NEAR Protocol Review 2026: Intents, AI, Tokenomics and Risks

NEAR Protocol has evolved far beyond its original role as a scalable Layer 1 blockchain. Today, its strategy centers on chain abstraction, cross-chain execution and AI-agent infrastructure, with NEAR Intents emerging as its strongest source of real-world adoption.

This NEAR Protocol review looks at how the network works, where Intents and Chain Signatures fit in, how NEAR's tokenomics are changing, and whether its growing infrastructure stack is translating into meaningful value for the NEAR token.

Editor's Note (Sept. 4, 2026): We fully updated this NEAR Protocol review in September 2026 to reflect the network’s latest developments, including NEAR Intents, Chain Signatures, Nightshade sharding, AI infrastructure, staking, governance, tokenomics, ecosystem activity and security risks. We also refreshed key figures on Intents volume, DeFi TVL, validator activity, issuance and token buybacks.

NEAR Protocol At A Glance

Data is current as of Sept. 4, 2026.

What Is NEAR Protocol?

NEAR Protocol is a sharded proof-of-stake Layer 1 blockchain that positions itself as infrastructure for cross-chain transactions and AI agents through products such as NEAR Intents and Chain Signatures.

Alexander Skidanov and Illia Polosukhin founded NEAR after initially working together on an AI-focused project. Polosukhin previously worked on machine learning at Google and was one of the authors of the Transformer research paper Attention Is All You Need. Skidanov brought distributed-systems experience from MemSQL, now SingleStore.

NEAR began its staged mainnet rollout in 2020, with its Proof of Authority mainnet launching on April 22, 2020, before the network transitioned toward community-operated proof of stake. The NEAR Foundation, a Swiss nonprofit, became one of the main organizations supporting ecosystem development, grants and coordination.

The original idea was fairly straightforward. NEAR wanted to make blockchain applications easier to build and use without running into the scaling limits that emerge when one network has to process every transaction through the same execution path. Human-readable accounts, low transaction costs and Nightshade sharding all came out of that design philosophy.

Its strategy widened as crypto itself became more fragmented. Capital, users and applications spread across Ethereum, Solana, Bitcoin, Layer 2 networks and dozens of other ecosystems.

That is why NEAR is harder to place in the usual Layer 1 comparison today. It still competes for applications and users on its own chain, but an increasing share of its infrastructure is designed for people who may never think of themselves as NEAR users.

What Is NEAR Trying to Solve?

NEAR's current architecture revolves around three problems that feed into one another: blockchain scalability, poor Web3 user experience and fragmentation between networks. AI-agent infrastructure is now developing on top of the same foundation.

Scalability was the original challenge. If every validator processes the same workload, more users eventually bring congestion or force the network toward heavier hardware. NEAR's sharding architecture divides network state and execution so multiple shards can process activity in parallel.

Scaling alone does not solve the usability problem. Crypto still asks users to manage wallet addresses, gas tokens, approvals, network switching and repetitive signatures before completing relatively ordinary actions. NEAR's account system and access keys were designed to reduce some of that friction at the account level.

Fragmentation became harder to ignore as the industry spread across more chains. Traditional cross-chain bridges move assets between ecosystems, but the user still needs to understand the source network, destination network, bridge support, gas requirements and liquidity route. NEAR Intents approaches the same problem from the opposite direction: the user specifies the result, then solvers compete to figure out how to get there.

AI agents fit naturally into that structure. An agent that needs to pay for a service, move assets or interact with several applications faces the same cross-chain mess as a human, while also needing tighter controls over credentials, privacy and permissions. NEAR increasingly sees the infrastructure it built for people as infrastructure agents can use too.

How Does NEAR Protocol Work?

NEAR combines sharded execution, proof-of-stake consensus and a flexible account model. The three pieces are closely related. Nightshade spreads workload across the network, validators secure that execution, and the account system gives users and applications more precise control over how transactions are authorized.

Nightshade Sharding

Nightshade is NEAR's sharding architecture. Instead of requiring every validator to execute and maintain the complete state of the network, NEAR splits state across multiple shards that process workloads in parallel while remaining part of the same blockchain.

The production network currently operates with nine shards, roughly 600-millisecond blocks and around 1.2-second finality.

One of the biggest changes from older versions of NEAR is stateless validation. Validators can verify shard execution using state witnesses without every validator maintaining the entire state for every shard. As the blockchain adds shards, this reduces the storage burden that would otherwise grow alongside network capacity.

Dynamic resharding takes the same idea a step further. NEAR can split or merge shards without shutting the network down, so the network does not remain tied to a shard configuration selected years earlier. If activity becomes concentrated in one part of state, the architecture can adjust how that workload is divided.

The technical benefit is horizontal scalability. NEAR can add aggregate capacity through more shards instead of trying to push every new transaction through one increasingly crowded execution lane. Whether it needs all that capacity is ultimately an adoption question rather than an engineering one.

Proof of Stake and Validators

NEAR uses proof of stake to secure the network. Validators commit NEAR to the protocol and participate in producing blocks, producing chunks or validating chunks. Holders who do not want to run validator infrastructure can delegate their tokens to staking pools.

Stake plays a role in determining who participates. Larger validators take on broader block and chunk-production responsibilities, while smaller participants can contribute primarily through chunk validation.

The active validator set currently sits around 420. As of Sept. 4, 2026, a NearBlocks snapshot showed roughly 610.9 million NEAR staked, a seat price around 9,800 NEAR and protocol version 86. These figures change from epoch to epoch, so they are better treated as a live snapshot than as permanent network constants.

The validator count has grown materially from earlier versions of NEAR, but the number itself only tells part of the decentralization story. Stake concentration, infrastructure providers, validator roles and software dependence all influence how much control is actually distributed. We return to that when assessing the validator set in more detail later in the article.

NEAR's Account Model and Access Keys

NEAR's account model remains one of its more useful design choices because it changes how users and applications handle permissions.

Instead of representing every account primarily through a long hexadecimal string, NEAR supports human-readable accounts, implicit cryptographic accounts and subaccounts. A user might interact with an address such as alice.near, while applications can create structured subaccounts beneath an existing namespace.

The larger difference appears in the keys attached to those accounts.

A Full Access Key can transfer NEAR, add or remove keys, deploy contracts and perform other privileged actions. A Function Call Key can be limited to particular contracts, methods and gas allowances, and it cannot make ordinary NEAR transfers.

That gives developers room to grant an application exactly the permissions it needs. A game, for instance, could receive permission to execute a narrow set of gameplay actions without receiving authority over the player's full account balance.

The security logic is similar to the principle behind protecting private keys: the less authority an exposed credential has, the smaller the damage if something goes wrong.

Function Call Keys are not automatically safe simply because they are scoped. A badly configured permission can still authorize more than intended, while losing a Full Access Key carries the same fundamental ownership risk as losing a high-privilege key on another blockchain. The advantage is that NEAR gives developers finer tools for separating those permissions.

Smart Contracts and Developer Experience

NEAR supports smart contracts written primarily in Rust and JavaScript, which compile into WebAssembly for execution.

Rust remains useful for developers who want tighter control and mature blockchain tooling. JavaScript lowers the barrier for people coming from conventional web development. WebAssembly provides the common execution layer underneath both.

Contracts live within NEAR's account system and can be upgraded when account permissions allow it. Developers can combine that with low transaction costs, Function Call Keys and meta-transaction patterns to reduce signature friction for users.

This is where NEAR's account design becomes more than a naming convenience. Human-readable accounts, scoped keys and programmable permissions were built into the network long before account abstraction became a larger industry theme. The combination gives developers more room to make a DApp behave like a familiar application without removing the user's ability to control the underlying account.

NEAR Intents and Chain Abstraction

NEAR Intents is now NEAR's strongest product-level differentiator because it turns chain abstraction into an execution market that people are already using at scale.

What Is Chain Abstraction?

Chain abstraction means letting users interact with crypto based on what they want to accomplish rather than forcing them to understand every network, bridge, gas token and transaction route involved.

Suppose a user owns ZEC but wants USDC on another network. A conventional workflow could involve finding a bridge or exchange that supports both assets, selecting the destination chain, arranging gas and then completing another swap.

The actual request is simpler: turn this asset into that asset and deliver it here.

Chain abstraction pushes the route construction into the application and execution layer. The blockchains underneath do not disappear. Bitcoin still follows Bitcoin's consensus rules. Ethereum still executes transactions through Ethereum. Solana still has its own runtime, validators and fee market.

The difference is that the user no longer needs to manage every one of those systems directly. NEAR's approach is to let applications, solvers and signing infrastructure handle more of the chain-specific work behind the transaction.

How NEAR Intents Work

NEAR Intents turns a desired outcome into a market where competing solvers try to fulfill it.

User Intent → Solvers → Best Execution → Verification → Settlement

1. The user states the desired outcome. Instead of prescribing a bridge and swap route, the user specifies what asset they want to receive and where.
2. Solvers compete to fulfill the intent. They evaluate liquidity and submit quotes based on the user's constraints.
3. A suitable route is selected. Execution can be optimized around price, speed and available liquidity.
4. The result is verified and settled. Settlement logic checks the agreed conditions and completes the transaction.

This differs from conventional AMM routing. An automated market maker calculates trades against liquidity sitting in pools. An intent system creates competition around the requested outcome itself. A solver can source liquidity from different venues and construct an execution path without exposing every leg of that path to the user.

That architecture only becomes interesting if people actually use it. As of Sept. 4, 2026, NEAR Intents reports more than $27 billion in all-time volume across 34 chains. DefiLlama tracks about $27.14 billion in cumulative DEX volume and $2.79 billion over the previous 30 days.

The fee structure also explains why high volume and token value capture are not the same thing. The base NEAR Intents protocol fee is 0.0001% per transaction. Swaps executed through near-intents.org add a 0.2% distribution fee, and unauthenticated 1Click API requests also carry an additional 0.2%. Authenticated integrations can operate with the underlying protocol fee plus market-maker spreads and any application-level fee.

A great deal of economic activity can therefore pass through Intents without all of those fees becoming revenue belonging to NEAR. That becomes central when we get to token value capture.

What Are Chain Signatures?

Chain Signatures gives NEAR accounts and smart contracts a way to authorize transactions on external blockchains.

The system uses multi-party computation, or MPC. Instead of storing one complete external-chain private key with one signer, a distributed group participates in generating the required signature.

NEAR's documentation says Chain Signatures supports signature schemes that allow NEAR accounts to sign transactions for networks including Bitcoin, Ethereum and Solana.

A multichain application can therefore use logic associated with a NEAR account to request signatures for external accounts rather than building an entirely separate account system for every chain it touches.

That is where Chain Signatures connects with Intents.

Intents determines what outcome should happen and which solver can execute it. Chain Signatures provides a way for the NEAR account or contract to authorize actions that need to happen elsewhere.

Together, they let applications coordinate activity across chains without requiring the user to manually sign into a different wallet environment for every step.

This can also reduce dependence on wrapped assets in some workflows. Instead of manufacturing a representation of a foreign asset on NEAR, an application may interact directly with the native asset on its original chain.

Are NEAR Intents Better Than Crypto Bridges?

NEAR Intents can offer a cleaner experience than a conventional bridge, particularly when the user's actual goal is a swap rather than simply moving the same token from one network to another.

The UX improvement is fairly obvious. Intents can hide manual bridging, gas management and route selection.

The underlying risk does not vanish with the interface. Smart contracts still verify settlement, solvers still need to honor quotes, liquidity still needs to be available and the external chains involved still need to function.

So Intents is better understood as a different way of organizing cross-chain execution. It can remove a large amount of work from the user's side by pushing that complexity into an execution and settlement network. Whether that is preferable to a bridge depends on the route, assets and trust assumptions involved.

NEAR's AI Strategy

NEAR has a credible technical connection to AI because many of the problems it is already solving for cross-chain users also appear when autonomous agents start making payments and moving assets. The infrastructure exists. The unanswered part is whether enough agents will use it to create meaningful economic demand.

Why NEAR Is Building for AI Agents

AI agents become much more complicated once they are allowed to act rather than simply generate text.

An agent may need:

• Payments to buy services, settle invoices or compensate another agent.
• Wallets to hold and move digital assets.
• Credentials to access APIs, accounts and private resources.
• Cross-chain execution to reach liquidity and applications spread across several networks.
• Private computation to protect user data, prompts and authentication secrets.
• Verifiable access to prove which code or model performed an action.

NEAR already has infrastructure aimed at several of these problems. Intents can translate an agent's desired outcome into an executable transaction. Chain Signatures can authorize activity on external networks. Function Call Keys can limit what the agent is permitted to do. Confidential computing can keep sensitive prompts and credentials away from the ordinary host environment.

The bet is therefore less about running an entire AI model on a blockchain and more about giving agents a transaction, permission and privacy layer they can operate through.

That makes the AI strategy easier to understand. NEAR is taking infrastructure built to hide blockchain complexity from humans and asking whether autonomous software will need the same abstraction even more.

NEAR AI and IronClaw

NEAR AI is the project's AI infrastructure effort, while IronClaw is an open-source agent runtime designed to execute autonomous workflows with tighter permission boundaries.

IronClaw 1.0 launched on July 27, 2026 with persistent state, centralized permission checkpoints and deployment through command-line, web, Slack and Telegram interfaces.

Its architecture separates decision-making from execution. Sensitive actions can pass through a guard layer, while credentials can be issued with narrower permissions and prevented from leaking through ordinary logs or agent output.

The release is already deployed internally across the NEAR Foundation and NEAR AI. That shows the product is usable in production, although internal deployment is still a much weaker signal than broad outside adoption.

External integrations are beginning to fill in that gap. Corbits, for example, introduced NEAR AI private inference for enterprise agent workflows in July 2026.

The AI stack is therefore past the point where it can be dismissed as a roadmap story. IronClaw, NEAR AI Cloud and confidential inference are live. Distribution and enterprise usage are developing. The unanswered question is whether these products eventually support a large economy of autonomous agents rather than a smaller market for confidential AI infrastructure.

Private Inference and Confidential Computing

NEAR AI uses Trusted Execution Environments, or TEEs, to run supported workloads inside hardware-isolated environments.

A TEE protects data while it is being processed, not only while it sits in storage or moves across a network. NEAR AI Cloud uses confidential computing so prompts can be decrypted and processed within isolated hardware rather than the ordinary host operating system.

Its current architecture describes prompts being decrypted inside Intel TDX and NVIDIA confidential-computing environments.

That is useful for agents because credentials, wallet information, proprietary data and private prompts are exactly the kinds of things users do not want exposed to an ordinary cloud operator.

Private Shard extends the same idea toward blockchain execution. It creates an environment where certain transactions can remain confidential while still interacting with NEAR's public network.

That privacy comes with a different trust model.

Public blockchain execution is built around transparent state transitions that independent validators can verify. TEE confidentiality depends partly on hardware isolation, firmware, attestation systems and the manufacturers behind them. A hardware vulnerability can therefore compromise privacy even when NEAR's proof-of-stake consensus continues operating normally.

The model also differs across AI services. NEAR AI identifies some closed-source partner models as Incognito models, where requests are processed outside NEAR AI's GPU TEE and do not receive hardware-level privacy or attestation.

Is NEAR's AI Adoption Real Yet?

There is enough evidence to say the infrastructure is real. There is not enough evidence to say the autonomous-agent economy has arrived.

IronClaw 1.0 is live. NEAR AI Cloud offers confidential inference for supported models. Hardware-backed attestation is available for TEE workloads. Private inference has outside integrations. Private Shard brings confidential execution into the blockchain stack. Intents and Chain Signatures already provide payment and cross-chain transaction infrastructure that agents can use.

The missing piece is scale.

NEAR has not yet shown a broad population of autonomous agents generating sustained transactions, fees and payments large enough to materially change network economics. It is also possible for customers to use parts of the NEAR AI stack without creating much direct activity on NEAR Protocol.

That puts AI in a different category from Intents. Intents already has substantial economic activity behind it. AI has working infrastructure and early distribution, but its contribution to network demand still needs to be demonstrated.

NEAR Tokenomics and How the NEAR Token Captures Value

NEAR's token economics improved in 2026 because issuance fell and Intents began funding NEAR purchases. The mechanisms are better aligned with token holders than they were before, but their current scale remains too small to make NEAR structurally deflationary.

What Is the NEAR Token Used For?

NEAR sits at the center of network execution and security.

Its main uses include:

• Gas and transaction fees on the Layer 1.
• Staking through validators and staking pools.
• Economic security for proof-of-stake consensus.
• Storage requirements for accounts and contracts.
• Payments throughout the ecosystem.
• Participation in House of Stake governance.

The governance role has become more concrete as House of Stake has taken responsibility for parts of NEAR's economic policy. Token holders still do not directly control every protocol upgrade, Foundation decision or ecosystem product.

NEAR Supply, Inflation and Token Burns

NEAR has approximately 1.305 billion tokens in both circulating and total supply. The network began with 1 billion NEAR at genesis, while validator issuance expanded supply over time. NEAR has no fixed maximum supply.

Most of the early token allocations are now substantially vested, so the current supply equation depends much more on issuance and burns than on old unlock schedules.

The larger 2026 change is the issuance rate.

NEAR now targets annual validator rewards equivalent to 2.5% of total supply, down from the previous 5% model.

At roughly 1.305 billion tokens, that works out to around 32.6 million NEAR in gross annual issuance before transaction-fee burns.

Transaction fees push supply in the opposite direction. Under the currently active protocol version, most fees are burned, while part of eligible smart-contract execution fees can still be returned to contracts.

That final part is on its way out.

House of Stake ratified HSP-027 on June 15, 2026, approving removal of NEAR's 30% developer gas rebate so the eligible execution fee can instead be burned.

The policy has been approved, but the protocol change still needs to reach mainnet. A NearBlocks snapshot showed mainnet running protocol version 86, while nearcore's unreleased changelog places the removal of gas rewards in an upcoming version. The code change removes the 30% contract reward and sets it to zero.

That timing affects the current supply model. The governance decision is settled; the stronger burn mechanism is not yet part of the live economics at this research cut.

NEAR Intents Fees and Token Buybacks

NEAR Intents gives the token another route to value capture outside ordinary Layer 1 gas.

As of Sept. 4, 2026, DefiLlama tracked roughly $44.5 million in cumulative Intents fees, $5.28 million in captured revenue, $3.51 million in holder revenue and $27.14 billion in cumulative swap volume.

Those figures diverge because solvers, distribution channels and other participants take part of the economics. High trading volume does not automatically become high NEAR revenue.

Since Feb. 23, 2026, captured Intents revenue has been used to buy NEAR on the open market.

This is where Intents changes the token thesis. A swap can happen between assets on external networks and still generate revenue that eventually creates demand for NEAR. That connection did not previously exist in the same form.

It also solves only part of the value-capture problem.

The mechanism works, but its scale remains small. Around $3.5 million in cumulative buybacks proves that Intents usage can feed into NEAR demand. It does not yet come close to offsetting annual token issuance.

The relevant question is therefore no longer whether Intents can create value for NEAR. It can. The question is how quickly the amount captured grows relative to the amount of new NEAR entering circulation.

Could NEAR Become Deflationary?

NEAR can become deflationary only if token sinks eventually exceed new issuance.

The basic equation is:

Net Supply Change = New Issuance − Token Burns − Permanent Supply Reductions

At roughly 1.305 billion tokens, 2.5% annual issuance produces around 32.6 million NEAR of gross yearly issuance.

Base-layer transaction burns reduce that amount. Intents buybacks add market demand, although bought-back tokens should not automatically be counted as burned unless they are permanently removed from supply.

For sustained net deflation, fee burns and any other permanent supply reductions would need to outweigh issuance.

That becomes more plausible if several things happen together:

1. Higher base-layer activity produces more transaction-fee burns.
2. Removal of the developer rebate directs more eligible fees toward burning.
3. Intents volume and retained revenue continue growing.
4. More protocol revenue begins creating NEAR demand.
5. Issuance stays near or below the current 2.5% target.

Those conditions are possible. They are not present at the required scale today.

NEAR's token design is therefore moving in a more favorable direction for holders, but the network still needs substantially more economic activity before deflation becomes a realistic base case.

NEAR Staking, Validators and Governance

NEAR holders can currently earn around 5% through native staking, while House of Stake has made token holders more relevant to economic policy. The quality of that system depends on more than yield, though. Validator concentration, entry economics and the limits of governance authority all shape how control is distributed.

How NEAR Staking Works

NEAR holders can delegate tokens to validators or staking pools without running validation infrastructure themselves.

Validators operate the hardware and software required to participate in consensus. Delegators contribute stake and receive a share of rewards after validator commission.

As of Sept. 4, 2026, a NearBlocks snapshot recorded last-epoch network APY around 5.27%. The actual return varies by validator commission, performance and the percentage of total NEAR staked.

Roughly 47% of supply is currently committed to validators.

Native staking also has an unstaking delay. Users who undelegate usually wait four epochs before withdrawing, which NEAR documentation estimates at approximately 24 to 28 hours.

Liquid staking reduces the liquidity problem by giving users a transferable token representing the staked position. It also adds another smart-contract and protocol layer between the holder and native staking.

How Decentralized Is NEAR?

NEAR has a considerably broader validator set than it did a few years ago, although stake is still concentrated enough that the headline validator count can overstate how evenly power is distributed.

Current explorer data shows roughly 420 active validators and around 610 million to 620 million NEAR staked, depending on the epoch.

One recent snapshot placed the largest validator at roughly 6.5% of active stake, with the top eight collectively crossing one-third of staked NEAR.

That concentration does not mean eight validators can simply rewrite the network. Consensus security depends on thresholds, validator roles and protocol rules. It does show why validator count and stake distribution need to be read together.

Entry requirements create another layer. The current seat price is around 9,800 NEAR, although NEAR now distinguishes large block and chunk producers from smaller chunk validators. That gives smaller operators routes into the active set without requiring them to compete directly with the largest staking providers for every role.

Stateless validation helps here because validators no longer need to maintain every shard's entire state. At the same time, the network still depends on a mix of stake distribution, validator infrastructure and software implementations.

House of Stake and NEAR Governance

House of Stake has become NEAR's main token-holder governance system for economic policy.

Its 2026 mandate gives it responsibility for economic parameters, value accrual, treasury allocation, emissions, fees and incentive design.

That means token holders can now influence decisions with direct economic consequences.

HSP-027 shows how the process works. House of Stake approved eliminating the developer fee rebate, but that vote did not instantly modify mainnet. Protocol contributors still need to write and test the code, and validators need to adopt the corresponding software version.

The governance structure therefore has several layers. House of Stake can approve economic policy. Protocol contributors implement changes. Validators activate those changes by running the relevant protocol version. The NEAR Foundation remains a separate organization supporting ecosystem functions rather than sitting underneath every token vote.

House of Stake's remit may expand further. An August 2026 mandate update proposed moving control of NEAR's top-level account namespace into House of Stake governance.

This gives NEAR holders a more tangible role than they had under older DAO experiments, while avoiding the fiction that one token vote controls every part of the network.

NEAR Ecosystem and Adoption

NEAR's adoption numbers tell two different stories. The Layer 1 has users and applications, but its native economy remains relatively small. Intents, meanwhile, is moving far more value across chains. That split may eventually say more about NEAR's future than its raw transaction capacity.

NEAR Network Activity

As of Sept. 4, 2026, NEAR had approximately:

• 77,130 active addresses over 24 hours
• 670,529 transactions over 24 hours
• Roughly $107 million in DeFi TVL
• Around $104 million in stablecoin market capitalization
• Roughly $19 million to $22 million in daily DEX volume, depending on the intraday snapshot
• About $1,600 in daily base-chain fees
• Around $1,120 in daily base-chain revenue

The user and transaction numbers show that the network is active. The fee numbers tell a different story about how economically dense that activity is.

NEAR is cheap to use, which is good for users but means transaction counts can grow without creating proportionate protocol revenue. A chain processing hundreds of thousands of daily transactions may still produce relatively little economic value for token holders if each transaction contributes very little in fees.

Stablecoins help fill in that picture because they show how much dollar-like transactional capital is sitting inside the ecosystem. With roughly $104 million in stablecoin supply, NEAR has usable liquidity but remains far behind the deepest smart-contract ecosystems.

This is why theoretical TPS figures add little to the adoption discussion. NEAR has demonstrated that it can process much more activity. The real question is how much valuable activity users actually bring to the chain.

NEAR DeFi

Native NEAR DeFi remains modest compared with the largest Layer 1 ecosystems.

Network TVL sits near $107 million, while seven-day DEX volume is around $110 million. That is enough to support an operating ecosystem, but it is a long way from the depth available on Ethereum and Solana.

Rhea Finance is one of the main native liquidity hubs and carries functions inherited from earlier NEAR DeFi infrastructure. The ecosystem also includes wallets, staking applications, consumer products and payments rather than revolving entirely around lending and trading.

Readers looking for the application layer in detail can use Coin Bureau's Top NEAR Projects guide.

The liquidity question is harder to route around. DApps become more attractive when they can plug into large pools of neighboring capital, users and protocols. Thin native markets weaken that feedback loop even when the underlying blockchain works well.

That makes NEAR's move toward cross-chain infrastructure easier to understand. If it cannot outcompete the largest ecosystems on native liquidity alone, it can try to become useful to liquidity wherever that liquidity already sits.

Intents Growth vs Native NEAR Activity

This is where NEAR's 2026 strategy becomes clearest.

NEAR Intents processed around $2.79 billion in swaps over the previous 30 days and more than $27 billion cumulatively. Native NEAR DeFi held around $107 million in TVL, while the network generated roughly $110 million in DEX volume over seven days.

These metrics are not directly interchangeable. TVL measures capital sitting in protocols, while volume measures activity over time. Their relative scale still shows where NEAR's strongest momentum currently lies.

Our interpretation is that this is exactly what chain abstraction was supposed to achieve.

An Ethereum wallet can use infrastructure that ultimately routes through Intents. A multichain application can use Chain Signatures. An AI agent can interact with several networks without becoming a conventional NEAR DApp. NEAR can therefore become more useful without every user moving their assets onto NEAR itself.

Under that model, the project starts to resemble middleware or a universal execution layer rather than a conventional destination chain.

The alternative interpretation is that native ecosystem gravity remains too important to dismiss. Liquidity attracts traders. Traders attract applications. Applications attract more developers and capital. Thin native markets can limit those network effects even if cross-chain infrastructure succeeds elsewhere.

The next stage of NEAR's development will show which interpretation carries more weight. If cross-chain usage keeps growing and NEAR captures a meaningful share of its economics, weak native TVL becomes less damaging. If Intents grows while value capture remains thin, NEAR may end up owning useful infrastructure without extracting enough economic value from it.

NEAR Protocol Security and Risks

NEAR's security surface now stretches well beyond proof-of-stake consensus. The base blockchain, MPC signer network, Intents solvers, verifier contracts, private validators and TEEs can all fail in different ways, so treating them as one security system gives a false sense of simplicity.

NEAR Layer 1 Security

NEAR's base layer is secured through proof of stake.

Validators commit NEAR to the protocol and participate in producing and validating blocks and chunks. Attempts to violate consensus can put validator stake at risk through slashing.

Nightshade distributes execution across shards, while stateless validation lets validators verify shard execution from state witnesses rather than maintaining every shard's complete state.

Audits, public security work and bug-bounty programs sit around that core protocol. They reduce the chances that obvious vulnerabilities remain undiscovered, but they cannot guarantee that exploitable code never reaches production.

Users also face risks that have nothing to do with consensus. A compromised wallet or malicious approval can cause losses while the Layer 1 continues working exactly as intended.

Chain Signatures and MPC Risk

Chain Signatures introduces a different security model because external-chain assets can depend on NEAR's MPC signer infrastructure.

No single MPC participant is meant to hold the complete signing key. The signer network cooperates to generate a valid signature only after a request has been authorized.

That removes the obvious single-key failure point, but it creates a distributed one.

Enough signer participants colluding, a threshold-protocol flaw or a serious implementation bug could threaten the integrity of signatures. A less dramatic failure can affect availability instead: if enough signers are offline, valid transactions may simply stop being signed.

The software handling requests is part of the same security boundary. So are the operators maintaining individual MPC nodes.

NEAR governance has already been funding MPC-node infrastructure, which is a useful reminder that the signer network is not an abstract cryptographic layer with no operating costs. It is a live system that needs enough independent, functioning participants.

For external assets to be directly exposed, an attacker would generally need to break enough of the authorization, contract or threshold-signing assumptions to produce a signature that should never have existed.

NEAR Intents and Solver Risk

Solvers improve execution by competing with one another, but competition alone does not make the system secure.

A solver can fail to complete an execution, withdraw liquidity, quote poorly or experience an outage. Intents therefore depends on verification and settlement rules that enforce the outcome the user agreed to.

NEAR Intents uses smart-contract-enforced settlement, with failed intents able to revert and return assets to the designated refund address.

The main risks sit across several layers:

• Solver failure can prevent a quoted route from completing.
• Liquidity can disappear before execution.
• Verifier or settlement contracts can contain vulnerabilities.
• Tokens moving through the system can contain their own contract risks.
• External chains can experience congestion or failure.
• Prices can move while an intent is being filled.

Having several solvers makes the market more competitive and reduces dependence on one execution provider. The contracts underneath them still need to work correctly.

Private Shard and TEE Trade-Offs

Private Shard addresses a very different problem: how to keep certain transactions and computations confidential.

Its current design uses a permissioned validator set together with Trusted Execution Environments, allowing sensitive activity to stay out of the public mempool while still interacting with NEAR's public network.

That can be useful for trading strategies, institutional workflows and AI-agent credentials, where public execution would reveal information the user would rather keep private.

The price of that confidentiality is a narrower trust model.

Public NEAR mainnet allows validator participation according to the network's staking rules. Private Shard currently relies on permissioned validators. TEE security also depends on hardware isolation, firmware and attestation systems.

A flaw in a TEE could therefore compromise confidentiality without affecting NEAR consensus. The opposite can also happen: mainnet can continue operating normally while the private environment suffers a separate security problem.

Users choosing Private Shard are accepting those additional assumptions in exchange for privacy.

Ecosystem and Smart Contract Risk

A blockchain can execute a vulnerable application perfectly.

The SWEAT exploit in April 2026 shows why that distinction needs to stay clear. Fireblocks estimated that attackers stole approximately $3.5 million after exploiting missing access control in the SWEAT token contract's ft_resolve_transfer callback.

The attackers did not break Nightshade, compromise NEAR consensus or steal a validator key. They exploited application logic running on top of the network.

That is the basic shape of smart-contract risk. Validators are responsible for executing the code according to protocol rules. They are not responsible for deciding whether the business logic inside every contract is safe.

The same separation applies to DeFi protocols, staking products, wallets, tokens and other applications. NEAR's Layer 1 can be functioning correctly while users still lose money higher up the stack.

Is NEAR Protocol a Good Investment?

• Bull case: NEAR can grow by becoming widely used cross-chain infrastructure, with Intents adoption, improving token economics and AI-native tooling supporting that thesis.
• Bear case: NEAR still has a small native economy, modest value capture and growing competition, while its AI and cross-chain strategies remain unproven at scale.

The Bull Case for NEAR

• Intents already has real usage: More than $27 billion in cumulative swap volume and around $2.8 billion in 30-day volume show sustained adoption.
• Growth does not depend on users moving to NEAR: NEAR Intents can serve wallets and applications on other chains.
• Chain Signatures expand NEAR's reach: They let NEAR infrastructure authorize transactions across external networks.
• This changes the Layer 1 growth model: NEAR can grow as cross-chain infrastructure rather than relying only on native users, TVL and applications.
• Token economics are improving: Validator issuance has fallen to 2.5%, transaction fees are burned and the developer gas rebate is being removed.
• Intents can support token demand: Revenue from Intents is now used for open-market NEAR purchases.
• The scaling stack is live: Nine shards, stateless validation and dynamic resharding are already operating.
• AI could create another growth channel: Autonomous agents may need cross-chain signing, payments and privacy infrastructure, areas NEAR is actively targeting.
• NEAR does not need to beat Ethereum or Solana: The bull case works if NEAR becomes widely used cross-chain infrastructure and captures more value from that activity.

The Bear Case for NEAR

• Native DeFi remains small: Around $107 million in TVL is modest relative to NEAR's age, valuation and technical ambitions.
• Base-layer revenue is limited: Low transaction revenue means ordinary network activity currently contributes little to token burns.
• Cross-chain infrastructure is highly competitive: Wallets, bridges, aggregators and solver networks are all targeting similar interoperability and chain-abstraction problems.
• Value capture is still modest: NEAR Intents has processed more than $27 billion in swaps, but cumulative token buybacks remain only a few million dollars.
• The AI thesis is still speculative: NEAR has real AI infrastructure, but large-scale autonomous-agent adoption has not yet arrived.
• Strategic focus has shifted repeatedly: NEAR has moved from a usability-focused Layer 1 toward chain abstraction and now AI agents, which can make long-term product-market fit harder to assess.
• More infrastructure means more dependencies: MPC signers, solvers, verifier contracts, TEEs and private validators all introduce additional security and economic risks.
• Technical progress has not yet translated into a large native economy: NEAR's infrastructure is increasingly capable, but adoption and token value capture still need to catch up.

Who Is NEAR Best Suited To?

NEAR may suit investors who believe chain abstraction will become an important part of crypto infrastructure and who see current Intents volume as evidence that the market is starting to form.

It may also suit:

• Developers building applications that need to interact with several chains.
• Investors who expect cross-chain settlement to grow without users permanently migrating between ecosystems.
• Investors interested in AI-agent payments and confidential execution.
• Users comfortable evaluating infrastructure adoption separately from native DeFi TVL.

NEAR may be less compelling for:

• Investors who want strong current protocol revenue.
• Users seeking the deepest native DeFi liquidity.
• Buyers whose thesis depends mainly on the AI narrative.
• Investors uncomfortable with MPC, solver and TEE dependencies.

The more useful indicators are Intents volume, retained revenue, NEAR buybacks, native liquidity, validator concentration and measurable AI-agent transactions.

If those numbers begin improving together, the investment case becomes much stronger without needing an arbitrary price target.

How to Buy and Store NEAR

NEAR is available on major centralized exchanges and can be held through exchange custody, NEAR-compatible software wallets or hardware wallets.

Where to Buy NEAR

Major crypto exchanges currently supporting NEAR include Binance, Coinbase, Kraken and several other global trading venues.

There is little reason to select an exchange solely because it lists NEAR. Trading fees, jurisdiction, fiat support, withdrawals, security and account requirements are more useful decision points.

Coin Bureau's best crypto exchanges guide provides the broader comparison.

Once NEAR is purchased, users should verify the withdrawal network and destination address before moving funds off the exchange.

Best Wallets for NEAR

The right wallet depends mainly on how often the holder intends to use the assets.

• Hardware wallets suit longer-term storage because signing keys stay isolated from an ordinary internet-connected device. A hardware wallet is usually a better fit when reducing key exposure matters more than frequent DApp interaction.
• NEAR-compatible software wallets such as Meteor, HOT Wallet, HERE Wallet and Sender are easier for everyday transfers and applications.
• Staking-capable wallets let holders delegate NEAR directly through the wallet interface.

NEAR Protocol Review: Final Verdict

NEAR Protocol has much clearer technological identity today.

The chain-abstraction strategy now has real infrastructure behind it. NEAR Intents has processed more than $27 billion in swaps. Chain Signatures lets NEAR accounts and contracts authorize transactions on external networks. Nightshade has progressed to nine shards, stateless validation and dynamic resharding. NEAR AI has moved into working products through IronClaw, confidential inference and Private Shard.

The strongest version of the NEAR thesis is becoming easier to see. The project may be more valuable as infrastructure that applications consume across several ecosystems than as a chain that tries to pull every user and every dollar of liquidity into its own native environment.

The token still has more to prove.

Native DeFi liquidity remains modest. Base-layer revenue is small. Intents buybacks create real value capture, but the amounts remain far below gross annual issuance. AI-agent infrastructure is running, yet large-scale economic demand from autonomous agents has not arrived.

If those pieces start moving together, NEAR could occupy a role that ordinary Layer 1 rankings fail to capture.