Eigen Labs Founder: Blockchain Must Catch Up to AI or Agents Will Stall
Sreeram Kannan argues that autonomous AI agents are being held back not by a lack of intelligence but by coordination infrastructure still operating at human speed. EigenCloud (formerly EigenLayer), the rebranded Eigen Labs platform, is his proposed fix, and a new Google partnership is the latest sign that enterprise players agree.
Eigen Labs founder Sreeram Kannan published an argument on May 23 that positions blockchain as the coordination layer AI agents currently lack. In a post titled "Intelligence Is Accelerating. Coordination Has to Keep Up" on the EigenCloud blog, Kannan contends that AI has effectively made raw intelligence cheap and abundant, but the systems agents must work through to get anything done, including contracts, property rights, legal standing, and capital formation, remain slow and human-dependent.
His conclusion: blockchains are the missing coordination layer.
"An agent reaches a decision in a second, then waits three days while a human gets around to countersigning," Kannan wrote. The bottleneck, in his framing, is not compute. It is trust infrastructure.
His thesis rests on the idea that programmable intelligence requires programmable institutions. Traditional institutions verify behavior through audits and reputation. Kannan argues blockchain replaces those mechanisms with cryptographic proof, allowing an AI agent to demonstrate that it performed a specified task using authenticated data and produced unaltered outputs, all verifiable on-chain without requiring a human intermediary to vouch for it. He describes crypto simply as "a technology for trust."
The longer-horizon version of this argument involves what Kannan calls "agentic companies," software entities that hold wallets, own assets, enter contracts, and hire contributors without any traditional corporate structure. Reporting from blockchain.news has quoted Kannan describing this category as a potential trillion-dollar asset class.
He also raises a distributional point: because early AI companies like OpenAI and Anthropic remain private despite massive user bases, the value they generate is largely inaccessible to the public. On-chain agents could, in theory, issue ownership stakes from inception through tokenization, something legacy tech companies never did.
Kannan also addresses where human economic value goes as intelligence commoditizes. His answer centers on interpersonal work: collaboration, leadership, and trust-building. He acknowledges, however, that these qualities do not distribute evenly or fairly, a caveat with particular weight for workers in labor-intensive emerging economies.
The Google Partnership and On-Chain Context
The blog post expands on an interview Kannan gave to a16z crypto, titled "What AI Agents Need Before They Can Act Alone." a16z crypto led Eigen Labs' $70 million Series B, making the firm both a lead investor and a natural platform for the ideas Kannan is now bringing to a wider audience.
The post arrives alongside a concrete infrastructure announcement. Google unveiled a product called AP2 (Agent Payments Protocol), designed to let AI agents transact using instruments ranging from credit cards to stablecoins. Google selected EigenCloud as the verifiable compute and slashing-backed coordination layer for the rollout. More than 60 organizations, including Coinbase and the Ethereum Foundation, are participating. According to The Defiant, the EIGEN token jumped roughly 25% on the announcement, though the date of that announcement relative to this article could not be independently confirmed, and the token remains well below earlier levels.
As of May 23, EIGEN trades around $0.19 to $0.21, giving the token a market cap of approximately $143 to $151 million against an all-time high of $5.65.
EigenCloud's total value locked sits at approximately $6.33 billion according to DeFiLlama, a significant contraction from a reported peak near $18 to $19.5 billion in March 2026. A security breach at Kelp DAO, which reportedly triggered roughly $5.4 billion in withdrawals, and broader market withdrawal appear to account for a large portion of that decline.
The protocol still commands roughly 93% of the total restaking market, which stands near $11.1 billion across all platforms. As of February 2026, the network reported over 1,900 active operators, including Google Cloud and Coinbase.
What This Means Outside the United States
Kannan's coordination problem is not equally distributed. In markets where legal enforcement is slow, banking infrastructure is thin, or contract finality is uncertain, the gap between what AI agents can decide and what they can actually execute is wider than in the US or Europe.
Sub-Saharan Africa recorded more than $205 billion in on-chain transaction value between July 2024 and June 2025, a 52% year-over-year increase, according to Chainalysis data reported by ecofinagency.com. Crypto adoption across Nigeria, Kenya, and South Africa has been driven primarily by remittance needs and inflation hedges rather than speculation. Kenya passed its first comprehensive Virtual Asset Service Providers Act in late 2025.
A May 2026 report from the Digital Frontiers Institute warns that African nations risk using AI-agent outputs without owning or controlling the infrastructure underneath them, a scenario EigenCloud's verifiable stack could, in theory, address.
India presents a different configuration. A TechWire Asia report from April 2026 found that India leads all surveyed countries in agentic AI return-on-investment realization at 50%, with the highest concentration of expert-level agent users globally. Indian fintech and SaaS developers are already deploying agents in production, alongside Brazil, which the same report identifies as a parallel leader in transitioning AI pilots to production.
EigenCloud's verifiable compute layer could serve as the trust layer those pipelines need to clear enterprise procurement. The constraint, for now, remains regulatory: a 30% flat tax on crypto gains and a 1% Tax Deducted at Source (TDS) on transactions continue to suppress domestic on-chain activity despite internationally competitive developer capacity.
What Comes Next
With over 280 crypto-AI projects now requiring some form of verifiable inference, according to BlockEden.xyz, and the AI agents market projected to grow from $7.84 billion in 2025 to more than $52 billion by 2030, according to Blockchain Council projections, the infrastructure question Kannan is raising will not stay theoretical for long. The AP2 rollout with Google is the first large-scale test of whether cryptoeconomic coordination can actually slot into enterprise AI pipelines. If that integration delivers, the ideas in "Intelligence Is Accelerating. Coordination Has to Keep Up" stand to move from a timely blog post to the baseline infrastructure underpinning the next generation of autonomous systems.