yoav igra

[Omer Goldberg @omeragoldberg]

Undercollateralized Onchain Lending

In July, @KintoXYZ was hacked for ~$2M, and the token collapsed.

To recover, the team borrowed $750K through Wildcat Finance at 50% APR, planning to relaunch and repay users.

This week, Kinto shut down, and the loan defaulted.

[Omer Goldberg @omeragoldberg]

$500K gone in 180 seconds.

Chainlink just proved oracles are one of the weakest links in DeFi.

Here’s what happened:

[Chaos Labs @chaoslabs]

1/ Following the launch of the Chaos SVR Monitoring Platform on @Aave, we’ve released an additional dashboard that tracks the timing and accuracy of SVR feed updates.

This extends observability into timing-sensitive oracle dynamics, providing key oversight during market volatility.

[Chaos Labs @chaoslabs]

1/ Introducing the Pendle PT Risk Oracle

A pricing and risk framework that enables secure, capital-efficient integration of @pendle_fi's PTs on lending protocols like @Aave.

Since going live on April 30th, PT-eUSDe-29MAY2025 has surpassed $200M+ in deposits.

[Chaos Labs @chaoslabs]

1/ Introducing the Smart Value Recapture (SVR) Monitoring Platform.

Track real-time metrics on SVR liquidation performance, OEV recapture, and revenue distribution. Now live on @Aave.

Check it out: https://t.co/ei0HRFUPCh

[Omer Goldberg @omeragoldberg]

Introducing: ChaosAI

In crypto, sophisticated traders, institutions, and insiders have always had an unfair advantage.

While you rely on public dashboards and Twitter sentiment, they analyze proprietary datasets and signals you can't access.

Today, we're changing that with Chaos AI—the first AI-powered crypto researcher.

The Blockchain Paradox

Blockchain promised transparency but delivered a paradox: more data than ever, with fewer people able to make sense of it. General AI models can't help here—they fundamentally lack an understanding of CeFi/DeFi market risk, Oracle risk, MEV, liquidity dynamics, protocol behaviors, and the complex interplay between on-chain and off-chain events. This gap in clarity is perpetuating the same information asymmetry crypto was supposed to eliminate.

Bridging the gap between general models and highly optimized crypto-specific AI required embedding our deep domain expertise directly into the models. Through our role as risk managers responsible for billions in crypto assets, we've developed extremely reliable, real-time datasets and signals. We've combined proprietary data with our domain expertise to create the product we wish existed: an AI that genuinely understands crypto markets.

Chaos AI answers questions that matter:
- What's driving unusual market movements?
- How do protocol risks interconnect?
- Which projects show concerning metrics?

For the first time, everyday users can access the same quality of financial intelligence previously reserved for institutions and insiders. This isn't just a product—it's a step toward the transparent financial system we've always envisioned.

The Problem with General-Purpose Models

Models like ChatGPT have shown remarkable general reasoning abilities. They're excellent for broad use cases, from creative writing to generating code snippets. However, their effectiveness in specialized fields like blockchain finance falls short. These general models do not understand critical, domain-specific concepts like on-chain liquidity dynamics, MEV, protocol-specific behaviors, liquidation cascades, and nuanced DeFi and CeFi interactions. Without this specific knowledge, general models stumble precisely where accuracy and clarity are needed most.

Building Chaos AI

Our insight was simple yet powerful: to bridge this gap, we needed to integrate our domain expertise into the models themselves. At Chaos Labs, we've spent years accumulating unique, proprietary datasets, building specialized reasoning models, and validating signals in the real world, powering billions of dollars in transactions and risk management. We had both the unique expertise and the tested, robust data infrastructure necessary to do this right.

Chaos AI was born from these strengths—a specialized AI designed specifically for crypto and financial blockchain data. However, blockchain data is notoriously complex, distributed across many networks, protocols, and ecosystems. We realized that to unlock the power of blockchain intelligence fully, an open-source standard was necessary—one that allows anyone to contribute specialized agents, endpoints, and insights. Initially, this beta will showcase capabilities powered by Chaos Labs' curated reasoning models, proprietary datasets, and internal tooling. However, we deeply believe blockchain intelligence must ultimately evolve into an open standard and ecosystem. As Chaos AI matures, we plan to transition into an open standard, enabling anyone in the community to contribute to the underlying large language models, agents, endpoints, and data networks.

The Impact and Implications

Today, we're announcing the closed beta launch of Chaos AI. Now is the ideal moment to build a dedicated, crypto and finance-specific blockchain model that is open, continuously up-to-date, and thoroughly data-driven. We'll soon share more about our long-term vision and how we plan to engage and empower the broader community.

We're excited to share this innovation with the blockchain community, which will enable you to make accurate, timely, and confident data-driven decisions.

The beta version is open today, and we’re beginning to onboard the first cohorts of users. If you’re a student, researcher, investor, or transact onchain, sign up today for early access.

[Chaos Labs @chaoslabs]

1/ @renzoai is integrating Edge Risk Oracles to protect over $852M of Renzo assets from a range of potential protocol risks.

Edge will enable a real-time safeguard, detecting disruptions, withdrawal failures, and systemic risks before they impact users.

Learn more: https://t.co/zFbN6iagrc

[Chaos Labs @chaoslabs]

https://t.co/em0zI6Jf5u

[Chaos Labs @chaoslabs]

Introducing Artificial Financial Intelligence (AFI)

An open initiative to enable purpose-built artificial intelligence for crypto and onchain finance.

[Omer Goldberg @omeragoldberg]

@ChatGPTapp can write Shakespeare.

@cursor_ai can write code.

But neither can tell you if an onchain protocol is about to collapse.

Introducing AFI—the first AI built specifically for crypto finance.

Why vertical AI changes everything and what @chaoslabs is building towards 👇

[Omer Goldberg @omeragoldberg]

https://t.co/S3iWnD24dw

[Chaos Labs @chaoslabs]

Predictable Fixed Income in DeFi: Our Risk Analysis of Pendle’s Principal Token (PT)

PTs are a novel approach to DeFi, enabling users to achieve predictable, fixed-yield income on their assets over specified time frames.

We worked with @pendle_fi to develop a comprehensive mechanism design and PT risk assessment framework.

[Chaos Labs @chaoslabs]

The Edge Oracle: Adoption Update

• Price — $125B Secured Volume
• Risk — $1.2B Secured Value
• Markets — 74 Live Across 4 Protocols

[Chaos Labs @chaoslabs]

1/ @use_corn, a protocol redefining Bitcoin’s role in DeFi, has integrated Edge Price Oracles!

With over $60 billion in secured volume, Edge delivers intelligent, context-aware price data to the Corn Network, significantly boosting Corn’s resilience and operational efficiency in dynamic market conditions.

[Chaos Labs @chaoslabs]

Smarter, More Efficient, and Risk-Aware Perps Trading.

>> Edge Risk Oracles power @GMX_IO's Dynamic Price Impact updates — enabling real-time parameter adjustments and reduced price impacts for an improved perps trading experience.

[Chaos Labs @chaoslabs]

Edge AI resolved the 2024 U.S. Presidential Election for OutcomeMarket on November 7, 2024 12:08 AM EST

View the tx proofs below.

Congratulations to @wintermute_t, @Backpack, @bebop_dex, @vertex_protocol, @_WOOFi, and @LayerZero_Core for a successful pilot!

Ethereum ✅
https://t.co/UgKRNL7zBn…

Base ✅
Request → https://t.co/VsnkHy6FM3…
Response → https://t.co/W33WPlrOyK…

Arbitrum ✅
Request → https://t.co/ExLL81Xx46…
Response →
https://t.co/6eqW37ooGA…

[Chaos Labs @chaoslabs]

1/ Reliable data resolution has never been more important.

Chaos Labs is excited to announce the alpha release of the Edge AI Oracle, the world’s first truth-seeking oracle, multi-agent oracle.

[Omer Goldberg @omeragoldberg]

1/ With the U.S. Presidential Election less than a week away, prediction markets are taking center stage.

At @chaoslabs, our mission is to build trust and transparency in the markets of the future through high-quality data and risk management tools.

[Chaos Labs @chaoslabs]

1/ Following our recent risk assessment, Agora @withAUSD has integrated the Edge Price & Risk Oracle as its primary oracle to ensure efficient onchain risk management and resilient price data for the $AUSD stablecoin.

🧵

[The Chaos Summit @ChaosSummit]

Join us in Bangkok on Nov 12, 2024, for The @ChaosSummit by @chaoslabs! 🇹🇭

Here’s what you can expect at the event:

→  30+ speakers from projects like @aave, @PayPal, @wintermute_t, and @ethena 🤓
→ Technical lectures and roundtable discussions on Data Integrity, Risk, and AI 💻
→  Local Thai cuisine and wellness drinks 🍛

Apply to attend: https://t.co/rQOWkLD3ND