GROWING AI AGENTS WITHIN L2 TEAMS - METIS' TOM NGO

A conversation with Tom, CEO of Metis

“It’s easy to make simple things complex, but it’s hard to make complex things simple.”

Guest Introduction

Tom is the CEO of Metis, one of the earliest scaling solutions built for Ethereum. A former trader and investor across equities, commodities, and forex, Tom brings a market-driven mindset to blockchain infrastructure, with a focus on decentralization, ownership, and now AI-native systems.

Key Takeaways

• Metis has been building toward AI-native infrastructure for over 18 months, not reacting to hype
• LazAI focuses on verified, monetizable data through the DAT standard
• Hyperion and Andromeda are unified by one vision, serving different performance needs
• AI agents are central to abstracting blockchain complexity for mainstream users
• Long-term success means usable technology that disappears into daily life

Introduction

Metis has made a bold statement this year: it is no longer just an L2.

In this AthenaX Roundtable, Tom explains why Metis has been quietly building toward AI-native infrastructure for over 18 months, how LazAI, Hyperion, Andromeda, and Lazbubu fit into one unified vision, and why experience, verified data, and abstraction are the next frontiers for blockchain adoption.

This conversation explores how AI and blockchain intersect—not as hype, but as infrastructure.

Metis Is No Longer Just an L2

Tom explains that Metis has always been more than a single product.

“We have always been more than just a product. We’re a vision, an intent to really contribute to personalization of ownership, truth, and decentralization.”

The open letter published by Metis was not a pivot, but a reaffirmation.

It highlighted a focus on AI that had already been underway—grounded in blockchain values, not trend-chasing.

Metis has been building a modular tech stack, with Hyperion and LazAI already in testnet. The goal is speed, especially for on-chain inference, where legacy blockchain architectures fall short.

Experience Over Complexity

A recurring theme throughout the discussion is experience.

“It just needs to work. It just needs to feel natural, not complex like how blockchain has been.”

Through LazPad and Lazbubu, Metis is intentionally blurring the line between technology and experience—making interaction intuitive, playful, and accessible, even when the underlying tech is advanced.

This philosophy showed real traction:

• ~100 hackathon participants
• 90,000+ signups
• Strong industry reception to LazAI-focused tooling

AI Agents and Abstracting Blockchain Complexity

One of the strongest use cases discussed is AI agents abstracting blockchain UX.

“AI agent, you can really abstract all that with the right kind of framework. Bridging, swapping, signing transactions.”

The long-term vision is not users choosing chains—but agents acting on their behalf, safely and transparently.

Leadership, Innovation, and Staying Convicted

Tom emphasizes that innovation requires conviction—especially during narrative shifts.

“Don’t chase the narrative, because it might be short term. Build in what you believe.”

He draws parallels to companies like Kodak and BlackBerry that failed to evolve, stressing the importance of listening to the market while staying grounded in vision.

Q\&A Highlights

Q1: How do you usually introduce yourself to people meeting you for the first time?

Tom:

“Usually I tell people that I first meet, I’m the CEO of Metis. We built one of the first scaling solution for Ethereum. At our heart and our drive for our company and organization is embracing the ethos of blockchain.”

Q2: Why announce the AI shift now, after building for so long?

Tom:

“Our testnet for Hyperion is in testnet. LazAI is also in testnet. So you just can’t build this within four weeks. We’re building a tech stack, a modular kind of tech stack to really support different developers and their needs.”

Q3: What does it mean for Metis to be an AI-native ecosystem in practice?

Tom:

“We’re building something where developers can really take that and build a high-speed game, advanced deep-in network. They can integrate their data set on chain and utilize LazAI to validate the data and also record it on chain.”

Q4: What is LazAI and how does it relate to Metis?

Tom:

“Metis has been the core building infrastructure for blockchain. We brought in new talent, new AI researchers, a new team to build that tech stack out.”“We’re looking to build LazAI as a unified experience with Metis.”

Q5: What problem does LazAI solve with data?

Tom:

“Without data, there is no model. People don’t know how to monetize data. They don’t know the value of data.”“Blockchain is one of the technology that can provide verified data. It’s transparent, it’s open.”