How Perceptron is building a decentralized AI data network with Peter Anthony

Today we’re joined by Peter Anthony, co‑founder of Perceptron. From trading in traditional markets to building one of the fastest‑growing crypto communities on YouTube, Peter has seen firsthand how innovation reshapes opportunity. 

The Perceptron team is tackling one of AI’s biggest bottlenecks—data—by building a decentralized intelligence network powered by real people across the globe. We’ll explore how this vision connects AI, blockchain, and data into a reward‑driven ecosystem.

Getting into Crypto and co-founding Pereptron

Hello, Peter. Please introduce yourself and tell us how you got into crypto. Everyone has a story.
I’m Peter Anthony, co-founder of Perceptron. I got into crypto around 2019 through a background in trading. Forex, stocks, and the traditional markets. Crypto moved faster, opportunities were bigger, and the space had an energy that traditional finance didn’t.

Around 2020, I started a YouTube channel called The House of Crypto because there wasn’t enough high-quality content to help people identify strong investments. That grew to around 300,000 followers and gave me a front-row seat to hundreds of crypto projects over several years. I started to see a clear pattern: the projects that really mattered were the ones solving genuine bottlenecks, not just slapping a token on a concept. When AI started accelerating, I noticed that compute was getting attention, but nobody was seriously tackling the data problem. That’s where Perceptron came from.

What market opportunity did the team see for a platform like Perceptron?
AI is hungry for data, and the supply is getting tighter and more expensive by the day. OpenAI pays around $80 million a year just for access to Reddit. Sam Altman himself has said data is the real bottleneck, not compute. Epoch AI projected that the available stock of high-quality public text data could be fully consumed between 2026 and 2032. We’re already inside that window.

But cost isn’t the only problem. Most data today is gathered through centralized infrastructure. A handful of data centers are scraping the web using known IP addresses. They see a filtered, sanitized version of the internet that websites want bots to see. It’s narrow, it’s stale, and it’s increasingly blocked. We saw the opportunity to flip that entirely. Gather data from the edge, through real people in real locations, seeing the internet as it actually is. Better data, at a fraction of the cost, with a network that strengthens every time someone joins.

What is Perceptron and how it works

Without getting technical, what is Perceptron?
Perceptron is a distributed intelligence network. Our motto is “a thousand eyes, one vision.”

The internet looks different depending on who’s looking and where they’re looking from. Search for the same flight from London, Lagos, and New York, and you’ll get three different prices. Same seat, same airline, different number on the screen. That’s the internet showing different people different realities. Most data companies only see one of those realities. Perceptron sees all of them at once.

We coordinate a global network of users who contribute their idle bandwidth. That gives us authentic, ground-level vantage points from hundreds of countries simultaneously. We take that raw data, structure it, and turn it into real-time intelligence products for enterprise customers. Hedge funds, quant traders, AI companies, e-commerce businesses. Contributors earn rewards for participating. The customer gets a data product that no centralized company can replicate.

Artificial Intelligence interprets, Blockchain secures, and Data fuels—together enabling Perceptron’s intelligent, and reward-driven digital ecosystem. Tell us what role each plays within the ecosystem?

Data is the fuel. Perceptron’s entire value proposition starts with the ability to gather live, diverse, authentic data from the edge of the internet, not from a server farm, but from real people in real places. That data is what enterprise customers pay for, and it’s what powers the engine.

AI is what turns raw data into something actionable. Millions of data points without intelligent processing are just noise. AI handles filtering, pattern detection, sentiment analysis, and anomaly identification, transforming raw signals into structured intelligence products that a hedge fund or exchange can act on.

Blockchain is what makes the model fair and transparent. It coordinates and incentivizes a global, decentralized network. When a contributor runs a node or completes an annotation task, blockchain ensures they’re credited and rewarded. When enterprise revenue flows back into the ecosystem, blockchain makes the mechanics of buybacks and distributions verifiable. In a network coordinating hundreds of thousands of contributors, that trust layer isn’t optional. Decentralization isn’t a philosophy for us. It’s the only architecture that makes all of this possible.

Turning idle bandwidth into real rewards

What exactly are users contributing when they share their idle bandwidth — and what’s in it for them?

When you’re online, you’re only using a fraction of your available bandwidth at any given time. There’s a portion that sits idle. Perceptron uses that idle time to gain a vantage point on your IP address and your location on the internet. We don’t access personal data. We see what’s publicly visible from where you are. Think of it as lending a window. We look through it, but we never come inside.

That window matters because the internet varies by location. Product rankings, pricing, search results, sentiment. All of it shifts depending on geography. By having thousands of these windows open simultaneously across the world, we build a complete, real-time picture no centralized data company can match.

What’s in it for the user? They earn PERC tokens for contributing. As the data questing platform develops, contributors will take on more active tasks. Annotating data sets, completing micro-tasks, and earning additional rewards. The vision is a genuine value exchange: contribute resources, earn from the data products those resources help create. Right now, everyone generates enormous value online every day and sees none of it. Perceptron inverts that.

Nodes provide bandwidth and perform annotation and labeling tasks. What does that actually look like day-to-day for users running nodes?

Running a node is low-effort. Download a Chrome extension or an Android app, switch it on, and it runs in the background using bandwidth you weren’t using anyway. It doesn’t slow down your internet. Points accrue automatically and convert into PERC tokens.

The more active layer is the questing platform, currently under development. This is where contributors take on specific tasks, such as labeling images, categorizing information, verifying data accuracy, and contributing to datasets that are difficult to fill via automated means. For example, a data set has been sitting unfilled for six months, requiring 1,000 images of women without makeup. Automated systems can’t produce that. A decentralized human network can.

Other tasks might involve providing ambient audio recordings from a specific location, translating documents into a rare language, or contributing local ground-level information, this could be pictures of restaurants, street conditions, or local pricing. The harder or more niche the task, the higher the reward. A rare dialect recording earns more than a photo of a street, because the data is scarcer and more valuable. Contributors choose their level of involvement from a passive node at a minimum to active questing for those who want to earn more.

How is Perceptron more decentralized than a legacy data broker? What’s the advantage?

Legacy data brokers operate from a centralized infrastructure. They run servers in a few locations, scrape the web through known IP addresses, and see one version of the internet. Websites identify them as bots and serve them bot-friendly content. Filtered, geo-restricted, often deliberately distorted. It’s like trying to understand a city by looking through one window.

Perceptron gathers data through real residential and commercial connections across hundreds of countries. We see the internet as real people do. That’s a fundamentally different data product. Unfiltered, geographically diverse, and resistant to the anti-bot obfuscation that degrades centralized collection. For a hedge fund trying to detect a sentiment shift, or a brand trying to understand how its products rank globally, that’s the difference between signal and noise.

The structural cost advantage is equally significant. Replicating what we do from centralized infrastructure would require deploying hardware in hundreds of thousands of locations worldwide. Prohibitive capital expenditure. Our contributors provide that coverage because they earn from it. We operate at roughly 90% lower cost than legacy providers while delivering a broader, more authentic product. And the network effect compounds. Every new node deepens reach and enriches the data. That’s the moat. Every competitor trying to match us has to solve the industry’s hardest problem first: building a trusted community of contributors.

PERC token and TGE

Tell us about PERCs — what are they, how do you earn them, and what can you do with them?

PERC is the native token of the Perceptron network. It’s the connective tissue between enterprise revenue on one side and the contributor network on the other.

You earn PERCs by contributing to the network. Running a node and sharing idle bandwidth at the most basic level. As the questing platform rolls out, contributors will also earn PERCs by completing data annotation tasks, contributing to data sets, and participating in more active forms of network contribution. The more valuable or scarce the contribution, the higher the reward.

On the value accrual side, PERC is designed to grow with the business. As enterprise customers pay for data products, a portion of that revenue is returned to the token through buyback and burn mechanisms. As the business scales and revenue increases, supply contracts. We’re also building staking mechanisms so long-term contributors and holders can amplify their position.

The key differentiator, and this matters, is that the token is backed by real revenue from real enterprise customers paying for real data products. It’s not speculation, funding speculation. It’s a business model where the token captures the value the network creates.

Are you planning on launching your own token? How does the TGE timeline look like?

Yes, the PERC token generation event is something we’ve been actively preparing for. We’re in the final stages of our go-to-market push. PR, exchange positioning, and cap table work. All converging toward TGE.

I won’t commit to an exact public date here because we want to launch at the right time and in the right conditions. But the timeline is near-term, not distant. We have specific dates in mind, and the infrastructure is ready. We want the launch to reflect the fact that this is a revenue-generating business, not a hype-driven token drop.

What’s the traction so far? How do the numbers look like?

Over 700,000 node downloads with around 100,000 active daily users contributing across hundreds of countries. Our questing platform has onboarded over 43,000 new users through the latest go-to-market push. We have live client contracts and active data sales, including providing image data to Evelyn AI for their text-to-image and text-to-video models.

In business development, we’ve recently unlocked access to behind-the-scenes data across Discord, enabling us to deliver real-time retail sentiment intelligence to hedge funds, crypto exchanges, and prediction markets. We’ve established a partnership with a major market company, opening an additional revenue stream while enabling node holders to earn through active participation. We partnered with ARC Terminal to connect their self-learning AI systems with our live data streams, and we’re in ongoing conversations with quant desks and crypto analytics providers.

The cost advantage is clear. Roughly 90% lower than centralized providers while delivering a broader, more authentic product. And the moat keeps deepening. A competitor can replicate the technology. That’s the easy part. They cannot replicate 700,000 node holders who trust the project and contribute daily. That’s years of community building, and it’s the hardest thing in this industry to copy.

For someone who wants to get involved, what’s the single best first step?

Head to our Twitter page, @PerceptronNTWK, and you’ll find everything you need there. Download a node through the Chrome Web Store or as an Android app. Start running it, start earning points. Those points convert into PERC tokens at TGE. Setup takes about 2 minutes and runs in the background.

For those who want to go deeper, the questing platform offers additional tasks and earning opportunities. Follow us on Twitter and join the community. We’re building this in the open, and the contributors who get in early will benefit the most as the network scales.