These are the categories that we could likely see further convergence of Crypto x AI, showcasing & solidifying Crypto-native AI use cases

First up, the most obvious synergy—where AI and Crypto converge around what matters most: capital efficiency and yield optimization

DeFi - On-chain Yields

• Stablecoin
• RWA
• Spot & Perps
• Lending/Borrowing
• Yield Markets (Interests / Points)

DeFi has always been the heart of Crypto w/ its on-chain yields & trading opportunities accessible from anywhere across the world. With the rise of AI, these value can be more efficiently captured & optimized, enabling idle capital to be further optimized i.e. Defi can be used as tools to hedge against inflation and/or earn outsized alpha returns from it.

• Stablecoin as one of the top Crypto use cases—as medium currency for literally everything on-chain
• RWA / tokenized real-world assets and bring them on-chain—T-Bill, bonds, real estate, DePIN loans, GPU, etc
• Spots & Perps trading fees / yields
• Lending/Borrowing higher capital efficiency through borrowing & enhancement of yields via lending
• Yield Markets introduction of additional interest rate and yield markets enabling further yield optimization

Crypto = DeFi = people care about making, moving, and multiplying money. Web3 AI could end up doing this better than any closed Web2 system, thanks to the open, incentivized nature of blockchains & tokens.

While still very nascent, we’ve seen exciting progress in DeFAI

• @gizatechxyz hit $1M TVL milestone & generated $6M volume with its hyper-personalized stablecoin yield optimizer agent that enables users to earn yields 83%+ more than traditional lending strategies.
• @Cod3xOrg kickstarting Sophon Spark agent trading competition campaign where Cod3x agents compete for a share of $1.5M rewards (data goes towards improving the agents ability to trade better)
• @autonolas‘ Modius & Optimus that serve as your personal portfolio manager agent. The only team that facilitate agents to run locally on users desktop via “Pearl” agent app store. Recently launched $1M Olas Accelerator program.
• Abstraction layers like @HeyAnonai, @AIWayfinder, @slate_ceo getting better at providing accessibility into DeFi (still very nascent though).

Why agents are good for Defi? They work 24/7 to optimize yields & risks by moving positions around. MCP push the boundaries on Defi / apps integrations allowing agents to tap into on-chain data & more protocols. In a year, we might see agents handle significant number of transactions on chain due to how efficient they’re at automating on-chain activity / optimizing yields.

Things to watch out for

• Teams driving real tech progress and enabling developer ecosystems (hackathons/competitions/workshops)
• Teams that focus on confidentiality/privacy, verifiability, non-custody i.e. human users in control of the agent (truly own the agent)
• Agents growth metrics i.e. Asset Under Agent (AUA) / TVL Under Agent

But beyond DeFi, Crypto AI is also fueling an evolutionary race—one where only the strongest AI agents / teams survive and thrive

Darwinian (Evolution of AI through Natural Selection)

• @opentensor (for Everything AI)
• @AlloraNetwork (for Machine Learning / Predictions)
• @BitRobotNetwork (for Robotics)

Darwinism — “the theory of the evolution of species by natural selection advanced by Charles Darwin.” In other words, the Hunger Games for AI teams. Advance your innovation/tech and get incentives or die.

Web3 AI offer the best infra to facilitate the natural selection process, incentivizing the strong, culling the weak through the use of token incentives, emissions, slashing mechanics, etc

Bittensor popularized this as seen from the number of teams building & pushing technological boundaries on their subnets (especially with SN6, 41, 44 that kickstart the narrative for GambleFAI, utilizing AI/ML prediction capabilities to gain an edge in prediction markets)

Allora been capitalizing on the power of ML, accelerating & enhancing their models across wide varieties of prediction use cases (Allora similar vibes to Bittensor but only for financial prediction use cases, instead of Subnet, Allora has “Topics” which represent a certain financial use case that dev teams can work on… best performing models / teams get most incentives). Best case study was with @steerprotocol for AI-driven LP strategy generating higher returns & lower IL for the position

Bit Robot by @frodobots team, the guys behind @SamIsMoving on @virtuals_io. Not much info is out on this but the team plans to create a Bittensor-like eco but purely for Robotics where each subnet represent each segment of Robotics e.g. data, hardware, vision models, LLMs, etc

Things to watch for

• $TAO performance, dTAO ecosystem growth, and consumer apps/agents leveraging subnets tech.
• Allora integrations, case studies, and its TGE.

Decentralized Infrastructure

• Data
• Model Creation / Training
• Verifiability
• Confidentiality
• DePIN (GPU)

Infra that supports open collaboration, open innovation, and prevents innovation from staying within the hands of couple of centralized players. Talked about some of the sectors of infra in my past articles and I think we’ll continue to see continued adoption of these infra especially as DeFAI & Darwinian progress and we see more mature/clearer use cases come out

What interests me the most for short to medium term

Social & sentiment data:

• @KaitoAI Yap Leaderboard and their recently launched Yaps Open Protocol which allow team to build on top of Yaps scores
• @aixbt_agent tracking & mapping project alpha / social trends based on Twitter
• @cookiedotfun offering market / social intelligence on AI agents

Onchain data:

• Haven’t seen clear leaders to the certain extent as social & sentiment data segment

Other Data players:

• Scraping: @getgrass_io leverage unused bandwidth to collect data
• Ownership: @vana incentivizing data ownership through DataDAOs
• Confidentiality: @nillionnetwork blind compute, their use cases, and the upcoming $NIL TGE (soon)

More read on Data:

For DePIN (GPU):

• There are two interesting protocols emerging that enable the financialization of GPU, allowing data centers / operators to hyperscale their GPU operations through on-chain loans.
• Since there’ll always be demand for more compute due to the growth of AI, data centers are always in need of capital to scale their ops, projects like @gaib_ai and @metastreetxyz connect Defi liquidity with borrowers demand—bringing DePIN yields on-chain while offering capital for operators to scale.

Gaib AI Dollar:

MetaStreet USDAI:

Bottom Line

Crypto-native AI solves challenges that Web2 AI can’t.

Crypto AI provides financial rails that allow agents to not just think but transact. It gives AI the infrastructure to handle assets, optimize capital flows, and operate autonomously on open, permissionless networks.

Crypto AI empowers a world where agents can:

• Move capital across DeFi without centralized gatekeepers
• Tap into decentralized data streams that Web2 players don’t have access to
• Leverage open models and collaborative ecosystems to evolve faster than closed silos

In short, Crypto AI unlocks use cases that Web2 AI simply can’t replicate at scale: programmable money meets autonomous agents, fully verifiable and composable. As DeFi, Darwinian AI, and decentralized infra mature, expect to see AI not just assisting but fully participating in the on-chain economy.

It’s not just about building smarter agents—it’s about enabling them to own, trade, optimize, and create value directly on-chain. That’s the moat.

Web2 kickstart the AI race. Web3 accelerate the autonomous Agentic Economies

Disclaimer:

1. This article is reprinted from [0xJeff]. All copyrights belong to the original author [0xJeff]. If there are objections to this reprint, please contact the Gate Learn team, and they will handle it promptly.
2. Liability Disclaimer: The views and opinions expressed in this article are solely those of the author and do not constitute any investment advice.
3. The Gate Learn team does translations of the article into other languages. Copying, distributing, or plagiarizing the translated articles is prohibited unless mentioned.