Blockchain needs AI; AI needs blockchain, confidentiality, and trustlessness.

We’ve invested some time to research current developments in the Crypto x AI space and identified over 80 projects building in this space. Lots of off-chain compute job, some on-chain mechanics and tons of hype, organized into a chart below.

There are a plethora of ways to utilize AI that improve different markets in the crypto space and beyond. For example, using AI to analyze on-chain data can expose misconduct like wash trading and illegal funds. Blockchain technology is also enhancing creative work, such as intelligent gaming NPCs and collectibles, while autonomous agents improve cooperation through payments or smart contracts.

Those are just a few examples of what crypto can do for AI and who is building so far. Considering its expansive nature and incredible growth, the list is by no means conclusive and we will update and expand it regularly.

Let’s explore four highlighted projects who are building AI technology for the space and how it can be crucial for the future of crypto and blockchain-based ecosystems.

Development Platforms

Development platforms form the backbone of any technological ecosystem, providing the tools and infrastructure necessary for developing new applications. These platforms offer a range of services from secure data sharing to smart contract development, facilitating the seamless integration of AI capabilities into blockchain applications.

The list of relevant projects ranges from established L1s&L2s like Oasis, NEAR and Phala to very young but specific ecosystems like Morpheus and Talus.

Highlight: Oasis

Oasis is bringing their privacy expertise for crypto x AI by mainly offering:

1. On-chain Confidentiality
2. Off-chain Verifiability

Oasis has built Sapphire, a Confidential EVM (TEE) that among other things enables AI model results to be posted on-chain while being fully confidential and thus unlock unique use cases and monetization methods for model builders that wouldn’t be viable on any other EVM.

Following the expansion of TEEs, Oasis aims to capture more through identifying a compelling use case where AI projects can leverage Sapphire/ROFL capabilities. ROFL, an extension of Sapphire runtimes that support off-chain components, which can be more flexible in the kind of computation they support. This means that ROFL will be able to support not only AI expansions, but any tech that needs off-chain support in order to be decentralized. Read more about ROFL in Decentralize All The Things!

Once TEE support is expanded with Intel TDX and NVIDIA GPU TEE support, crypto can have real local model inference. With this, developers can prove that the correct models were used and also, for example, have private prompts that could be sent to models so that not even the operator would see the prompts or responses.

Decentralized Compute

Another crucial building block for Crypto x AI is decentralized compute. The demand for computational resources is skyrocketing with the rise of AI, and decentralized computing Platforms are providing scalable solutions.

They distribute computational tasks across decentralized networks, optimizing resource use and enhancing performance.

Highlight: Aethir

Aethir is connecting the GPU supply and demand side by aggregating and intelligently redistributing new and idle GPUs from enterprises, data centers and consumers. This has resulted in an already impressive partner and node network.

For example, Aethir has access to 4000+ of the latest NVIDIA H100 GPUs and allows projects to leverage it for use cases like gaming, AI and virtualized compute.

Predictions/Trading

The financial sector is one of the most prominent areas where AI and blockchain converge. Predictive analytics and automated trading systems are transforming how we approach investments and market strategies.

These platforms leverage AI algorithms to analyze market data, forecast trends, and execute trades, improving trading efficiency and profitability.

Highlight: Ocean’s Predictoor

Built on Sapphire, Predictoor leverages on-chain confidentiality to allow AI models to post and monetize inference results on-chain. The platform is currently focused on AI-based predictions for cryptocurrency prices that are predicted every 5 and 60 minutes into the future. Future prospects are looking to expand to a world-to-world model that predicts everything from predicting the weather, traffic patterns and more.

‍Read more about Ocean's Predictoor on Sapphire  

Creative Applications

Projects have been using development platforms and decentralized compute providers to build AI applications with real world use. Currently, most applications are assistant models like on-chain agents or API interfaces for off-chain AI models.

Highlight: AI Arena

AI Arena is educating the masses on AI development with a fun but competitive game. Players train an NFT character via Imitation Training where an AI learns the game  by copying your actions. It highlights how human-in-the-loop training can work, mimicking RLHF methods by AI researchers.

By leveraging the strengths of both technologies, decentralized solutions have the possibility to benefit everything from finance and computing to creativity and data management. AI development itself is currently accelerated due to the capabilities of how quickly AI could benefit some industries.

With ROFL and other new technologies, the potential for new use-cases with decentralized compute and AI-tech breakthroughs is limitless.