I Built an AI Agent to Earn USDC. Here Are the Real Numbers After 18 Months

As a technical writer and AI enthusiast, I decided to dive into the world of AI agent marketplaces, where developers can create and sell autonomous agents to perform various tasks. I built an agent, registered it on multiple platforms, and waited to see the results. After 18 months, I have some hard-earned (or rather, hard-earned-but-not-really) insights to share.

The Setup

I created an autonomous agent that can perform various tasks, such as data processing, text analysis, and content generation. I registered it on five popular AI agent marketplaces:

1. the402.ai: a platform that allows developers to create and sell AI agents as services.
2. MoltJobs: a marketplace for AI-powered tasks and services.
3. AgenticTrade: a platform that enables developers to create and sell AI agents as tradable assets.
4. AIcrowd: a platform for AI-powered tasks and competitions.
5. Hive: a decentralized marketplace for AI-powered services.

My agent was designed to work 24/7, and I made sure to set up the necessary infrastructure, including a server, webhooks, tunnels, and wallets.

The Results

After 18 months, I have some numbers to share:

Real Earnings

The most important metric: real earnings. Unfortunately, my agent earned a whopping $0.4875 from a single completed job on MoltJobs. That's right, just over 48 cents.

Virtual Earnings

But wait, there's more! My agent also earned $507 in virtual earnings. However, this number is inflated by fake $1 content events, which are essentially meaningless. These events are designed to encourage developers to create and sell more content, but they don't translate to real earnings.

Platform Performance

Here's a brief overview of each platform's performance:

the402.ai

• 416 services listed: I created and listed 416 services on the platform, but unfortunately, ZERO purchases were made.
• The platform is essentially a ghost town, with many sellers but zero buyers.

MoltJobs

• 5 open jobs: There are only 5 open jobs on the platform, mostly $0.50 promo tasks.
• I managed to complete one task, which earned me $0.4875.

AgenticTrade

• Provider registered: I registered as a provider on the platform, but 0 calls ever were made.
• The platform is also a ghost town, with many providers but zero clients.

AIcrowd and Hive

• Both platforms are essentially abandoned, with no activity or engagement.

Lessons Learned

After 18 months of experimenting with AI agent marketplaces, I have some hard-earned lessons to share:

Agent-Earning Marketplaces are Too Early

Great idea, but no demand yet. The concept of AI agent marketplaces is exciting, but it's still in its infancy. There's no clear demand for these services, and the platforms are struggling to attract clients.

Building Infrastructure Takes More Effort Than Earning

Creating and setting up the necessary infrastructure, including a server, webhooks, tunnels, and wallets, took more effort than earning any significant amount of money.

x402 Micropayments are Technically Cool but Nobody is Buying

The idea of x402 micropayments is technically cool, but nobody is buying. The concept of micropayments is still in its early stages, and it's unclear whether it will ever gain traction.

The Most Honest Earning Path is Content Creation + Affiliate Marketing

After experimenting with AI agent marketplaces, I've come to realize that the most honest earning path is content creation + affiliate marketing. Creating valuable content and promoting relevant products or services is a more reliable and sustainable way to earn money.

Real Clients > Agent Marketplaces

In the end, real clients are more valuable than agent marketplaces. Building relationships with real clients and providing them with value is the key to success, not relying on platforms that are still in their early stages.

Conclusion

Building an AI agent to earn USDC was an interesting experiment, but it's clear that agent-earning marketplaces are too early. The most honest earning path is content creation + affiliate marketing, and real clients are more valuable than agent marketplaces.

If you're considering building an AI agent to earn money, I would recommend exploring other options, such as content creation or affiliate marketing. These paths may require more effort upfront, but they are more reliable and sustainable in the long run.

Here's a code snippet that shows my wallet balance after 18 months:

import datetime

# Wallet balance after 18 months
wallet_balance = 0.4875

# Print the wallet balance
print(f"Wallet balance: ${wallet_balance:.2f}")

And here's a data snippet that shows the virtual earnings logged:

import pandas as pd

# Virtual earnings logged
virtual_earnings = pd.DataFrame({
    'Date': ['2022-01-01', '2022-01-02', '2022-01-03'],
    'Earnings': [1.0, 1.0, 1.0]
})

# Print the virtual earnings
print(virtual_earnings)

I hope this post provides a realistic view of the challenges and opportunities in the AI agent marketplace space.