The Difficulty of Stablecoins

Introduction

Stablecoins are the talk of the town. Everyday there is some breaking news. Just today, Stripe announced they are acquiring Privy, a Wallet as a Service Company. Paypal also announced they are natively minting PYUSD on Stellar. It’s impossible to keep up with all of it. With more companies coming into the space, the need to track and access stablecoin data is growing. But from customer calls, people keep coming back to the same four questions:

1. What are stablecoins being use for?
2. Who is using stablecoins?
3. What opportunities exists?
4. Where in the world are stablecoins being use?

At Artemis I spend every day collecting, organizing and aggregating stablecoin data so I can answer these questions. Today, it’s time to debunk a few myths and get to the bottom of how hard are these questions REALLY to answer.

Myth 1: The data is accessible, open and transparent to everyone.

Getting access to blockchain data without working with a service is prohibitively expensive and technically complex. While today it is much easier to get access to raw blockchain data than it was 5 years ago, this problem still persists. The raw chain data leaders Dune, Flipside, Allium, and Goldsky all have their strengths and weaknesses. However, the most challenging aspect is no single provider supports every chain that matters.

Reality: Everyone and their mother is launching a blockchain with their own quirks, making data analytics extremely complex

To truly understand how your stablecoin is being used and where new opportunities lie, you need visibility across allchains — not just the ones you've launched on. As you expand to more ecosystems and ask more complex questions, the scope of what you need to track grows quickly.

Take PYUSD as an example: after integrating with LayerZero’s OFT protocol, getting the full picture means understanding Ethereum, Solana, LayerZero’s bridging logic, and now newer chains like Berachain and Flow. And this challenge only gets harder as users bridge the token to even more places.

The challenge isn’t just access to the chains you’ve launched on — it’s keeping up with a constantly expanding, uncapped universe of chains. Which brings us to problem number two: architecture fragmentation.

Each blockchain has its own data format and architecture

Think back to the early 2000s, when sending a file to a colleague didn’t guarantee they could open it. A PowerPoint might not open on a Mac. A video file might need a special codec. Everyone used different software, and nothing worked together seamlessly. Teams wasted time converting files, fixing formatting, or chasing down the right version. I was only in lower school but I faced these problems every day.

We’re in that era for blockchains — where every major chain speaks a different language. The most-used chains today — Solana, Tron, Ethereum, TON, Stellar, Aptos — all have fundamentally different architectures and represent data in radically different ways.

To understand a simple transfer on Solana, you need to grasp token accounts and owner accounts. On Ethereum, you need to distinguish between smart contracts, externally owned accounts, and ERC-20 tokens. On Aptos and Sui, it gets even more complex with their object-oriented models. Then there’s Stellar and TON — long-tail but high-usage chains for stablecoins — which don’t fit into any of the common architectural paradigms.

Understanding activity across chains means untangling a growing web of unique technical foundations.

To make this concrete, let’s return to the PYUSD example. Until recently, understanding PYUSD meant understanding Ethereum, Solana, and the LayerZero protocol. But now, with its launch on Stellar, you also need to understand Stellar’s architecture — including its new smart contract platform, Soroban. It’s an entirely different model, with its own virtual machine and a fundamentally different way of handling transfers and balances.

You need to be a domain expert just to access and parse the data — long before you can extract any meaningful insight.

Myth 2: The job is done once you get access to the blockchain data, insights can now be made

We are now getting into the secret sauce of Artemis where we blow our competition away.

Say you have access to all of this blockchain data. You have built the datasets that get you balances and transfers across the entire ecosystem. What do you actually have? Well it turns out you have a whole lot of noise. Users are represented with letters and numbers, and wallet balances are inaccurate or misleading (we will touch more on this in Myth 3).

Reality: Context and off-chain data are MUST-HAVES when understanding what is happening onchain

Even after doing all the hard work to collect onchain data, you’re still flying blind when it comes to answering key questions: Who is using your stablecoin, and where is it being used? The only thing you can confidently say is, “My stablecoin is being used.” But that’s not actionable. That doesn’t help you understand user behavior, market penetration, or opportunity. To get there, you need off-chain context. So the real question becomes: what offchain data do you actually need — and how do you get it?

1. Application and protocol labels: There’s no single source of truth for tagging onchain activity. Flipside, Dune, the Open Label Initiative, block explorers, Arkam — they all offer pieces of the puzzle, each with their own schema and limited coverage. To even begin answering questions like “What application is this address using?” or “What kind of usage are we seeing?”, we’ve had to unify fragmented label sources and tag important wallet addresses by hand ourselves. Without this concerted effort, onchain data is just noise.
2. Geolocation: This is the holy grail — and probably the question I get asked most often: Where are my users? We approach this using a mix of timezone heuristics and more advanced techniques to infer geographic distribution. More importantly, we work with data partners to acquired proprietary offchain geo data to better triangulate what country the wallet most likely resides in.

This problem only gets worse as new blockchains gain traction, each with their own architecture and data model. Labels don’t translate across ecosystems, which means every new chain puts you back at square one — rebuilding context from scratch.

As a result, only Artemis, an independent neutral data analytics company, can get around these issues. Artemis is the “Switzerland” of data. By partnering closely with every major L1s and protocols, we have the most context and most complete labeled data set in the industry, allowing us to cut through the noise better than anyone else, even if it’s not perfect.

Myth 3: Blockchains are immutable

Blockchains are more complicated than they seem. Over the last few years the industry has begun to standardize and follow particular design patterns for token transfers. But previously this has not been the case. When bridging first became popularized, there were no community standards around how to track this information, and it created issues when trying to track both balances and transfers. You need to understand the bespoke history of each chain.

Reality: Blockchain “database schemas” change all the time — you need to be a blockchain historian to have accurate data

It is easy to forget that each of these ecosystem and in flux and constantly changing. For example, Solana has had major upgrades to both its architecture (how the blockchain works) and token program (how tokens are created and transferred).

1. Architecture Upgrades: When Solana first launched the chain did not store timestamps in longer term storage. This caused a major issue when trying to account for balances over time. In 2020 Solana changed this, but it poses the question: how do you rebuild historical balances without timestamps? (Read more here)
2. Token Program Upgrades: Last year Solana launched the Token Program 2022 to solve issues around fragmentation of the original design (read more here) but this requires you to understand the nuances of this new token program in order to track fungible tokens accurately.

Both solutions to these issues are non-trivial and only a subset of the full set of issues we face every day at Artemis.

Piggy backing off of this, people are constantly told that blockchains are immutable public append only databases. And while this has turned out to be the case now, in the beginning this was quite different. Optimism is a great example of this, they not only had a genesis event and launched once but actually relaunched a few months later.

1. What was the outcome? There is no complete dataset in the world of all the token transfers on optimism.‍
2. Why does this matter? This missing data is crucial to understand and analyze both current and past history of major stablecoins on OP Mainnet including USDC, USDT, and DAI. Without this data, you no longer have the complete dataset and wallet balances become impossible to get accurate.

I previously wrote an article that explains some of these edge cases when thinking of how to track this information. You can read more here.

Building these datasets meant becoming blockchain historians. We've spent years in the space doing the work to understand the nuanced evolution of each chain.

Conclusion

Blockchain data presents unique challenges that don’t exist outside of crypto. Even though crypto data is “open,” extracting anything actionable, surprisingly, requires offchain data, integrations with over a dozen data providers, context that’s spread across “crypto Twitter” and thousands of pages of documentation. Otherwise, you're just chasing shadows in a space that moves at the speed of light.