Layer One: Building AI-native infra with Kite AI's Lei Lei

Layer One•January 22, 2026, 1:18AM EST
Make us preferred on Google
Layer One: Building AI-native infra with Kite AI's Lei Lei
Partner offers

We'd love your feedback.

Advertisement

Welcome to Layer One: the newsletter and podcast on the intersection of crypto and the real world. Hosted by Steven Gates and Kelvin Sparks, Layer One brings you inside the conversations driving blockchain forward.

This week on the podcast, we were joined by Lei Lei, Head of Business Development at Kite AI, to talk about the future of onchain AI agents and the infrastructure developments needed to turn sci-fi speculation into near-future reality.

AI-native infrastructure the next step in expanding onchain AI agent use cases

Despite lingering fears of a bubble forming in the artificial intelligence markets, major AI equities have seen continued growth, while start-up funding has kept flowing strong. The Financial Times reported that US VCs invested upwards of $161 billion into AI companies in the year ending October 2025.

In the cryptosphere, the AI token boom followed a similarly explosive trajectory, peaking at a $25 billion market cap at the beginning of last year then closing the year at around a quarter of that value. This trend was buoyed by AI evangelists on crypto Twitter, predicting a new era of autonomous agents transacting directly with one another onchain, and dominating network activity.

In a previous edition of the newsletter, we argued that this has largely failed to materialize. And this week Lei Lei, Head of Business Development at AI-native L1 project Kite AI, gave us some insights into why this has been the case. In short, there remain some major obstacles when trying to reliably run autonomous agents on the blockchain: namely payment safeguarding and trust.

General-purpose blockchains have proven too expensive and unwieldy to provide a reliable base layer for scalable agentic networks. Network congestion drives up fees, which undermine the economic viability of agents that may need to conduct hundreds or thousands of transactions each hour.

Likewise, Steven Gates pointed out the need for “higher standards of accuracy” when multiple agents are working in tandem as a swarm, otherwise a single hallucination may proliferate throughout the group and amplify into chaotic results. Without such safeguards, many complex use cases will remain unviable.
 
These are the core problems Kite AI is setting out to solve, by building their own L1 chain on top of the Avalanche C-Chain architecture, custom-made for machine-native transaction and data flows. The firm raised over $20 million last year alone, with investors including Coinbase Ventures and Paypal Ventures.
 
Our everyday tasks — like going to buy groceries, ordering Uber Eats — all of these can be done by your AI agents. Unleashing these AI agents obviously increases human productivity. But in order to unleash that, we need to break through some of these barriers.
– Lei Lei
 
Kite AI makes use of the x402 web payment protocol: a joint initiative between Coinbase and Cloudflare launched in May 2025, built to allow AI agents standardized pay-per-call access to data and other web resources. Later in the year, Coinbase likewise assisted Google in releasing their own open-source agentic payments protocol, covering stablecoins and card networks.

With major players moving to establish these frameworks for reliable, predictable agent-to-agent transactions, Kelvin predicted that “the next wave of non-institutional capital is probably going to come from these AI agents transacting onchain”. We may be closer than we thought to a world in which AI agents can reliably manage the minutiae of our daily, financial and onchain lives.

Lei Lei in turn likened this to other major advancements throughout the history of human work, akin to the development of agriculture, or the new machine-driven economies of the industrial revolution. But it’s worth noting that during both of those surges of progress, there were many individuals on the losing end of the equation. Progress isn't always distributed equitably.

A January report from consultancy firm McKinsey stated that 84% of US employees expect to be using generative AI for a third of their daily tasks within five years. This could pose major problems for the incoming cohorts of new graduates who would normally progress into minor organizational roles — roles which may largely cease to exist in the near future.

On a more optimistic note, our hosts offered up the suggestion that AI tools may drive down the barriers for entry for entrepreneurship. We may well see entirely new types of opportunities emerge, in a world a small two-person startup, backed up by swarms of autonomous agent assistants, can match the output and efficiency of major companies.

Podcast Recap: Building AI-native blockchain infrastructure with Kite AI's Lei Lei

In the latest edition of the Layer One podcast. We were joined by Lei Lei, Head of Business Development at AI-native L1 project Kite AI, to discuss how AI agents are primed to bring a massive leap in human productivity.

Subscribe to Layer One on Youtube, Apple, Spotify or wherever you get your podcasts.

In the Headlines: The stories driving the conversation this week

  • Avalanche has launched Build Games: a six-week development contest with $1 million in prizes. Applications for the contest are open now, with submissions welcome for any kind of project from core infrastructure to experimental consumer DApps. Successful applicants will gain access to experienced industry mentors, wide exposure and potentially hundreds of thousands of dollars in funding for their idea.
  • California startup Human API envisions a world in which humans can earn through completing tasks delegated to them by AI agents. The firm is building a platform which will allow AI agents to recruit humans for parts of complex tasks which are not machine-accessible, such as performing deliveries or accessing offline data. If this sounds familiar, that’s perhaps because it’s a core plot point in 1984 cyberpunk classic Neuromancer.
  • Analytics platform Nansen has launched its own AI-powered trading solution across Base and Solana. In a spin on the “vibe coding” trend, Nansen’s platform works on a system of “vibe trading” where users submit their requests in plain language. They can then sign off on various trade suggestions drawn from the platform’s vast onchain analytics engine.

Top of the Charts: AI use growing among the world's biggest firms

Keep up with the latest in tokenization, DeFi, and institutional adoption by subscribing to Layer One's weekly market insights.

Layer One is brought to you in collaboration with Avalanche.


Disclaimer: The Block is an independent media outlet that delivers news, research, and data. As of November 2023, Foresight Ventures is a majority investor of The Block. Foresight Ventures invests in other companies in the crypto space. Crypto exchange Bitget is an anchor LP for Foresight Ventures. The Block continues to operate independently to deliver objective, impactful, and timely information about the crypto industry. Here are our current financial disclosures.

© 2026 The Block. All Rights Reserved. This article is provided for informational purposes only. It is not offered or intended to be used as legal, tax, investment, financial, or other advice.