Near is a 2021-era L1 that pivoted twice and failed both times. The L1 scalability story (2021) lost because developers didn’t migrate from the lack of liquidity, and the Blockchain Operating System pivot (2023) lost because decentralized frontends solved a problem nobody felt acutely enough to switch for. NEAR peaked above $20 in early 2022, flatlined through most of 2024-2025, and is down 93% from all-time-highs today, trading at $1.3 ($1.68B market cap and FDV). The asset is priced as if it's a ghost L1 that nobody uses. Except that this is no longer the case. The most recent roadmap prioritizes chain and wallet abstraction, together with intents. This positions NEAR as one of the potential market leaders for onchain AI agentic infrastructure while still playing a positive-sum game versus other L1s thanks to its crosschain capabilities.

The current thesis is that AI agent transactions will become the largest new source of crosschain routing volume. Agents are not constrained by UI/UX in crypto execution. Instead, they need infrastructure that takes an instruction (intent), and produces the completed transaction (abstracting complexity away). Near has already quietly built (and now officially shipped) a complete execution infrastructure for AI agents in crypto. The thesis bets on the probability that if the agent economy develops, even with a modest trajectory, NEAR is mispriced as an L1 (that is inactive), rather than crosschain infrastructure for agents.
The x402 protocol serves as a good gauge for crypto-native AI agent financial activity. It recently averaged ~$2.5M daily volume in December (~$900M annualized). This amount declined into 2026 with broader crypto market. On top of that, the ATXP protocol (Stripe crypto alum founders and backed by Stripe) is breaking into the agentic commerce scene, raising $19.2M in Sep 2025. Keep in mind, Stripe processed $1.9T in total payment volume in 2025.
The agents deploying today fall into various categories like trading, treasury management, DeFi strategy/yield farming agents. The common thread across all of these is that any agent touching more than one chain immediately runs into the same infrastructure problem. Each chain has its own gas token, its own transaction format, potentially different wallets, and its own bridge connections to other chains.
Without a unified execution layer, every agent developer builds custom integrations for each network independently, which means the agent's compute is spent solving infrastructure problems rather than executing its strategy. CEX infrastructure doesn't solve this either, because exchanges carry freeze risk and require KYC on API provisioning, both of which are risks to the 24/7 execution model agents require.
The demand for a single layer that accepts an outcome and handles the routing underneath is an operational bottleneck that every multichain agent developer is already working around with custom code, and the volume of agents that need this will scale with the broader AI deployment wave that is already underway.
NEAR Intents is the only live infrastructure built for how agents actually execute. Agents submit outcomes and solvers compete to fill them, handling routing, gas, and settlement underneath. The agent simply specifies what it wants (the intent) and gets the outcome.
Chain Abstraction gives the agent a single identity across 35+ chains, meaning the developer doesn't write custom integrations for each network. IronClaw, launched at NEARCON 2026, provides a hardware-secured runtime on NVIDIA Confidential Computing where credentials stay encrypted and secrets never reach the model, addressing the trust barrier that has hindered enterprises from deploying agents to manage money.

Since launch, $14B have already flowed through the Intents solver network, with $2.4B of that occurring in the last 30 days across 556,000 unique addresses, concentrating 16% of all-time volume in the most recent month. Of that, $25M was generated as fees, implying an average take-rate of roughly 0.18%. Ledger integrated Near Intents via SwapKit as its crosschain swap backend, and Rango and CoW Protocol route through the solver network as referrals.
SwapKit accounts for 77% of total intent fees with most of that volume originating from Ledger users who likely don’t realize Near’s underlying infrastructure. This is another part of the infrastructure thesis, where Near is embedded in third party apps and grows through distribution it doesn't own or control. Each new wallet or agent framework that integrates Intents brings an entire user base into the solver network without any of them seeing the Near infrastructure underneath.

Crosschain routing and bridging currently processes roughly $50B per month at peak volumes, making it one of the largest fee-generating layers in DeFi. This is the addressable market if NEAR Intents becomes the canonical routing option for both agents and the wallets embedding it. At the current run rate of $2.4B monthly, NEAR captures ~4% of that existing market. Each new wallet integration that embeds Intents brings an entire user base into the solver network without those users or agents knowing NEAR exists.

This is a positive sum system as Near is not competing with other L1s, since solvers are routing to all of them. Every intent sends value to Solana, Arbitrum, Base, or wherever the instruction points. Every chain in the network benefits from NEAR being the routing layer because NEAR's growth is their growth too. There is no other L1 that occupies this position, and the low TVL that makes NEAR look dead on an L1 metric is the product working as designed, because NEAR's job is to route value out, not to park it.
Additionally, NEAR is fully unlocked. The 4-5 year vesting schedule across all allocations (backers, core contributors, community grants, foundation endowment) is completed with no cliff events ahead or scheduled supply shocks. The structural overhang that kills narratives for most 2021 L1 tokens wouldn’t apply here.
The remaining supply-side dynamics are token inflation and the fee switch. Maximum inflation was halved to 2.5% annually, with 70% of transaction fees burned onchain. At current rates, annual inflation produces roughly 32M new NEAR (~$41M at current prices).

Ultimately, the thesis rests on Near growing as an embedded piece of infrastructure rather than routing flows from the near.com frontend user-facing app. Partner share should remain above 75% share of total crosschain volumes otherwise this is an invalidation. The market prices a ghost L1 while the underlying product is a positive-sum routing infrastructure for an agent economy that is already deploying capital, has no other execution layer at this scale, and sends value to every chain it touches. If that framing is confirmed with agentic transaction volumes, the current price is paying 2021 ghost-chain multiples for a protocol who is capturing a trillion-dollar agentic commerce market.


