Developer activity reveals infrastructure focus
I was looking at this data from CryptoDep and Santiment, and it’s pretty interesting how it shows what’s actually happening behind the scenes. You know, when you see token prices jumping around, it’s easy to forget that real development work follows a different rhythm. This snapshot of GitHub activity over 30 days gives us a clearer picture of where engineering effort is actually going.
ICP really stands out with a score of 237. That’s a huge gap compared to everyone else. Filecoin comes in at 36.3, Livepeer at 31.2, and The Graph at 24.4. The rest of the list includes Bittensor, QUBIC, Oasis Network, FLUX, Swarms, and Virtuals Protocol, all with scores under 20.
What the numbers actually mean
Now, I should be careful here because GitHub activity doesn’t tell the whole story. It misses private repositories, research work, and contributions that happen outside of public platforms. But when you see numbers this far apart, it usually means something concrete is happening.
For ICP, that 237 score suggests major development work. Maybe they’re pushing through protocol upgrades, building new tools, or just have a very active community contributing code. It’s hard to say exactly without digging into the repositories themselves, but that level of activity doesn’t happen by accident.
Infrastructure projects dominate
What strikes me about this list is how infrastructure-heavy the top projects are. Filecoin handles decentralized storage, which requires constant work to improve reliability and performance. Livepeer focuses on video streaming infrastructure, another area that needs regular iteration. The Graph’s strong showing makes sense too – indexing networks need maintenance to support growing numbers of decentralized applications.
These aren’t flashy consumer apps. They’re the plumbing that makes other things work. And maybe that’s why they show such consistent developer activity. When you’re building infrastructure, you can’t just launch and forget about it. There’s always something to fix, improve, or scale.
The emerging AI and compute category
The middle of the pack is interesting in a different way. Projects like Bittensor, QUBIC, and FLUX are trying to create decentralized marketplaces for compute power, data, or AI model training. They’re essentially building alternatives to centralized cloud providers.
Their scores are lower, but I don’t think that necessarily means they’re less important. Smaller teams, more private experimentation, or different development workflows could explain the difference. Sometimes meaningful progress happens in private testnets or through smart contract audits that don’t generate many public GitHub events.
Why this matters
For developers trying to decide which technologies to learn, or for anyone trying to understand where real engineering work is happening, this kind of data is useful. It’s not a perfect metric, and it certainly doesn’t predict token prices. But it does show where teams are putting in the work.
What I take from this is that the most interesting developments in Web3 often happen quietly, in code repositories rather than on exchanges. When you see sustained developer activity, it usually means features are being built, bugs are being fixed, and foundations are being strengthened.
Maybe we should pay more attention to these metrics. They give us a glimpse into the actual building happening behind the scenes, which ultimately matters more than short-term price movements.
