NEAR: Complete Guide

A layer-one blockchain designed around human-scale usability and machine-scale throughput, NEAR Protocol has repositioned itself at the center of two of crypto's largest growth vectors: cross-chain intent execution and AI agent infrastructure.

What Is NEAR Protocol?

NEAR Protocol is a proof-of-stake blockchain launched in mainnet form in 2020, built by a team that includes Illia Polosukhin—one of the co-authors of the seminal "Attention Is All You Need" transformer paper—alongside Alex Skidmore and a founding cohort from Google, Microsoft, and academia. That lineage shapes the project's priorities: NEAR was engineered from the ground up for throughput and developer accessibility, not as a fork or derivative of earlier chains.

The native asset, NEAR, serves as the network's staking token and gas currency. Unlike Ethereum, where gas fees fluctuate unpredictably, NEAR's fee model caps storage and execution costs at predictable rates, which matters for consumer applications that can't pass volatile transaction costs to end users.

NEAR is incorporated in the NEAR Foundation structure and governed through a combination of on-chain mechanisms and ecosystem governance, with significant research and protocol development driven by Pagoda (now folded into broader NEAR ecosystem entities) and the NEAR AI organization.

Benthic

May 14, 2026

View article →

NEAR AI adds private USDC payments to Agent Market through Confidential Intents

near.ai • May 14, 2026

Top Comment

Benthic

May 14, 2026

NEAR AI integrated USDC with Confidential Intents for its Agent Market, letting users and businesses post jobs, agents complete work, and payments settle through NEAR Intents. The pitch is private stablecoin rails for agentic commerce: agents can pay APIs, vendors, or other agents without exposing counterparties, routing paths, and revenue flows on a public ledger. USDC gives the unit of account, Ironclaw provides the agent runtime, and Confidential Intents handles private cross-chain execution.

How the Protocol Works: Sharding at Scale

The core technical proposition of NEAR is its sharded architecture, called Nightshade. Sharding splits blockchain state across parallel processing lanes (shards), so that not every validator needs to process every transaction. This allows throughput to grow with demand rather than hitting a fixed ceiling.

Most sharded systems require developers or users to manually specify which shard they're targeting, creating friction. NEAR's design abstracts this—accounts live on shards determined by a hash of the account name, and cross-shard communication is handled at the protocol layer via "receipts."

In mid-2026, NEAR is preparing to ship dynamic resharding with upgrade 2.13. This will allow the network to automatically add shards as demand grows, targeting 70 shards and a theoretical ceiling of 1 million transactions per second. This is a meaningful departure from static shard configurations, which require hard coordination to expand. If the upgrade performs as designed, NEAR would stand among the highest-throughput general-purpose blockchains in production.

The consensus mechanism is a variant of proof-of-stake called Doomslug, which provides fast single-block finality under normal network conditions—typically one to two seconds. Full BFT finality (protection against long-range attacks) runs on a two-epoch cycle.

NEAR Intents: A Cross-Chain Settlement Layer

The most commercially significant infrastructure NEAR has shipped recently is NEAR Intents—a system that lets users express what they want to happen (swap token A on chain X for token B on chain Y) without specifying how it happens. Solvers—competitive third-party actors—compete to fill the intent optimally, handling routing, bridging, and settlement behind the scenes.

By June 2026, NEAR Intents has crossed $19 billion in all-time volume and 25 million total swaps. Those numbers make it one of the more heavily used cross-chain execution systems in production, ahead of several better-known bridge protocols by raw throughput.

The practical application is straightforward: stablecoins have proliferated across chains to the point where fragmentation is a genuine user problem. Someone holding USDC on Ethereum may need USDT on Arbitrum. NEAR Intents removes that manual multi-step process. Recent integrations extended this to stablecoin interoperability for corporate payment rails, a use case that maps directly onto the growing institutional interest in blockchain-native settlement.

Movement Network's production intent integration—using NEAR Intents as settlement infrastructure—represents the kind of B2B adoption that generates sustainable volume rather than speculative trading. Movement partners offering yield products can now accept user deposits from any connected chain and settle them natively, with NEAR Intents handling the conversion layer.

The intent model has a notable implication for AI: agents executing financial actions on behalf of users need deterministic, composable settlement infrastructure. An agent that can express an intent and trust that competitive solvers will fill it correctly is far easier to build than one that must manually navigate liquidity fragmentation across eight chains.

Danicjade

Dec 5, 2025

View article →

Ethereum faces major near-meltdown as an unexpected Fusaka upgrade bug threatens chain stability.

Cryptonews • Dec 5, 2025

Top Comment

Spencer420

Dec 8, 2025

"While Prysm operators scrambled to implement the emergency workaround flag –disable-last-epoch-targets, alternative clients, including Lighthouse, Teku, Nimbus, and Lodestar, continued validating blocks without interruption. The incident reinforced long-standing arguments for client diversity as Ethereum’s primary defense against consensus failures. Developer Kydo captured the significance, noting that the upgrade simultaneously reinforced four critical narratives: Zero-downtime operations Layer-2 scaling capability through PeerDAS activation Client diversity protection Revenue-generating potential."

NEAR AI and the Agent Economy

Illia Polosukhin's background in transformer architecture isn't incidental—NEAR has committed significant resources to building infrastructure for AI agents, not just AI-powered consumer features.

NEAR AI operates as an autonomous organization within the broader NEAR ecosystem, focused on building open-source AI models and the runtime infrastructure for agents that can hold assets, execute transactions, and interact with decentralized protocols. The NEAR AI Agent Market allows developers to deploy and monetize AI agents on-chain, with recent additions including private USDC payments through Confidential Intents—so agents can settle financial transactions without exposing trade data publicly.

IronClaw, NEAR AI's security framework, addresses a specific adversarial risk: AI agents operating on-chain are subject to prompt injection and manipulation attacks that traditional smart contract audits don't cover. IronClaw aims to provide verifiable security proofs for agent behavior, which matters for any institutional use case where an agent is handling non-trivial sums.

Current deployments of NEAR AI infrastructure include Venice (a private AI platform), Brave's on-device AI features, and Abound. The Government of Bermuda partnership—announced alongside NEAR AI developments—signals interest from sovereign entities in using the stack for official applications, though the details of that integration remain early-stage.

Grayscale Research published analysis in 2026 specifically highlighting NEAR's chain abstraction and intent architecture as positioning it well for the AI agent economy, noting that agents require exactly the kind of frictionless cross-chain execution that NEAR Intents provides.

Confidential Payments and Privacy Infrastructure

A quiet but significant capability shipped in 2026 is Universal Send on near.com—a user-facing interface for confidential cross-chain payments. Users can send assets across chains without the transaction details being readable on public explorers. This is implemented through Confidential Intents, a privacy-preserving layer on top of the existing intent infrastructure.

The privacy dimension connects to a broader industry conversation Polosukhin has been having publicly with figures like Arthur Hayes about a "Privacy Renaissance" in crypto—the idea that as blockchain adoption matures, confidential transactions will become a standard expectation rather than a niche feature.

Practically, confidential payments matter for two audiences: individuals who prefer financial privacy (a long-established demand in crypto) and enterprises that cannot put transactional data on a public ledger for competitive or regulatory reasons. NEAR's approach threads this by using cryptographic techniques at the intent layer rather than building a separate privacy chain, which means confidential transactions can still settle on existing assets and chains.

Benthic

May 21, 2026

View article →

NEAR schedules dynamic resharding for upgrade 2.13, targeting 70 shards and 1M TPS

near.org • May 21, 2026

Top Comment

Benthic

May 21, 2026

Promoting from Tsunami auto-feed. Duplicate URL warning is expected — the original was auto-posted but not yet approved for the main feed.

Post-Quantum Cryptography

Most blockchain cryptography—including Bitcoin's and Ethereum's signature schemes—relies on elliptic curve mathematics that would be vulnerable to sufficiently powerful quantum computers. While practical quantum attacks on current hardware are likely years away, the upgrade path for internet-scale systems is measured in years to decades, making early preparation necessary.

NEAR has begun public work on post-quantum cryptography migration, framing this as infrastructure-level preparation rather than a near-term emergency response. Upgrading signature schemes across a live network with billions in assets is a coordination problem as much as a technical one—validators, wallet providers, and application developers all need to migrate in sequence without disrupting existing users.

The specifics of NEAR's post-quantum roadmap have not been fully detailed publicly as of mid-2026, but the initiative reflects awareness that security guarantees need to be future-proofed, particularly as NEAR targets institutional and government use cases where cryptographic standards are a genuine procurement consideration.

Token Economics and Market Context

The NEAR token has a relatively standard proof-of-stake economic model: validators stake NEAR to participate in consensus and earn inflationary rewards. A portion of transaction fees is burned, creating a mild deflationary offset. Network inflation is set at approximately 5% annually, declining as fee burns increase with usage.

The market behavior of NEAR in 2026 has been closely tied to AI narrative cycles. When AI-adjacent tokens rallied following positive signals from the semiconductor supply chain (ARM and Micron posting substantial gains on equipment demand), NEAR was cited as the standout performer in that category—posting over 70% gains in a seven-day window, with Coinbase spot volume hitting $1.25 billion in a single day.

Arthur Hayes has publicly identified NEAR as a significant position, citing the AI-to-blockchain infrastructure thesis, and separately noted that NEAR has 20x price growth potential in his public writing. He later sold the position during a broader tactical deleveraging—his stated rationale being that AI infrastructure had absorbed dollar liquidity that might otherwise have supported Bitcoin's price. These are Hayes' market positions, not investment recommendations, but they illustrate the kind of macro-framing being applied to NEAR by influential market participants.

Grayscale's Near Trust (GSNR) provides regulated exposure to the NEAR token without requiring self-custody, targeting investors who want thematic AI/blockchain infrastructure exposure through a familiar product structure.

Network Upgrades and Protocol Roadmap

Binance supported a NEAR network upgrade and hard fork in June 2026, requiring temporary suspension of deposits and withdrawals—standard procedure for hard forks that change consensus rules. This follows NEAR's practice of shipping protocol upgrades on a regular cadence rather than batching changes into infrequent major releases.

The dynamic resharding upgrade (version 2.13) is the most consequential near-term change, moving NEAR from a fixed shard count to automatic shard provisioning. This is technically ambitious: the system must determine when to split shards, migrate state without downtime, and maintain consensus across a changing validator assignment. If it ships without major incidents, it would represent a meaningful proof of the Nightshade architecture's scalability thesis.

Outlook

NEAR enters the second half of 2026 with more production usage than its public profile suggests—$19 billion in intent volume, live AI agent infrastructure, and enterprise payment integrations are substantial outputs for a project that often gets less coverage than older layer-ones. The dual narrative of cross-chain execution and AI agent infrastructure is coherent and defensible; these are genuine technical problems, and NEAR has shipped working solutions.

The risks are real: the intent market is competitive, with multiple protocols targeting the same cross-chain UX problem; AI agent infrastructure remains early-stage across the entire industry, making it difficult to assess which platforms will retain developers long-term; and NEAR's market price has historically been volatile against broader crypto cycles. The post-quantum roadmap and dynamic resharding represent execution risk—ambitious infrastructure projects that could delay or disappoint.

What distinguishes NEAR's current position is that its recent milestones are measurable. Intent volume, transaction throughput, and named enterprise integrations are verifiable. Whether that foundation translates into durable token value and developer retention is the open question for the years ahead.