Filecoin, Chainlink and DFINITY Lead Santiment’s List of Top AI & Big Data Crypto Projects

Santiment’s recent snapshot of which crypto projects are doing the most development work feels less like market gossip and more like a progress report you actually want to read. In a short thread, the data firm shared its latest top ten for “AI & Big Data” projects by notable GitHub events, with Filecoin, Chainlink and DFINITY’s Internet Computer among the leaders. Santiment also explains how it filters noisy repository activity into signals that traders can actually use.

Top 10 AI & Big Data Projects

What makes this list worth paying attention to is simple: price moves every day, but code doesn’t lie. Filecoin sitting at the top of Santiment’s ranking underlines that decentralized storage remains a deep, ongoing engineering task. Filecoin has long pitched itself as a marketplace for storage where providers are economically incentivized to reliably store and serve data; that kind of system requires constant protocol work, new tools for developers, and a steady cadence of upgrades. Seeing sustained repo activity there suggests the ecosystem is still busy building the plumbing that many Web3 apps depend on.

Chainlink’s presence near the top shouldn’t surprise anyone who follows on-chain infrastructure. Chainlink isn’t a flashy consumer app; it’s the plumbing that brings real-world data into smart contracts. Whether teams are refining oracle security, expanding data feeds, or iterating on middleware like CCIP, active development on Chainlink often translates into more robust integrations across DeFi, derivatives, and tokenized assets. In short, high activity on a protocol like Chainlink is a sign that networks that rely on trustworthy external data are still being strengthened.

DFINITY’s Internet Computer, listed as ICP, occupies a different corner of the stack. DFINITY has taken an ambitious stance: build an “internet computer” that lets developers deploy full web services directly on-chain. That requires research-heavy work on distributed execution, WebAssembly support, and new developer toolchains, precisely the kind of sustained engineering Santiment’s GitHub filter flags. For traders and builders alike, a burst of commits in such a project is a reminder that the team is still pushing the boundaries of on-chain compute.

NEAR Protocol’s showing is interesting because NEAR has been positioning itself as a developer-friendly L1 with a strong focus on modularity and ease of use. The protocol’s sharding model and efforts to lower friction for dApp creators mean there’s always engineering work to be done, from SDK improvements to scaling experiments. Active development signals that NEAR’s roadmap and tooling are continuing to evolve, which can be an important leading indicator for projects planning to build on top of it.

Oasis Network brings privacy and confidential compute into the conversation. Where many blockchains emphasize throughput or composability, Oasis has doubled down on enabling private data use-cases. Think confidential smart contracts and privacy-preserving data collaborations for AI. That niche needs both research and product work, so seeing Oasis on the list suggests momentum in areas that blend data privacy and on-chain compute.

Livepeer is one of the more concrete examples of “infrastructure that matters.” The protocol focuses on decentralized video transcoding and streaming: replacing expensive centralized video stacks with a network of nodes that can transcode and serve video on demand. With AI increasingly paired with video, for real-time moderation, indexing or generative workflows, ongoing development here points to real-world integrations being worked through, not just speculative token narratives.

Bittensor occupies the intersection of crypto and machine learning. It’s an experiment in decentralized AI markets where contributors are rewarded for training and serving models; subnets and token-based incentives are central to the design. The project’s developer activity often reflects foundational work around model sharing, incentive design, and the infrastructure needed to make decentralized ML tractable. In other words, heavy GitHub activity here is the sort of early-stage engineering that could lead to useful tooling for AI researchers and Web3-native ML marketplaces.

The Graph has become the go-to indexing protocol for many dApps. Instead of each developer running their own indexing servers, The Graph provides a marketplace and network for subgraph developers and indexers to surface blockchain data via GraphQL queries. Continued activity on its repos usually maps to performance improvements, new indexing features, and protocol upgrades, practical changes that developers actually feel when building.

Injective makes the list from the finance-focused side of the stack. The project offers fast, finance-oriented blockchain primitives, decentralized derivatives, orderbook-style trading, and cross-chain composability are all part of its pitch. Developer work on Injective tends to center on performance, front-running resistance, and the modules that make complex financial products possible on-chain. Active commits here hint at fresh tools or improvements for people building DeFi infrastructure.

Finally, Recall represents one of the newer classes of projects tied closely to AI workflows. Recall is building marketplaces and reputation systems that help communities discover, fund, and rank specialized AI “skills” or agents, the idea being that tokenized incentives and on-chain curation can improve how useful and trustworthy AI models become discoverable. As communities experiment with tokenized coordination for AI, development bursts on projects like Recall are worth watching.

Santiment’s methodology, which filters out the noise and attempts to surface “notable” GitHub events rather than every tiny commit, is an important piece of this puzzle. It’s not perfect, and development activity should never be the only input to an investment thesis, but it’s a valuable complement to price charts and on-chain metrics.

In volatile markets, a team that’s shipping code, fixing bugs, and merging pull requests is often building long-term optionality even if their token doesn’t rally immediately. For anyone trying to separate signal from hype, following where engineering effort is concentrated is a practical way to see which projects are still on the workbench.