This space is available Advertise with CryptoOverlook → 160 × 600
This space is available Advertise with CryptoOverlook → 160 × 600
This space is available Advertise with CryptoOverlook → 728 × 90

Coins Explained

What Is Near Protocol? A Sharded Chain Now Selling Itself on AI

How Near's sharding design actually works, and how the project has shifted its public pitch from everyday apps toward AI agents, with an honest look at what's changed and what hasn't.

What problem Near was built to solve

Near Protocol is a Layer 1 blockchain, founded by Illia Polosukhin and Alexander Skidanov and launched on mainnet in 2020, built around a specific technical bet: that the way to make a blockchain handle more transactions without becoming more centralized is sharding, splitting the network into multiple parallel pieces that each process a portion of the total activity, rather than having every single node process every single transaction. If the basic idea of a blockchain and why agreement across many computers takes real work is new to you, What Crypto Actually Is is worth reading first.

Sharding is a genuinely hard engineering problem for blockchains specifically because security depends on having enough independent validators checking each other’s work; split the network into pieces carelessly and you can end up with each individual shard being easier to attack than the network was as a whole. Near’s answer to that problem is called Nightshade.

The mechanism, in plain language

Picture a single, old-fashioned assembly line where every worker has to inspect every part of every product before it moves on: thorough, but slow, since nothing can happen in parallel. Sharding is the alternative: multiple assembly lines running at once, each handling a slice of the total work. The hard part is making sure each individual line is still trustworthy on its own.

Nightshade’s approach is to keep a single underlying blockchain (one shared record, one native token, one set of validators securing the whole system) while having different shards produce their own piece, called a “chunk,” of each new block in parallel. Validators are randomly and periodically reassigned across shards, and a chunk from any given shard only gets accepted once it’s been validated according to the network’s overall rules, so an attacker can’t simply target one shard in isolation and treat it as an easier, separate target from the rest of the network. In practice, this lets Near increase its transaction capacity by adding more shards as demand grows (a process called resharding) rather than requiring every validator to individually handle the entire network’s transaction volume forever. Near’s more recent network upgrades have moved toward dynamic resharding, where shards can split automatically as network usage grows, rather than requiring a manually coordinated upgrade each time capacity needs to increase.

The pivot: from “everyday apps” to AI

For its first several years, Near’s public pitch centered on being a fast, low-fee, developer-friendly chain for mainstream consumer apps, deliberately positioned as more approachable than Ethereum for people without a crypto background. More recently, Near’s messaging has shifted substantially toward artificial intelligence, branding itself around what it calls “User-Owned AI” and a system called NEAR Intents. The pitch is that AI agents will increasingly need to transact on someone’s behalf (paying for data, executing trades, coordinating across services) and that doing this through a blockchain, rather than through a single company’s servers, lets a user keep verifiable control and a transaction record of what an agent did in their name, instead of trusting a company’s black box. NEAR Intents is built around chain abstraction: letting a user or an agent express what they want to happen (say, “swap this asset for that one, wherever it’s cheapest”) without manually bridging assets between different blockchains themselves.

It’s worth being direct about how to read this shift. Sharding and fast, cheap transactions are a real, working piece of underlying technology, verifiable independent of any marketing narrative. The AI framing on top of it is a newer, deliberately timed repositioning that rides the broader AI hype cycle sweeping tech generally, and it’s genuinely unclear yet how much real usage (agents actually transacting through NEAR Intents at meaningful volume) will materialize versus how much of this is branding built ahead of the product. Both things can be true: the technical foundation is real, and the current pitch built on top of it is unproven.

Strengths, and real limitations

Near’s strength is a working, live sharding implementation that has been operating and iterating for years, not just a whitepaper concept, along with fast confirmation and low fees that have made it usable for the kind of high-frequency activity chain abstraction and AI agents would actually need. It also, unlike some Layer 1s built primarily around trading and speculation, has consistently emphasized approachability for regular developers and users, with human-readable account names instead of long cryptographic addresses.

The honest limitations include real governance friction: in late 2025, Near’s core team implemented a reduction in the network’s token issuance rate (halving annual inflation) despite the change failing to pass a formal community governance vote, which drew legitimate criticism about how much weight community votes actually carry on this network. There’s also an open, unresolved debate within Near’s own leadership about long-term token supply policy, including disagreement over proposals to burn tokens versus simply capping future issuance, that hasn’t been fully settled. And as with any narrative pivot, there’s a real risk that “AI chain” ends up being a repositioning that outruns actual product usage; the size of that gap is something to check with current, independent usage data rather than take from Near’s own marketing.

How to actually trade or hold it

NEAR trades on spot markets at most major exchanges; see our exchange comparisons to check availability and fees. If you’re new to buying crypto, Spot Trading Explained covers order types and custody basics first. Because Near’s current price narrative is tied heavily to AI-agent adoption and usage of NEAR Intents specifically, it’s worth tracking actual on-chain usage data for that product, not just the marketing around it, separately from the underlying sharding technology, which is a more stable, independently verifiable fact about the network.

For the current price, market cap, and 24-hour chart, see our live NEAR page.

Frequently asked questions

Is Near actually an AI project? No, not originally, and not at the protocol level. Near is a sharded Layer 1 blockchain. Its more recent public positioning leans heavily on AI agents and “chain abstraction” as a use case for that existing technology, but the underlying network is general-purpose infrastructure, not AI software itself.

What is sharding, in one sentence? Splitting a blockchain’s workload across multiple parallel groups of validators instead of requiring every validator to process every transaction, which can increase total capacity without concentrating trust in fewer hands, if done carefully.

Did Near’s community actually vote for the recent inflation change? No. The reduction in Near’s token issuance rate was implemented in late 2025 despite failing to pass a formal community governance vote, which is a legitimate point of criticism about how much practical authority Near’s token-holder governance actually has over the core team’s decisions.

How is Near different from Ethereum’s approach to scaling? Ethereum has scaled primarily by pushing activity onto separate Layer 2 networks that settle back to Ethereum periodically, covered in What Are Layer 2s?. Near instead scales its own single base layer directly through sharding, splitting the work within one network rather than offloading it to separate chains on top.

Risk

Nothing on this page is financial advice, and it is not a complete or current statement of Near’s technical status, price history, or regulatory situation, all of which can change. Near’s current public narrative leans heavily on AI-agent adoption that is still unproven at scale, and its token issuance policy has already been changed once outside of a passing community vote, both of which are worth weighing separately from the underlying sharding technology. Do your own research before buying or holding any crypto asset.

This space is available Advertise with CryptoOverlook → 728 × 90