TL;DR
- The Revenue Engine Is Already Running and Accelerating: Protocol fee capture has scaled from near-zero to over 30% of fees in a matter of months, with a meaningful share of all lifetime net revenue earned in the most recent period. Live and verifiable at revenue.near.org.
- com Is the Biggest Revenue Lever in the Stack: NEAR’s owned front-end captures 100% of its fees and routes them to NEAR buybacks. Volume and fee generation have grown several-fold since April, ahead of internal targets. Every percentage point of total Intents flow that shifts toward near.com improves the blended capture rate across the entire protocol.
- Confidentiality Is Becoming Default Behavior: A large and growing share of near.com swap volume now runs confidential, with confidential TVL past $30 million. This is a differentiated fee surface that competitors cannot easily replicate.
- Deflation Is a Measurable Threshold, Not a Narrative: Net deflation is a defined function of protocol revenue against issuance, and the threshold is already being crossed on volatile days in a down market, before any cyclical tailwind.
- NEAR Protocol Is Getting Significantly More Powerful: Dynamic resharding and post-quantum cryptography just went live on testnet, making NEAR the first major L1 with one-click PQ key rotation and hardware wallet support from day one.
- The Next Wave Extends Beyond Today’s Fundamentals: NEAR’s current traction rests largely on Intents. The agent market, confidential inference, IronClaw, and the institutional volume ramp represent a separate layer of potential value accrual that sits on top of an engine that is already working.
Most foundational technology bets spend years looking early. The ones that were right eventually reach a point where the thesis and the traction converge, and it becomes difficult to argue against. In 2017, Illia Polosukhin was a machine learning researcher at Google when he co-authored a paper that would quietly rewrite the trajectory of the entire technology industry.
That same year, he and Alex Skidanov, who had built one of the first production-grade distributed databases at MemSQL, joined forces with a specific goal: build AI models that could write code from natural language descriptions. In attempting to pay global contributors for that training data, they ran into a coordination problem that existing financial infrastructure couldn’t solve at scale. The answer was blockchain. NEAR was born not as a crypto project, but as an attempt to build the infrastructure layer for a world where humans and machines transact at scale.
Everything since has been a straight line toward that original problem. Sharding and mainnet gave NEAR the throughput to operate at global scale. Chain Abstraction, a concept NEAR named before the industry had language for it in 2023, became operating infrastructure by 2025. Chain Signatures and Intents launched in 2024, giving humans and agents the ability to act across any chain without fragmentation.
2026 is where the line ends and the proof begins. Confidential Intents, IronClaw, and quantum-safe scaling aren’t new bets, they are the components of an architecture sketched nine years ago arriving in production simultaneously. The thesis hasn’t changed. The infrastructure to execute it is now live. What’s emerging is a new category: AI money. Private, programmable, agent-native financial infrastructure built on real revenue, real transactional volume, and a market that is still being created, one NEAR is uniquely positioned to define.
Where NEAR Is Now: Tokenomics and the Path to Deflation
Revenue mechanics
At the end of February 2026, NEAR activated the fee switch. Since then, a portion of every dollar flowing through Intents is taken as protocol revenue and used to buy back NEAR. Four main metrics tell the story as a single compounding chain: volume generates fees, capture rate determines what NEAR keeps, margin measures how efficiently each dollar of flow becomes revenue, and offset tracks how much of NEAR’s own emissions that revenue is neutralizing.
Volume is the starting point: $20+ billion cumulative across NEAR Intents, with daily flows continuing to increase especially in times of volatility. The lifetime capture rate average sits at 11.5%. Over the trailing 90 days it’s 17.9%. The trailing 30 days is 30.5%, roughly 2.7x acceleration in a single quarter. Margin is rising, meaning more net revenue per dollar of swap volume. To view live revenue metrics and capture rate visit revenue.near.org.
NEAR is getting better at capturing the revenue already flowing through the system. The same volume is generating meaningfully more protocol revenue than it was 90 days ago. As institutional and agent volume arrives the compounding effect on each of these metrics accelerates with it.
Institutional demand
The demand-side structure around NEAR is broadening across multiple channels simultaneously. Sovereign (SVRN), a NASDAQ-listed company, is accumulating toward 10% of NEAR’s total supply – a long-duration conviction position that anchors the strategic holder base, with other funds building alongside it. Institutional validators are earning native yield through staking, adding another layer of programmatic demand. On the listed vehicle side, five products are already live: Bitwise and 21Shares both offer NEAR Staking ETPs, Virtune offers a Staked NEAR ETP, Grayscale has the NEAR Trust (GSNR), and Valour carries NEAR SEK. Two additional filings are pending regulatory approval: a Bitwise NEAR Strategy ETF and a Grayscale NEAR ETF. The infrastructure for institutional capital to enter NEAR at scale, across custody, yield, and regulated exposure, is largely in place.
Where NEAR Is Going: The Stack and the AI Money Category
The transformation thesis
Every major computing era has been defined by what it connected. The internet connected information. Blockchains connected ownership. The agent economy will connect intention to action, and NEAR is the infrastructure that makes that possible at scale. Today, both traditional commerce and crypto require users to navigate several layers of friction to reach an outcome. The agent economy collapses this friction into a single flow. A human expresses intent. NEAR AI understands it. NEAR Intents executes it. Outcomes arrive without the user ever touching the underlying infrastructure. This isn’t a UX improvement, it’s a structural shift in how value moves. AI sits on the front end interpreting what people want; blockchain settles on the back end with finality and privacy. Together they turn both crypto and traditional commerce into outcome-driven, agent-matched activity, and every transaction in that new world runs through NEAR.
AI Money as a category
Every major technology category eventually settles on a default. You don’t deliberate over which search engine to use, which cloud to deploy on, or which tool to message a co-worker. The category and the default become synonymous: Google, AWS, Vercel, Slack. The strategic question for NEAR isn’t whether AI money becomes a category. It’s whether NEAR becomes the default when it does.
Start with what makes something a standard in the first place: not the best feature set at a moment in time, but the deepest integration across the stack that makes switching costly and adoption self-reinforcing. NEAR’s vertically integrated architecture, financial settlement, intent matching, private execution, and AI inference running together, is what makes the NEAR token function as AI money rather than a fee skimmed from the system on top of it. The token isn’t incidental; it’s the economic unit the system runs on.
The window for establishing this position in the market is now, while the category itself is forming and before any single standard has been declared. The infrastructure is live, the revenue mechanics are compounding, and the agent economy is beginning to route real activity. The companies that become category defaults rarely win on product alone. They win by being the most credible, most integrated option at the moment the market decides it needs one. That moment is approaching.
The seven-layer stack
NEAR can be broken down into seven distinct layers that depend on and reinforce each other. The compounding logic is the point. Each layer makes the layers above it more valuable and the layers below it more defensible. A competitor can replicate one layer. Replicating the integration across all seven, with live revenue already flowing through the stack, is a different problem entirely.
Consistency of Vision
Individual layers are infrastructure. What they unlock in combination is the product. This is the distinction that matters for evaluating NEAR’s competitive position. Post-quantum cryptography combined with Intents produces something no other L1 currently offers: a single NEAR identity, PQ-secure, holding and transacting assets across 35+ chains. The account model NEAR designed in 2018 was built specifically to make upgrades like this trackable. The security upgrade arrives without a migration burden, which is precisely the kind of architectural foresight that only compounds in value over time.
NEAR AI Cloud combined with Intents and IronClaw produces a different kind of guarantee, one that matters enormously as agent-executed transactions scale. A business can hire an agent from the market, hand it private data and execution rights, and have a cryptographic guarantee that the agent cannot act maliciously with either. The operator can’t see your data. You can verify what ran, on what data, on what model, on what hardware.”
Agent Market combined with Intents and IronClaw produces what Illia described simply as “Upwork with full audit logs,” a labor market where agents are hired, paid, and held accountable through built-in payment rails and verifiable execution records. And SHIELD, running across the full stack with AI monitoring and formal verification, means security properties are proven mathematically against what users actually expect the code to do, not arbitrary specs written after the fact. Across 30+ chains, catching exploits in flight before they settle.
Any one of these is a nice piece of technology. None of them changes much in isolation. It’s the combination that does the work, each one only becomes powerful because the others are already running underneath it. That’s where the real value is: not in the layers, but in the way they compound.
NEAR Intents: Cross-Chain Revenue Engine Today, Agent Rails Tomorrow
near.com: The Flagship Front-end
near.com occupies a strategically distinct position within the Intents ecosystem, as NEAR can shape its direction. The front-end had $209+ million in volume over the last 30 days (based on June 11, 2026), representing roughly 10% of all Intents flow. near.com is the only surface offering users confidential mode, and adoption is already the default behavior: 42% of its swap volume now runs confidentially, with confidential TVL past $30 million as of late June 2026.
As of June 24, NEAR-operated front-ends in aggregate now account for 44.9% of total Intents flow, a significant step above the 15% channel-mix assumption embedded in SVRN’s March valuation model, meaning the capture rate upside case is already arriving faster than the base case projected. For up-to-date metrics, visit revenue.near.org.
Intents Ecosystem Map
NEAR Intents operates as a single liquidity layer accessible from anywhere in crypto. Developers looking to integrate into their product can get started at docs.near-intents.org.
Intents Growth strategy
The Intents growth strategy operates across six compounding workstreams, each adding volume, fee surfaces, or both.
- B2B integrations: Expanding the distribution network through a deep Q2 pipeline of wallet providers, aggregators, and DeFi infrastructure. Every new integration is another channel routing volume through the same execution and capture layer.
- More assets, more markets: Systematically broadening the asset universe from tokens to tokenized RWA’s, yield instruments, and derivatives. Every new instrument expands the addressable volume without changing the underlying infrastructure.
- Institutional order flow: Routing Intents directly into institutional execution stacks, and compliance tooling being built to meet the eligibility requirements regulated entities need to participate at scale.
- Stablecoin Transport Protocol (STP). Building the rails for stablecoin movement across chains, starting with cross-chain USDT today and expanding to the full universe of regulated, GENIUS-compliant stablecoins in Q3, positioning Intents as the default settlement layer for compliant stablecoin flow as that market scales.
- New products. Building net-new fee surfaces on top of existing volume, like earn and yield, confidential derivatives, fiat on/off ramps, and AI-assisted trading, each one expanding what users can do within the same execution layer.
- Sharper commercials. Improving take-rate mechanics on existing flow through tighter quoting and solver margin discipline, a capture rate lever that compounds independently of volume growth.
Intents / near.com roadmap
Q2 delivered the foundational layer including the full confidential loop, the first tokenized RWAs, perpetuals via Hyperliquid, and the [email protected] incentive program. Q3 brings the institutional and partner surfaces: tokenized US equities and treasuries, B2B Earn, B2B Confidential Intents, and prediction markets. Q4 is where the AI layer arrives on top of the financial rails: agentic functionality embedded on near.com, AI portfolio management, and continued expansion of yield tools, RWAs, and custom assets. What began as cross-chain crypto infrastructure is becoming the money rail for AI finance. The product surface stays the same. The market it addresses does not.
NEAR AI: The Next Wave of Value Capture
Why agents need NEAR
Agents are emerging as economic actors, and economic actors have infrastructure requirements. They need identity that persists across sessions and carries reputation. They need capital they can deploy across any chain without bridging friction. They need execution that is both reliable and private, so that strategy isn’t exposed in the act of transacting. They need to be discoverable, composable, and monetizable. And the users who deploy them need to retain sovereignty over the intelligence those agents generate. NEAR’s is building the infrastructure stack to provide all of this across every chain while being secure, private, and reliable.
Chain Signatures give agents cross-chain reach without bridges. The NEAR Protocol’s dynamic resharding scales settlement with agent volume and survives the quantum transition. IronClaw provides reliable, rule-bound execution guardrails so agents act autonomously without going rogue. Confidential Intents combined with TEE inference means agents transact without leaking strategy to the market. The native account model gives agents persistent identity that carries reputation across sessions. Agent Market makes agents findable, composable, and monetizable.
What NEAR is building extends well beyond the agent payments narrative that has dominated recent industry conversations. Replacing a payment credential with a crypto wallet address is a feature-level improvement, useful, but not structurally significant. The more consequential opportunity is replacing the entire commercial flow that surrounds a transaction: partner discovery, contracting, escrow, billing, and dispute resolution. That is the scope of what NEAR’s architecture is built to handle, and it represents a meaningfully larger addressable market than the payments framing implies.
Two big bets
NEAR AI is making two distinct bets that share the same underlying infrastructure but are at different stages of market readiness. Agent Market is an early-market bet, first SMBs are using it now for specific verticals, but the strategic logic is about surface ownership. The comparison to selling books online in 1995 is apt: the near-term use cases are narrow, but the team building the marketplace when agent-to-agent commerce is nascent is the same team that owns the substrate when it matures at scale. IronClaw is a trust bet, and that framing is deliberate. The primary constraint on agentic adoption isn’t capability, it’s user willingness to hand an agent real money and real authority. The near-term work is security and visible reliability; the mid-term goal is a killer app that becomes the default operating environment for agentic work; the long-term vision is that IronClaw is the brain running across every surface a person owns. Feature velocity and engagement metrics are the wrong scorecard for this stage. The only metric that compounds into the long-term position is trust.
Agent Market 2.0
NEAR’s revamped Agent Market is launching early Q3 as a directory of independent agents hired by the job. In plain language, pay in NEAR, USDC, or USD, get the result back. The user writes the prompt. Agent Market handles everything in between: billing, escrow, identity verification, metering, payouts, refunds, dispute resolution, and discovery. The commercial infrastructure that would otherwise require assembling multiple vendors is bundled into a single surface.
Two structural advantages separate this from competing approaches. First, the scope: Agent Market is targeting full commercial flow replacement, the kind of work you would otherwise hire a person, contractor, or firm to do, with discovery, contracting, escrow, and dispute resolution handled end-to-end. Second, the integration model: a company connects its data to Agent Market once, and every agent it subsequently hires inherits permissioned access to that data automatically. Agent builders get instant distribution without per-customer integration work, and enterprises get a single connection point rather than a fragmented vendor landscape. V1 launched at NEARCON to validate the core concept; the learnings from that cohort were folded entirely into the V2 rebuild launching this quarter.
NEAR AI roadmap
Q2 established the foundation, IronClaw Reborn, private USDC on Agent Market, automatic PII anonymization, and IronHub, collectively hardening the trust and privacy layer that everything above it depends on. Q3 brings the collaboration and coordination surface: a revamped Agent Market, multi-tenant cross-agent collaboration, self-learning loops, and permission management, the infrastructure for agents to work together reliably at scale. Q4 is where the system becomes truly autonomous: advanced memory and context, autonomous work execution, and a proactive assistant that acts without being prompted.
Case Studies
Case study: Wholesale Gorilla
Wholesale Gorilla is a Shopify app serving approximately 4,500 merchants, and its wholesale application vetting process illustrates precisely the kind of problem NEAR Agent Market was built to replace. Every application required 15 to 45 minutes of manual review, verifying the business, checking website and social presence, screening for fraud and territory conflicts. At a small brand, that work falls on the founder. At scale, it becomes a bottleneck that limits how fast the merchant network can grow.
The solution wasn’t integrating a model vendor, building prompts, standing up a headless browser, or wiring together billing and monitoring infrastructure. Wholesale Gorilla posted a job. One SDK call funds $1 USDC in escrow; an agent reviews the applicant and gets paid on acceptance. No model lock-in, no infrastructure overhead, no engineering sprint. The entire build was replaced by a marketplace call.
NEAR + Venice: a full DeFi + AI privacy loop
Venice routes private inference through NEAR AI, selecting end-to-end encryption in Venice means hitting NEAR’s confidential infrastructure. Meanwhile the VVV token (used to pay for inference) is tradable through NEAR Intents.
NEAR + Abound: cross-border payments on autopilot
Abound integrated IronClaw directly into their app as a financial concierge. The flagship use case: “Send $1,000 to India when USD/INR crosses ₹86”, the agent watches FX and acts, valuable given rupee volatility.
The Real TAM of NEAR
NEAR’s addressable market is widening from the crypto-to-crypto swaps a person initiates today over NEAR Intents to agents transacting on users’ behalf, and all the new value channels and markets they create.
Conclusion
NEAR is entering the second half of 2026 with thesis and traction converging for the first time. Protocol fee capture has scaled materially over recent months, near.com captures 100% of its fees and routes them to buybacks, and the protocol is measurably approaching the threshold where buybacks outpace issuance. These mechanics are live and verifiable in real time at revenue.near.org. The integrated stack NEAR has built operates as a single system, and the agent economy it was designed to serve is in its earliest stages of routing activity through it. The same settlement and coordination layer generating revenue today is the one positioned to address agent-native commerce as that market develops toward the end of the decade.
Standards in any category tend to consolidate early and hold for a long time. NEAR built its integrated stack before the category had a name and wired value capture to the specific capabilities agentic commerce requires. Translating that positioning into category-default status is now NEAR’s focus. The infrastructure is in place, the revenue mechanics are compounding, and the window to establish the standard is open now.
