Blockchain Oracles are at a Crossroads, and AI is About To Change Everything
Oracles have always served as the bridge between blockchains and the real world, but in 2025, they’ve become something far more critical. Once limited to price feeds, oracles now underpin DeFi automation, enable intent-centric smart contracts, and increasingly interact with AI agents making decisions on behalf of users and DAOs.
This means that the blockchain oracle industry has evolved into a competitive arms race: speed, decentralization, trust-minimization, cross-chain interoperability, and now, AI-powered decision-making. Oracles are powering everything from real-time CDP liquidations and cross-chain messaging to automated intent resolution and autonomous agents.
As the DeFi landscape matures, blockchain oracle protocols like Chainlink, Pyth, UMA, Chronicle Protocol, and Supra are racing to re-architect this crucial layer. Some focus on publisher networks or optimistic dispute models; others pursue threshold signatures, low-latency consensus, and automation.
At Supra, we began our journey as an oracle protocol, focused on solving core issues like latency, manipulation, and decentralization. But as DeFi matured, we realized that oracles alone weren’t enough. Smart contracts needed faster consensus, secure automation, and native support for AI agents.
That’s why Supra evolved into a fully interoperable Layer-1 blockchain, with Threshold AI Oracles as a core part of our decentralized, vertically integrated tech stack.
In this article, we explore the state of blockchain oracles — from simple data feeds to the AI-native middleware powering intent-based DeFi.
The State of Blockchain Oracles in 2025: Who’s Leading?
Top blockchain oracles by TVS, June 2025. Source: Defillama.
The oracle market has consolidated significantly since the early DeFi boom, but it’s also evolving faster than ever. According to DefiLlama’s oracle leaderboard, Chainlink remains the undisputed leader with over $65 billion in Total Value Secured (TVS), while Chronicle Protocol — a rising contender that serves as the primary oracle for Sky (formerly MakerDAO), ranks second with over $7 billion. Other active players include UMA, Pyth, and Supra, which currently secures over $873 million in crypto assets and is live on 67 mainnets, placing it at #9 by TVS.
Notably, several early oracle protocols have lost ground. Band Protocol, once a top competitor, has fallen to #14 with just $143 million in TVS. DIA, despite early traction and significant partnerships, now ranks #20, securing less than $50 million. The rapid turnover highlights how the oracle market is becoming more competitive — and how newer designs can overtake legacy players if they address the right pain points.
And those pain points are becoming more visible.
While not as large as some of the oracle exploits seen in 2021 and 2011 (like the $112 million Mango Markets exploit), recently, a new wave of oracle-related exploits has rocked the industry.
The April 2025 KiloEx attack leveraged a price manipulation vulnerability that led to $7.5 million in losses, while, in August 2025, the Vow token contract was hacked, and its internal oracle was manipulated, leading to $1.2 million in losses.
These attacks didn’t just exploit poorly designed smart contracts — they revealed that oracle architecture itself must evolve to defend against manipulation, latency, and cross-chain inconsistencies.
To stay relevant, modern oracle protocols must now deliver on three fronts:
Speed, for low-latency DeFi actions like liquidations and arbitrage.Decentralization, to minimize single points of failure.Cryptoeconomic security, ensuring data accuracy through economic incentives and verifiable computation.
In this context, projects like Supra — combining high-throughput consensus, AI-native oracles, and smart contract automation — are stepping into the void left by older, slower middleware.
Threshold AI Oracles: Bringing Real Intelligence On-Chain
Traditional oracles were designed to relay facts — like ETH/USD prices or gas fees — but the DeFi and AI ecosystems of 2025 demand something more sophisticated: oracles that can interpret, deliberate, and act in real-time. That’s the driving force behind Supra’s Threshold AI Oracles, a new class of decentralized infrastructure designed to bring intelligence on-chain.
Rather than relying on a static data feed or a single validator, Supra’s Threshold AI Oracles use a multi-agent committee orchestration model.
Here’s how it works:
First, a decentralized committee of nodes is selected and randomized on demand to retrieve, verify, and evaluate off-chain or cross-chain data.Each node operates a lightweight AI agent capable of deliberating based on local context, historical data, or user-defined parameters.Instead of relying on a single output, the system uses threshold cryptography to generate a final consensus output only when a quorum of agents agrees.
This process ensures that no single agent, oracle, or attacker can manipulate the result. It also dramatically reduces latency, bias, and centralized points of failure, problems that continue to plague older oracles in high-stakes environments like lending, liquidations, or intent-based execution layers.
Because these agents can process more complex data (like volatility metrics, intent models, or event triggers), they’re especially well-suited for integration with AI agents, DAO governance, and autonomous DeFi applications. Essentially, they bridge not only off-chain to on-chain, but also human intention to smart contract execution.
Supra isn’t alone in exploring this convergence between oracle design and agent-based systems:
Bittensor (TAO), while not a traditional oracle, has pioneered a decentralized network of AI subnets. Some subnets are experimenting with economic-weighted consensus to deliver off-chain data, effectively functioning as AI-native oracles.Autonolas is building decentralized autonomous services (DAS) that allow multiple agents to collaboratively make decisions off-chain, including for governance, MEV mitigation, or data attestation. Their agent-to-agent messaging framework is especially relevant for future oracle coordination.Tellor has evolved from a basic oracle into a flexible, customizable data marketplace where users can specify data queries and dispute outcomes. While not AI-native, this design allows greater configurability and is often cited in discussions of oracle modularity.
Intent-Based DeFi and Autonomous Agents
A growing number of developers are rethinking the way we build decentralized applications, shifting from direct user-initiated transactions to systems that respond to intent. Instead of clicking “swap” or “borrow,” users express what they want to achieve (e.g., “maximize stablecoin yield over the next 30 days on Ethereum and Ethereum L2s”), and autonomous agents figure out the best execution path.
But here’s the catch: most oracle systems today can’t support intent-based DeFi. They’re too slow, too centralized, and too rigid — designed for simple data retrieval, not real-time negotiation between agents or complex automation logic.
That’s where oracles like Supra’s come in. With zero-block-delay automation, Supra’s infrastructure allows for instant liquidation triggers, cross-chain arbitrage execution, and on-chain intent resolution — all in sub-second timeframes. Rather than waiting for multiple confirmations or external scripts, agents can execute trust-minimized strategies in real-time.
This shift doesn’t just improve performance — it opens the door for a new class of apps: AI-native protocols, fully autonomous DAOs, and intent routers that optimize capital without constant user oversight.
Protocols like CowSwap and UniswapX already rely on intents to match orders, while Anoma and Skip are rebuilding the entire stack around intent-centric execution. Supra’s zero-block-delay architecture complements these systems by enabling sub-second oracle automation in intent resolution.
The Road Ahead: AI, Security, and Cross-Chain Expansion
As DeFi, AI, and cross-chain ecosystems converge, oracle infrastructure must evolve to meet three emerging demands: intelligence, automation, and interoperability.
We’re already seeing early signals. Protocols like UniswapX have introduced off-chain quoting mechanisms that rely on trusted execution environments (TEEs). As previously mentioned, the Bittensor network has launched AI-powered oracles that weigh inputs based on reputational staking. And intent-based transaction routing is becoming the norm across rollups, appchains, and Ethereum L2s.
But while these features advance specific parts of the stack, only cross-chain oracles — and especially AI-enhanced oracles — are positioned to support all chains and ecosystems.
That’s where Supra’s architecture comes in. While Threshold AI Oracles are designed to serve any chain, Supra’s native stack goes even further:
Moonshot Consensus: a high-speed, low-latency finality engine for real-time automation.Containers: modular, isolated blockspace optimized for AI agents, DAOs, and DeFi subprotocols.HyperNova: a bridgeless, trustless cross-chain messaging protocol.
These advanced features are tailored for builders deploying directly on Supra’s L1, offering an end-to-end environment for intent-based and AI-native dApps.
Together, these components support an emerging class of DeFi: one that’s not just faster, but also smarter, safer, and capable of reasoning in real-time.
The bottom line? In 2025, winning DeFi applications will rely on smarter oracle protocols — not just to scale, but to coordinate AI agents, intents, and cross-chain actions in real-time.
Learn more about Supra at Supra.com, or check us out here on X.
Blockchain Oracles in 2025: AI and Intent-Based DeFi was originally published in Coinmonks on Medium, where people are continuing the conversation by highlighting and responding to this story.