Written by: blocmates. (@blocmates) Compiled by: AididiaoJP, Foresight News

If you are like us and have been firmly tied to this industry in the past few years, you will obviously feel that the atmosphere has changed.

Things feel less exciting, and the only things that seem to grab attention are things with two words - AI and Agent.

The mainstream consensus is that the industry is being heavily optimized to serve AI agents, leading to the marginalization of products that still rely on direct human interaction or the "human layer".

Therefore, from a human perspective, the industry may appear a bit rigid, but the on-chain environment is still active and full of life on a new layer (Agentic layer) where humans are technically unable to directly intervene.

Efficiency is driving more users toward AI-led interactions. The platform originally designed for human clicks and operations is now being optimized to serve "non-humans".

Big players like Uniswap Labs launched 7 open source "Skills" for AI Agents in February. These tools enable autonomous AI-coded Agents (such as those in Claude, Cursor, or other Agent frameworks) to interact directly and reliably on-chain with the Uniswap protocol.

However, contrary to the "AI Agent will eat everything" argument flying around the timeline, a closer look reveals a slightly different story - the growth of Agent activity is more track-specific than industry-wide.

We decided to dig a little deeper to see which tracks had been "eaten" and which ones were still waiting to be eaten.

Our goal: to understand whether the human layer in encryption is truly dying, and to explore solutions that build on new layers of encryption to ensure control is not lost.

The track that AI Agent has dominated

In certain tracks, we observed a high level of activity driven by AI agents, while direct human interaction declined. Here are some examples:

Derivatives Trading (Perpetual Contracts)

The perpetual contracts market is the clearest bot-dominated liquid market in crypto. Speed, pattern recognition, and 24/7 execution are things machines do better than humans. No one would argue that humans should front trade manually.

The top 10 perpetual protocols generated approximately $592 billion in trading volume over the past 30 days, with Hyperliquid alone accounting for $248.8 billion, followed by Aster ($61.6 billion).

Aster’s “Human vs. AI” real-time trading competition, which ran for two weeks under highly volatile conditions, is a clear example of this: 43% of human participants were liquidated, while all 30 AI Agents completed the competition without liquidation, with a 100% survival rate.

The human trading team had an overall ROI of -32.22%, while the AI Agent limited total losses to approximately $13,000 for an overall ROI of -4.48%.

Arbitrage Trading (MEV)

This is the most absolute case of bot dominance in crypto, because profitable human MEV operators at scale simply don’t exist.

The cross-network MEV ecosystem has evolved into a highly competitive automated trading industry, with specialized robots and infrastructure tools scanning the blockchain memory pool.

In 2025, sandwich attacks accounted for 51.56% of total MEV transactions ($289.76 million). On Solana, sandwich bots captured between 1.7% and 5.4% of total daily trading volume (2.9% on average), executing $3.85 billion in sandwich trades across more than 3.9 million bundles.

A single bot accounts for 42% of all sandwich trading volume, executing more than $1.6 billion in trades in the past 30 days.

This also extends to DeFi protocols. The entire liquidation lifecycle – monitoring, triggering and execution – is handled by permissionless bots.

While this existed before the AI Agent craze, as the DeFAI category continues to grow, the entire process is now significantly automated by Agents.

Yield Optimization

This category is Agent-first by default. Data shows that 68% of new DeFi protocols launched in Q1 of 2026 include at least one autonomous AI Agent for trading, liquidity management, and risk monitoring.

Compared to 12 months ago, we saw a 15% increase in AI Agent adoption in the revenue space.

AI Agent continues to outperform on platforms such as Giza and ZyFAI - the latter achieving a +73.42% return outperformance over static strategies.

Giza has recorded more than 800,000 autonomous transactions and has assets under management of up to US$40 million.

In addition to Giza and ZyFAI, there are many more projects in this category, some of which we have already covered, and others we are happy to cover in depth upon request and further review - including:

Arrakis, Reflect, AFI, Lulo, Sail, Almanak, Surf, Infinit, AXAL, Superform, DeFi Saver, Kamino, Mamo, HeyAnon, etc.

Updates to headlining projects like Pendle, which includes deploying MCP connectors and building Skills that allow Pendle to easily integrate with crypto-native and non-native AI agents, are also evidence that the revenue industry is rapidly moving toward agent-first interactions.

Spot Trading and Portfolio Optimization

Automated trading bots currently account for an estimated 65% of global crypto trading volume. At the beginning of 2026, the number of daily active AI Agents on the chain reached 250,000, an increase of more than 400% compared with 2025.

On Solana specifically, AI Agents generated $31 billion in DEX trading volume in 2025, accounting for approximately 2% of total DEX trading volume ($1.5 trillion).

We are seeing an increase in agent-driven spot transactions, including meme coin transactions across the network.

Users increasingly rely on agent-first infrastructure for token issuance, trading and portfolio management, driving the popularity of platforms such as Virtuals, Bankr, Glider, Surf, Symphony and more.

Battlefield track (Agent + human activities coexist)

Prediction Markets

Polymarket is the most granular testing ground for AI vs. humans in crypto, and the data is hard to refute. We’ve all seen those posts boasting about making millions of dollars in prediction markets using Agent.

However, based on 10,582 active traders, 880 bots (8.3% of accounts) made an average profit of $119,156, compared to $12,671 for humans - a per capita difference of 9.4 times.

Agent achieved a profitability rate of 66.4%, compared to 45.3% for humans. The arbitrage window compressed from 12.3 seconds in 2024 to 2.7 seconds in 2026, with bots executing in sub-100ms capturing 73% of all arbitrage profits.

AI-powered Agents now account for approximately 18% of total prediction market trading volume, and over 30% of Polymarket wallets are already using AI Agents.

Here’s the subtlety, however: for markets lasting weeks or months, the gap shrinks significantly—and in some categories, humans actually perform better.

Bots have been shown to be poor at handling change, so they become confused when fundamentals shift dynamically. Instead, humans adapt.

So what we see is this: the short-term arbitrage game has been taken over by Agents, while the long-term judgment game still belongs to humans.

This infers that Agent activity and human interaction in prediction markets will continue to be balanced for the foreseeable future, until we may have more perceptive models capable of making nuanced decisions where humans still dominate.

DeFi Lending

Lending is another clear example of layered automation. As we mentioned in our Agent-dominated track, liquidation bots are deeply entrenched; however, the vast majority of deposit and borrowing decisions are still made by humans.

Aave leads with $12.4 billion TVL, followed by Morpho ($6.9 billion).

DeFAI Agent redeployed over $2 billion in TVL across lending and yield protocols—impressive in absolute terms, but less than 2% of total DeFi TVL ($130-140 billion).

This is a clear sign that deposit decisions, collateral selection and risk preferences are still primarily a human voice. Although AI agents handle the pipeline at the edge, humans are still in control of the core.

A race track dominated by humans

Stablecoins and card payments

As of March 2026, the total market value of stablecoins is approximately US$312 billion. Adjusted trading volume (which filters out bot activity, MEV, and wash trades) reaches $28 trillion in real economic activity in 2025. Growing at a CAGR of 133% since 2023.

Sub-$250 stablecoin transfers hit a new high of 5.84 billion in August 2025. We think these are users sending money to family, paying freelancers, or splitting bills. Over 80% of USD-backed stablecoin transactions occur outside of the United States, where Agent adoption is leading the way.

Real people in emerging markets are using stablecoins as USD access and economic hedges, making them directly key to stablecoin market share. In February 2026 alone, transaction volume reached $1.78 trillion.

Additionally, the card payment category is booming due to clearer regulations. The product allows users to spend crypto assets anywhere traditional cards are accepted, with funds remaining self-custody until the moment of purchase.

Only about 5% of this track is Agent driven. The rest is people moving money. Unlike bot-dominated tracks, users here often don’t know or care that they’re using encryption. That's exactly the point.

Wallet

The wallet is the last layer between humans and the blockchain, and it is also a layer that cannot be completely abstracted.

While abstraction attempts have been made, the approval process urgently requires human oversight. Someone has to sign. Someone has to decide whether to trust what's in front of them.

Phantom has over 15 million monthly active users. The entire wallet space is investing in human-centric improvements such as human-readable transaction previews, biometric security, and card-based spending.

The best wallets of 2026 have evolved from mnemonic + string storage containers to complete financial dashboards.

Enterprise-grade Agent wallets in 2026 include budget limits, whitelists, audit logs, and emergency stops—treating Agents as operators with limited permissions rather than omnipotent signers.

Human vs. Agent Authentication Layer: The more Agents, the more important this is

As more and more Agents flood into on-chain activities, the value of proving that you are a human being or that an Agent acts on behalf of a human being becomes increasingly valuable.

There are several projects developing along this line, ensuring that we don't get lost in the matrix of the machine world.

World & AgentKit

First mention: World (formerly Worldcoin - WLD) - These guys have verified over 17 million users via iris scanning Orb hardware.

World describes itself as a response to an AI-saturated world—building the digital infrastructure that makes being human truly count.

It then launched AgentKit. A toolkit that lets AI Agents carry cryptographic proof that they are backed by a unique human via World ID, and integrates with Coinbase and Cloudflare’s x402 protocol for stablecoin micropayments.

t54

Another project we are focusing on is t54, which is building a trust and security infrastructure (often referred to as the "trust layer") for the Agentic economy, a world where autonomous AI Agents handle real-world tasks such as managing money, making payments, and transacting on behalf of individuals or businesses.

Currently, AI Agents transferring real funds are risky (no verification, no accountability, prone to fraud or violation of compliance rules).

t54 addresses this issue with x402-secure, a dedicated trust layer that enhances the x402 protocol to enable secure AI Agent micropayments. x402-secure provides real-time risk scoring through its Trustline Engine and helps detect scams, including prompt injection, to ensure accountability.

t54 provides these guardrails so that institutions and users can truly trust the Agent with their finances.

Self Protocol

These guys are building a decentralized zk-proof human-Agent binding layer on top of ERC-8004 (On-chain Agent Identity).

Self Protocol uses zk technology to anchor each AI Agent to a verified human owner (human proof) without doxxing or data leakage.

It protects against Sybil attacks and enables self-hosted wallets, autonomous actions, and business protocols while maintaining human accountability.

Selfclaw has been integrated with ecosystems such as Celo/Google Cloud, and fee cycles are used to support verified Agents.

Kite AI

Kite is a base L1 (EVM compatible, using Proof of AI consensus) built specifically for the Agentic Internet.

It provides Agent Passport (verifiable identity, delegation, programmable spending rules or guardrails), autonomous stablecoin payments, governance and verification so that Agents can authenticate, transact and collaborate without intermediaries.

Conclusion

Seriously, we're not anti-Agent. The data is clear in the transaction, MEV and revenue areas; the bot has won those rooms and won't give them back.

A head-to-head race with 43% humans liquidated and zero bots liquidated tells you everything you need to know about who owns the speed game.

But data across the web still shows that humans are still doing most of the work that actually touches real life - in payments, identity and verification.

These are the layers that truly create value and truly generate revenue. They share a common characteristic: they require judgment, trust, physical presence, or cultural context—things that are not yet reducible to optimization functions.

We believe teams should not completely abandon building for direct human interaction in these areas and tracks.

Agent currently needs humans more than humans. We think the teams that understand this, and those building Agent and human proof systems, are worth paying attention to.