Crypto development often runs on a different clock from token markets, quieter, slower, but no less revealing. That was the clear takeaway from a short but telling tweet by CryptoDep this week, which shared a Santiment-sourced snapshot titled “Top AI & Big Data Projects by Developer Activity (30d).” The dataset looks at public GitHub events, commits, pull requests, issues and other visible activity across projects that sit at the intersection of blockchain, AI and big data.
At first glance, the numbers are stark. ICP towers above the rest with a development activity score of 237 for the 30-day window. The gap between that figure and the next entries is immediate. Filecoin recorded 36.3, Livepeer 31.2, and The Graph 24.4. Further down the list are projects that blend AI and decentralized compute. Bittensor is at 18.1, QUBIC at 17.8, and Oasis Network at 17.5. FLUX sits at 16.4, while Swarms and Virtuals Protocol round out the ten with 8.63 and 8.13, respectively. Santiment is credited as the data source.
Numbers like these don’t tell the whole story by themselves, but they do give you a useful compass. Public GitHub activity is a blunt instrument; it misses private work, research notes, and contributions that happen off-platform, yet when you see a spike or a big gap, it usually signals something concrete: major upgrades, developer sprints to fix bugs, or an influx of new contributors.
AI Blockchain Development Race
In ICP’s case, a 237 score screams “something is happening” in public repositories. That could mean a series of protocol improvements, new SDKs or tools, or simply a coordinated push from the community to move features forward. Filecoin and Livepeer finishing near the top makes intuitive sense.
Both projects are infrastructure-heavy. Filecoin handles decentralized storage, and Livepeer focuses on video streaming infrastructure. Those spaces demand constant iteration to improve reliability and developer ergonomics. The Graph’s strong showing also tracks with expectations, indexing networks require regular maintenance to support an expanding web of decentralized apps and data queries.
The mid-pack entries are interesting because they represent projects that try to mash together AI-style models and decentralized compute marketplaces. Networks like Bittensor, QUBIC and FLUX are prototypes of an emerging category. Think marketplaces for compute, data or model training that aren’t controlled by a single cloud provider.
They show steady developer momentum, even if they don’t yet generate the kind of public activity seen in larger infrastructure projects. That may simply reflect smaller teams, more private experimentation, or development that’s not reflected in a single public repository. Lower scores for projects such as Swarms and Virtuals shouldn’t be read as failures.
In many cases, smaller numbers mean quieter, focused development cycles: auditing smart contracts, private testnets, or groundwork that won’t appear as frequent GitHub events. The ratio of visible commits to meaningful progress isn’t always linear. For people watching the space, engineers considering which stacks to learn, or investors trying to separate hype from substance, developer activity is a useful input among many.
It’s not a prediction of price, but it does help you see where engineering energy is concentrated. This snapshot from CryptoDep, using Santiment’s metrics, is a reminder that the most interesting action in crypto often happens in code, not on exchanges. Keep an eye on these repos; when development accelerates, feature releases and ecosystem growth often follow.
