AI-Agent Adoption — August 2026

Executive view

August strengthened the infrastructure case for the onchain agent economy, but not yet the broad economic-adoption case. Payment rails, identity standards, data access, cross-chain execution and agent tooling are becoming usable production infrastructure. Activity, however, remains concentrated, and headline registration and transaction counts substantially overstate the number of economically active autonomous agents.

GLI24 Agent Adoption Score: 7.2/10 for infrastructure maturity. Real economic breadth remains closer to early-stage adoption.

Throughout this note, figures drawn from project disclosures, documentation or public explorers are described as reported or observed. They are not independently verified by GLI24.

1. Market reality check

x402: organic economic volume increased from roughly USD 428k in July to about USD 503k in August (+17% month-on-month), while economically relevant sellers fell from 173 to 124 and the top ten wallets represented roughly 71% of organic volume. That is genuine growth in machine-payment activity, in a concentrated market rather than mass adoption.

Raw x402 counts are far larger and should not be equated with independent AI-agent purchases. A current onchain dataset reported roughly 175.3M cumulative x402 transactions and USD 41.5M settlement volume across 12 chains by 30 August, against about 160M cumulative transactions at the end of July. Treat these as protocol activity, not as a count of autonomous agents.

ERC-8004: registrations accelerated sharply during 2026, reaching roughly 526,600 reported registrations across 24 chains by August. An analysis of 25,627 Ethereum and Base registrations found only 3,334 reachable advertised endpoints, and only 44 agents whose reputation could be connected to a paid onchain job.

The distinction that matters: Registered Agents does not equal Active Agents, does not equal Paying Agents, does not equal Profitable Agents.

2. GLI24 ranking — exposure to a growing agent economy

This is an Agent Exposure score, not a general investment ranking. BTC, for example, belongs in GLI24 primarily as a reserve and institutional crypto asset, not as an AI-agent token.

3. Highest-conviction observations

LINK

Chainlink currently has the most complete agent-infrastructure stack in GLI24: Chainlink for Agents combines Data Streams, CRE, CCIP, transaction management, gas abstraction and x402 pay-per-call. A public x402-indexed endpoint showed tens of thousands of calls but only one unique payer, so the technology is real while broad product-market fit remains unproven. Thesis impact: clearly positive, but not proven mass adoption.

VIRTUAL

Currently the cleanest GLI24 exposure to agent-to-agent commerce through ACP. Reported onchain activity is meaningful enough to justify monitoring jobs, repeat counterparties and agent revenue, but reported jobs and revenue must not be treated as independently verified economic output. Thesis impact: positive.

NEAR

One of the strongest relative upgrades this month. NEAR Intents and Chain Signatures make cross-chain agent execution more credible, and Shade Agents provide a framework for autonomous, TEE-backed financial automation. The missing metric is how much Intent volume is actually generated by autonomous agents. Thesis impact: materially stronger.

FET

Agentverse and ASI infrastructure remain one of the most direct agent-economy bets. Fetch reports roughly 2.7M agents in its Agentverse directory, but directory entries are not economically active agents. The chain of evidence to watch is active agents to paid A2A interactions to fees and staking to sustained FET demand. Thesis impact: positive, still requires better value-capture data.

GRT

Strategically one of the best-positioned data layers. The Graph can structure ERC-8004 identity, MCP/A2A endpoints, skills, feedback, validation and x402 support, and x402 pay-per-query is live. The most important public monetization baseline remains extremely small: 29 identified agents, 222 paid queries, 20 repeat agents and USD 2.22 revenue in the cited measurement. Technology leadership does not automatically create token demand. Thesis impact: strategically stronger, economically unproven.

SOL, POL, XLM, XRP, ETH

Increasingly relevant as execution and machine-payment rails. Solana and Polygon show significant reported x402 activity; Stellar now supports x402/MPP agentic payments and is a premier member of the x402 Foundation; XRPL has early x402 payment support; Ethereum remains the settlement and security environment around much of the EVM and ERC-8004 stack. The open value-capture question is whether stablecoin-heavy machine payments generate enough native-token fee or security demand to matter economically.

VVV

An important tokenomics improvement. Venice announced that 5% of purchased API credits are used to buy and burn VVV, while annual VVV emissions are being reduced. This creates a clearer path from inference and API demand to token demand. The missing metric is agent API consumption versus human API usage.

AAVE, UNI, ENA, ONDO, CC

Second-order agent-economy beneficiaries. Their importance rises sharply if agents begin autonomously allocating meaningful capital rather than merely buying API calls. Aave already appears in real agent workflow examples, while ONDO and CC represent tokenized-asset and institutional-finance destinations. Today the measurable agent-specific share of their activity remains small or unavailable.

TAO, RENDER, NIGHT, ETHFI, XDC

Optionality retained. TAO and RENDER provide intelligence and compute primitives, but August data does not cleanly isolate blockchain-agent demand. NIGHT has strategically relevant privacy and selective-disclosure properties without measurable production agent adoption. ETHFI could benefit from automated treasury and staking allocation, but no robust agent-specific evidence is available. XDC has early x402/gasless settlement and agentic-finance work, still at an early stage.

4. GLI24 stack view

• Direct Agent Infrastructure: LINK, VIRTUAL, FET, GRT, NEAR
• Machine Economy Rails: SOL, POL, ETH, XLM, XRP
• Resources: TAO, RENDER, VVV
• Agentic Finance / Institutional Destinations: AAVE, UNI, ENA, ONDO, CC, ETHFI, with ENS and NIGHT as identity and privacy infrastructure, and XDC and ARB as additional execution and finance optionality

5. What changed in August

• Upgrade — NEAR: more credible agent execution and chain-abstraction architecture.
• Upgrade — XLM: x402 and MPP support plus x402 Foundation governance make Stellar a serious machine-payment rail.
• Upgrade — VVV: API-credit-linked buy-and-burn materially improves the value-capture thesis.
• No conviction upgrade yet — GRT: infrastructure improved, monetized agent usage is still tiny.
• No mass-adoption declaration yet — LINK: strong product stack, extreme public payer concentration.

6. GLI24 conclusion

August strengthens the core GLI24 AI/agentic thesis, but it does not justify changing index weights on agent data alone. The infrastructure question is increasingly answered: much of the required stack now exists. The next question is economic.

The five leading indicators for the coming months:

• Organic x402 volume, unique payers and sellers, and concentration.
• ERC-8004 agents with reachable endpoints plus independent paid jobs and reputation.
• Chainlink for Agents unique payers and real CRE/CCIP economic notional.
• GRT paid agent queries, repeat customers, revenue and actual GRT value capture.
• Agent-driven financial volume through Aave, Uniswap, NEAR Intents, Virtuals and tokenized-asset venues.

Only sustained acceleration in these metrics would justify moving the framework from infrastructure build-out to economic agent adoption. August 2026 remains the transition between those two states.

Sources

Official project and specification sources:

• The Graph documentation and The Graph forum — agent data, x402 pay-per-query monetization
• Chainlink documentation — Chainlink for Agents, Data Streams, CRE, CCIP
• Fetch.ai — Agentverse and ASI agent directory material
• NEAR documentation — Intents, Chain Signatures, Shade Agents
• Solana — x402 and agentic payments documentation
• Polygon — x402 documentation
• Stellar — agentic payments, x402/MPP support, x402 Foundation membership
• XRPL documentation — early x402 payment support
• Venice — VVV tokenomics update (API-credit buy-and-burn, emissions reduction)
• ERC-8004 specification — trustless agents registry standard
• x402 — protocol documentation

Independent measurements: the x402 organic-volume, seller-count and concentration figures, the cumulative x402 transaction and settlement dataset, the ERC-8004 endpoint-reachability and paid-job analysis, and the cited GRT paid-query baseline are third-party or explorer-derived measurements. They are reported as observed and have not been independently verified by GLI24.

Disclaimer

Research only. Nothing in this analysis is investment advice, a solicitation or an offer. The Agent Exposure score is thematic research, not a price target, a forecast or an allocation recommendation. GLI24 is a portfolio-weighted thematic research index and not an investable product. This analysis does not change GLI24 constituent weights or index methodology.