Analyzing the Best AI Sectors in Crypto

What Are The Best AI Sectors? (Pt. 1)
If you haven’t already noticed, the buzz around AI is impossible to ignore.
In a market littered with both novelty & junk, identifying signals amid the noise is key.
Below are my opinions on each niche and the market dynamics at play.

1. AI Agents:

IMO, the most viable intersection of AI x Crypto are AI (autonomous) agents.
Vitalik agrees IMO, claiming (AI as the player is the most viable). Player = Agent.
The reason is simple and two fold:

1. Crypto helps agents (payment rails)
2. Agents help crypto (smart contracts 2.0)

Palantir Co-Founder believes AI agents are the key to mainstreaming crypto/Web3.
It’s a frictionless win-win. AI agents can operate on/off chain and in any industry.
They are the final user-touch point; providing action with ML/data inputs.
That said, in terms of competition, the barriers to entry are relatively low.
Inherently, agents are not super novel, and I expect dilution to intensify.
Differentiation based on flywheel, composability & scalability will be key IMO.
In terms of opportunities, you have:

1. Semi-Autonomous Agents
2. Autonomous Agents
3. Agent Marketplaces
4. Agent Simulation Systems
5. Neural Networks (NNs):

NNs are arguably the most ambitious niche, with perhaps the highest upside potential.
Simply, NNs are a style of ML designed to mimic the function of the human brain.
IMO, this is what @VitalikButerin classifies as 'AI as the objective."
@ARKInvest sees NNs as the most ‘important catalyst’ in modern technology.
Broadly, this could encompass training or inferencing compute/ML workloads etc.
Namely, using crypto to incentivize superior AI outcomes within a network.
Both speculatively and fundamentally, NNs are extremely infectious/impactful.
However, many existing mechanisms remain controversial in terms of viability.
In-house zkML, opML…?

1. Compute/GPU Networks:

DePINs built for general-purpose or ML-specific compute have proliferated.
According to @MessariCrypto, they are the oldest and most successful DePIN model.
Namely, using crypto-incentivized networks where compute is made more accessible.
$SOL is currently the leading network for compute DePINs by a wide margin.
There are several narratives at play:

• Disintermediating cloud oligopolists
• Easing the compute supply crisis
• GPU is the new oil of the AI economy
• Falls in the same bucket as NVDA

IMO, compute DePINs have the biggest TAM and strongest PMF.
While scalability is questionable, the cost advantages DePINs offer is undeniable.
They are mainstream-aligned with AI/NVDA and appeal to both sides of the bell curve.
First movers have clear network advantages and extremely deep moats.
Also flywheels, crucial mechanisms, often need significant time to be optimized.
Despite the rapid pace of dilution, the USPs maintain robust defensibility, in my opinion.
Clustering is a space to watch IMO.

1. Final Thoughts:

This is just part 1 of my series. Part 2 will be released in the coming weeks.
For those interested in projects, explore my altcoin report series (link in bio).
I’ve featured emerging innovators such as $MOR, @gensynai & @ionet_official.
As always, not financial advice. My content is 100% independent and educational.