but they will undoubtedly be a key part of our internet experience in the future.

It’s not just us who see massive potential of AI agents. Think about the software that powers their creation, tokenization, and management. 

That’s exactly what Virtuals Protocol set out to deliver—and the market was beyond excited.

The Virtuals Protocol, with its token $VIRTUAL, was the top crypto gainer of 2024, skyrocketing by an insane 231x. 

That’s life-changing money if you got in early! But let’s be honest…most of us didn’t.

And, in the past few weeks, most AI tokens have taken a serious hit (we are gonna show you a chart in a second). 

So the real question is: Does the AI agent narrative still have legs in 2025, or was that explosive rally just another bubble that’s already popped?

Well, that’s what we want to cover today. We’re breaking down the following:

• The five stages of AI agent evolution (and where we are now)

• The state of the AI Agent market after February's bloodbath

• Hype vs. Fundamentals: separating wishful thinking from real value

• Investment opportunities that might actually have legs

• How we’re managing our AI Agent investments

P.S. - We’re hosting a PRO AMA next week — scroll to the bottom for more info. 👀 

Buckle up—we’re here to help you make sense of the AI madness that’s taken the main narrative position over the past few months.

But to understand where we're going, we need to see where we've been. 

The AI agent ecosystem isn't standing still, it's evolving rapidly through distinct phases.

THE FIVE STAGES OF AI AGENT EVOLUTION

Think about how crypto itself evolved: from Bitcoin to smart contract platforms to DeFi to NFTs. 

Each phase built on top of the previous one, creating increasingly complex and valuable ecosystems. The same pattern is unfolding with AI agents.

👉 What we're witnessing is development at an accelerated pace, where years of technological evolution are compressed into months. 

The market is efficiently testing different models and use cases, revealing which ones deliver actual utility versus those built purely on speculation. 

Understanding these evolutionary stages helps you identify genuine value creation and helps you focus on projects with sustainable economics. 💪

These are the 5 stages of the AI Agent evolution:

1/ Initial Hype Phase (Meme Agent tokens)

Remember $GOAT (Goatseus Maximus)? It rocketed from zero to a $1.2B market cap based on a NSFW meme and an AI chatbot's obsession with it. 

✍️ Pure hype, zero utility, and predictably, it's now down 96% from its peak.

This phase is like the wild teenage years of AI agents. Lots of energy and excitement, but not much substance beneath the surface.

2/ Personality Phase (AI Agents)

Next came the "personality" phase. Luna from Virtuals Protocol gained 1 million TikTok followers, hosts livestreams, and posts chaotic tweets. Cool for grabbing attention, but is that enough for sustainable growth?

Spoiler: It isn't so far. The $LUNA token had a short-lived pump before crashing back to earth. Turns out pure "tokenized attention" doesn't hold value for very long.

3/ Specialization Phase (Utility AI Agents)

Now we're getting somewhere. Projects like $AIXBT provide actual trading alpha (plus a lot of hallucinations).

In January, AIXBT tweeted about $PIPPIN token, which pumped 8x in 53 hours. That's where AI agents start transitioning from "fun" to "useful" — the critical turning point where real value begins.

✍️ Think of it as AI agents getting their first job. They're not just partying on social media anymore; they're delivering something people will pay for (AIXBT terminal access for unlimited alpha is 600k tokens which is about $62k - obviously not sustainable prices but there is a business model here nonetheless).

4/ Infrastructure (Launchpads and Frameworks)

These are the pick-and-shovel plays: Virtuals, AI16Z, and others are building the foundations for creating and scaling AI agents. Think of them as the "L1 blockchains" of the AI agent world.

✍️ Just like how Ethereum created a platform for thousands of dApps, these frameworks provide the backbone for the next generation of AI agents.

5/ Full-Use Agents

The final evolution: AI that's seamlessly integrated into your daily workflow. Portfolio management, 24/7 customer support, personalized research — we're not there yet, but that's the promised land.

✍️ This is when AI agents aren't just neat toys or specialized tools, but essential parts of our digital lives.

So where are we now? Somewhere between stages 3 and 4, with the market pivoting from pure hype to actual infrastructure. 

👉 The party phase is over, and now the serious builders are sticking around.

With that clarified, let’s get into the hot stuff!

Starting with some context on where the AI Agent market is today. 🤝