The altcoin market witnessed a tectonic shift this week as NEAR Protocol decoupled from the broader market, recording a massive 50% surge to reach a six-month high. This breakout, fueled by the announcement of the v2.13 “Dynamic Resharding” upgrade and the launch of the “IronClaw” AI runtime, marks NEAR’s definitive transition into the primary infrastructure layer for the 2026 agentic economy.
By Jennifer Kim | May 24, 2026
Protocol Primer
For years, the “scalability trilemma” has forced blockchains to choose between decentralization, security, and throughput. While competitors like Solana (trading at $86) and Ethereum ($2,114) have pursued monolithic speed or Layer 2 fragmentation, NEAR Protocol has double-downed on its Nightshade sharding architecture. As of May 24, 2026, NEAR is no longer just a “fast blockchain”; it has repositioned itself as the AI-native settlement layer.
The protocol’s core purpose in the current cycle is to facilitate “User-Owned AI.” Unlike centralized AI models controlled by big tech, NEAR provides a decentralized environment where AI agents—autonomous bots capable of financial sovereignty—can execute complex tasks without human oversight. This “Agentic Web” requires a level of throughput and privacy that legacy L1s struggle to provide. By unifying Chain Abstraction with a self-scaling backend, NEAR aims to hide the complexities of blockchain from the end user, allowing the Near.com Super-App to manage assets across 35+ chains via simple FaceID or email authentication.
Key Innovations
The primary technical catalyst for the recent rally is the upcoming v2.13 Network Upgrade, scheduled for June 2026. This release introduces Dynamic Resharding, a milestone that moves the network from a manually managed system to an autonomous, self-scaling infrastructure. Historically, adding shards required intense coordination among validators and governance votes. With v2.13, the protocol monitors state size and compute load in real-time, automatically splitting or merging shards in a single block (under 2 seconds) when capacity thresholds are met.
Alongside scaling, NEAR has introduced the IronClaw AI runtime, a security-hardened framework built in Rust. Designed to run inside Trusted Execution Environments (TEEs), IronClaw provides “provider-blind” execution. This means that an AI agent can hold its own private keys and process sensitive financial data in an encrypted enclave, where even the GPU cloud operator cannot see the plaintext data. This is paired with Confidential Intents, a private execution layer that prevents MEV bots from front-running AI-driven trades.
- 1 Million TPS Benchmark — Recent public testing demonstrated the network’s path toward one million transactions per second as it scales toward 70+ shards.
- Post-Quantum-Safe Signing — v2.13 integrates quantum-resistant cryptographic signatures, making NEAR one of the first “quantum-ready” Layer 1s in the industry.
- WASM Sandboxing — IronClaw utilizes WebAssembly to isolate every AI “skill,” ensuring that a malicious tool cannot compromise the agent’s core vault or outbound network connections.
Tokenomics Breakdown
The economic model of the NEAR token underwent a radical transformation in late 2025 with the “Halving Upgrade,” which the market is only now beginning to fully price in. To combat the dilution concerns of the previous era, the community successfully reduced the maximum annual inflation rate from 5% to 2.5%. This has stabilized staking rewards at approximately 4.75%, a yield that has attracted significant institutional inflows via the Bitwise NEAR Staking ETP in Europe.
Crucially, NEAR’s burn mechanism has evolved into a “deflationary flywheel.” While 70% of gas fees are still burned, the protocol now also burns 100% of the protocol fees generated by NEAR Intents, the cross-chain settlement engine. Furthermore, the AI Agent Economy has introduced a new demand sink: State Staking. To store data or run long-lived agent instances, users must “rent” storage by locking NEAR tokens. As millions of autonomous agents launch, this “state rent” effectively removes a significant portion of the circulating supply from the open market, creating a net-deflationary pressure as network usage grows.
As of May 2026, the NEAR token supply is 100% fully unlocked. With the expiration of all VC and team vesting cliffs, the “overhang” that plagued the token in 2024 and 2025 has vanished, allowing the asset to respond directly to the fundamental demand generated by its AI and infrastructure partnerships.
Roadmap Reality Check
NEAR’s “Mission 10 Billion”—the goal of onboarding 10 billion users—is no longer a purely social target; it is a technical roadmap. The progression from 6 shards in 2024 to the current self-managing sharded architecture proves that the Nightshade design can deliver on its promises. However, the roadmap has faced challenges. The transition to sharded RPC and archival nodes has been slower than expected, leading to occasional latency issues for developers during the initial rollout of the Bermuda AI-powered public services initiative earlier this month.
While the 1 million TPS benchmark is impressive, it remains a “theoretical maximum” in controlled environments. The “reality check” for investors is whether the network can maintain these speeds as it moves from its current shard count toward the 70-shard target envisioned in the Nightshade 3.0 final phase. The upcoming House of Stake governance model, which uses AI-driven mechanisms to manage protocol parameters, will be the ultimate test of the network’s decentralization as it scales beyond human-managed limits.
Investor Takeaway
The bull case for NEAR is centered on its lead in the AI-Web3 convergence. By solving both the scaling bottleneck (via Dynamic Resharding) and the privacy bottleneck (via IronClaw TEEs), NEAR has positioned itself as the indispensable substrate for the next generation of autonomous commerce. The endorsement from Arthur Hayes, who recently named NEAR part of his “holy trinity” of infrastructure assets, has added significant momentum to the institutional rotation away from speculative memes and toward hard utility.
However, risks remain. The complexity of a sharded L1 architecture is inherently higher than that of monolithic competitors like Solana. Furthermore, as the CLARITY Act moves toward Senate passage in July, the regulatory spotlight on “AI-native” assets could intensify. Investors should watch the June v2.13 activation closely; a flawless execution of Dynamic Resharding would likely cement NEAR’s position as the primary competitor to the Ethereum-Solana duopoly. For now, the combination of halved inflation and surging AI demand suggests that NEAR is entering a fundamental re-rating phase that transcends simple price action.
The cryptocurrency market remains highly volatile. This article is for informational purposes only and does not constitute financial advice.
Disclaimer: This article is for informational purposes only and does not constitute financial advice. Always conduct your own research before making any investment decisions. Cryptocurrency markets are highly volatile, and past performance is not indicative of future results.
50% surge on a resharding upgrade is aggressive but the IronClaw AI runtime narrative is strong. NEAR positioning as the AI settlement layer is smart
sol at 86 and ETH at 2114 while NEAR surges 50%. rotation into AI infra tokens is the trade right now
1M TPS is the marketing number. real world throughput with AI workloads will be a fraction of that. still impressive for Nightshade v2 though
Anika the 1M claim was on a stress test with synthetic txs, not AI workloads. nobody is settling ML inference on NEAR today
IronClaw runtime is the actual story here. AI settlement layer narrative is what drove the pump, not the sharding upgrade itself
delegate_max IronClaw is riding the AI agent narrative but nobody is training models on NEAR. the settlement layer pitch needs actual AI workloads not just a runtime brand
50% on a sharding upgrade while ETH does nothing is the most 2026 thing possible. market punishes execution and rewards narratives
dynamic resharding is what ETH has been trying to do for years. NEAR just shipped it. 1 million TPS target is ambitious but the Nightshade architecture is legit
ETH abandoned sharding for L2s because the engineering was too hard. NEAR actually shipped it. whether the market rewards that is a different question
shard_skeptic ETH didnt abandon sharding because it was too hard, they pivoted because danksharding + L2s gave them more headroom without redesigning the L1. different engineering call
throughput_nerd even 100K real TPS would be impressive but nobody is running AI training workloads on NEAR. the 1M claim is pure marketing
been in this space since 2017 and i have seen plenty of AI-native chains come and go. NEAR has the dev activity to back it up though, will give them that
seen this AI-native chain narrative before with Fetch and SingularityNET. NEAR actually shipping code instead of just whitepapers is the difference
50 percent pump on resharding while Solana does 65k TPS without sharding. market rewards narratives over working products every time
1M TPS is the marketing number sure but even 100K real throughput puts NEAR ahead of every L1 except maybe solana under ideal conditions
50% in a week on an upgrade announcement is retail chasing the AI tag. seen this with FET AGIX OCEAN and now NEAR. the pump is the exit liquidity event
nocoin_aunt calling it exit liquidity but NEAR shipped 1M TPS testnet results last month. how many AI tokens actually have working infrastructure vs whitepapers
v2.13 dynamic resharding means shards split based on load without hard forks. thats actually a meaningful upgrade, not just a whitepaper bullet point
shard_count_ the Nightshade split mechanism is technically solid but the IronClaw runtime branding feels like AI washing. show me actual ML workloads settling on NEAR
Yuna K. the IronClaw AI washing concern is valid. show me one ML workload actually settling on NEAR and ill change my tune. until then this is a token narrative riding the AI wave
infra_realist exactly. show me one ML workload settling on NEAR. the IronClaw runtime is a branding exercise until proven otherwise
infra_realist asking for ML workloads on NEAR is fair. every chain claims AI integration but none show actual settlement data. IronClaw needs proof not branding
shard_count_ dynamic resharding where shards split based on load without a hard fork is a real engineering achievement. ETH gave up on L1 sharding and just piled L2s on top instead
ETH gave up on L1 sharding and just stacked L2s creating fragmentation. NEAR doing dynamic resharding without hard forks is genuinely different engineering
50% pump on a resharding announcement while Solana sits at 86 and ETH at 2114. this is narrative driven rotation not fundamentals