<p>Have you ever wondered how blockchain networks handle increasing demand without slowing down? </p>



<p><strong>Adaptive State Sharding</strong> is a cutting-edge solution that helps blockchains scale efficiently while maintaining security and decentralization. It dynamically splits and reorganizes data across multiple shards, allowing networks to process more transactions in parallel.</p>



<p>This approach is central to MultiversX (the network formerly known as Elrond), which deployed adaptive state sharding at the protocol level and has continued refining it through upgrades like Andromeda v1.9.6 in May 2025. By understanding Adaptive State Sharding, you can see how modern blockchains optimize speed, reduce congestion, and provide a smoother user experience.</p>



<h2 class="wp-block-heading tablecontent">Quick Answer: What Is Adaptive State Sharding?</h2>



<p>Adaptive state sharding splits a blockchain into smaller parallel segments called shards. Each shard handles its own slice of the network state, its own transactions, and its own validator group. The system automatically adds or removes shards as demand rises or falls.</p>



<p>Three things make it &#x27;adaptive&#x27; rather than plain sharding:</p>



<ol class="wp-block-list">
<li>State sharding: validators store only their shard&#x27;s data, not the full chain.</li>



<li>Transaction sharding: transactions route to the relevant shard and process in parallel.</li>



<li>Network sharding: validator assignments rotate across shards to prevent concentrated attacks.</li>
</ol>



<p>MultiversX is the primary live implementation. Its May 2025 Andromeda upgrade set 400 validators per shard and removed confirmation blocks entirely, cutting finality to near-instant from a previous target of around 6 seconds.</p>



<p>The rest of this article covers how each layer works, how it compares to other scaling approaches, and what the real tradeoffs are.</p>



<h2 class="wp-block-heading tablecontent">Adaptive State Sharding vs. Other Scaling Approaches</h2>





<p style="margin:0 0 16px 0;font-size:16px">
👉 <strong>Quick takeaway:</strong> Adaptive state sharding (MultiversX) and DAG-based architectures offer the fastest finality in the table. Layer-2 rollups on Ethereum deliver the deepest ecosystem depth but introduce withdrawal delays and L1 settlement dependencies. Fixed sharding remains a research baseline rather than a production architecture.
</p>
<div style="overflow-x:auto;margin:0 0 24px 0">
  <table style="width:100%;border-collapse:collapse;font-size:16px;line-height:1.5;min-width:1000px;border:1px solid #e5e7eb"><thead><tr style="background-color:#f8fafc;text-align:left"><th style="padding:14px 16px;border-bottom:1px solid #e5e7eb">Approach</th><th style="padding:14px 16px;border-bottom:1px solid #e5e7eb">How It Scales</th><th style="padding:14px 16px;border-bottom:1px solid #e5e7eb">Finality</th><th style="padding:14px 16px;border-bottom:1px solid #e5e7eb">Main Tradeoff</th><th style="padding:14px 16px;border-bottom:1px solid #e5e7eb">Best For</th></tr></thead><tbody><tr><td style="padding:14px 16px;border-bottom:1px solid #e5e7eb"><strong>Adaptive State Sharding</strong><br/><span style="font-size:14px;color:#6b7280">MultiversX</span></td><td style="padding:14px 16px;border-bottom:1px solid #e5e7eb">Splits state, transactions, and network across dynamic shards</td><td style="padding:14px 16px;border-bottom:1px solid #e5e7eb">
          <span style="color:#15803d;font-weight:600">🟢 Near-instant</span><br/>
          <span style="font-size:14px;color:#6b7280">After Andromeda v1.9.6 (May 2025)</span>
        </td><td style="padding:14px 16px;border-bottom:1px solid #e5e7eb">Cross-shard communication complexity</td><td style="padding:14px 16px;border-bottom:1px solid #e5e7eb">
          High-throughput L1 apps<br/>
          <span style="color:#15803d;font-weight:600">🏆 Most complete sharding implementation</span>
        </td></tr><tr style="background-color:#fcfcfd"><td style="padding:14px 16px;border-bottom:1px solid #e5e7eb"><strong>Fixed Sharding</strong><br/><span style="font-size:14px;color:#6b7280">Early Ethereum proposal</span></td><td style="padding:14px 16px;border-bottom:1px solid #e5e7eb">Fixed number of shards, static validator sets</td><td style="padding:14px 16px;border-bottom:1px solid #e5e7eb">
          <span style="color:#d97706;font-weight:600">⚠️ Slower; dependent on shard count</span>
        </td><td style="padding:14px 16px;border-bottom:1px solid #e5e7eb">
          <span style="color:#dc2626;font-weight:600">🔴 Inflexible under load spikes</span>
        </td><td style="padding:14px 16px;border-bottom:1px solid #e5e7eb">
          Research baseline only<br/>
          <span style="color:#d97706;font-weight:600">⚠️ Not deployed at production scale</span>
        </td></tr><tr><td style="padding:14px 16px;border-bottom:1px solid #e5e7eb"><strong>Layer-2 Rollups</strong><br/><span style="font-size:14px;color:#6b7280">Optimism, Arbitrum</span></td><td style="padding:14px 16px;border-bottom:1px solid #e5e7eb">Batches transactions off-chain, settles on L1</td><td style="padding:14px 16px;border-bottom:1px solid #e5e7eb">
          <span style="color:#d97706;font-weight:600">⚠️ Minutes to hours for full L1 settlement</span>
        </td><td style="padding:14px 16px;border-bottom:1px solid #e5e7eb">Withdrawal delays; relies on L1 security</td><td style="padding:14px 16px;border-bottom:1px solid #e5e7eb">
          DeFi apps on Ethereum<br/>
          <span style="color:#15803d;font-weight:600">🏆 Best for Ethereum ecosystem depth</span>
        </td></tr><tr style="background-color:#fcfcfd"><td style="padding:14px 16px;border-bottom:1px solid #e5e7eb"><strong>Near Nightshade Sharding</strong><br/><span style="font-size:14px;color:#6b7280">NEAR Protocol</span></td><td style="padding:14px 16px;border-bottom:1px solid #e5e7eb">Single-chain appearance with sharded execution</td><td style="padding:14px 16px;border-bottom:1px solid #e5e7eb">
          <span style="color:#15803d;font-weight:600">🟢 Fast (~1-2s)</span>
        </td><td style="padding:14px 16px;border-bottom:1px solid #e5e7eb">Chunk-only producer model; different trust assumptions</td><td style="padding:14px 16px;border-bottom:1px solid #e5e7eb">
          NEAR ecosystem dApps<br/>
          <span style="color:#15803d;font-weight:600">🏆 Best UX abstraction over sharding complexity</span>
        </td></tr><tr><td style="padding:14px 16px"><strong>DAG-Based</strong><br/><span style="font-size:14px;color:#6b7280">Fantom, Hedera</span></td><td style="padding:14px 16px">Directed acyclic graph replaces linear chain</td><td style="padding:14px 16px">
          <span style="color:#15803d;font-weight:600">🟢 Sub-second in practice</span><br/>
          <span style="color:#15803d;font-weight:600">🏆 Fastest finality in table</span>
        </td><td style="padding:14px 16px">DAG security assumptions differ from chain-based</td><td style="padding:14px 16px">
          High-frequency microtransactions<br/>
          <span style="color:#15803d;font-weight:600">🏆 Best for sub-second high-frequency use cases</span>
        </td></tr></tbody></table>
</div>




<p>Adaptive state sharding is the only approach in this table that combines all three sharding dimensions (state, transaction, network) in a single live L1 protocol. That integration is its main structural advantage. The main cost is cross-shard coordination, which grows in complexity as shard count increases.</p>



<h2 class="wp-block-heading tablecontent">How Adaptive State Sharding Works</h2>



<p>Adaptive state sharding combines three layers into one cohesive protocol. Each layer handles a distinct problem. Together, they allow a blockchain to process transactions in parallel at a scale that single-chain architectures cannot match.</p>



<h3 class="wp-block-heading">State Sharding</h3>



<p>State sharding ensures that <a href="https://www.thecoinzone.com/blockchain/ashdraked-meaning/" data-wpil-monitor-id="77">blockchain</a> data is divided across multiple shards, meaning validators only store and process a portion of the entire blockchain instead of the full dataset. This significantly reduces storage requirements and increases efficiency.</p>



<h3 class="wp-block-heading">Transaction Sharding</h3>



<p>Transaction sharding ensures that transactions are processed in parallel across different shards, rather than being queued for validation on a single chain. This speeds up confirmation times and increases overall throughput.</p>



<h3 class="wp-block-heading">Network Sharding</h3>



<p>Network sharding controls which validators talk to which nodes. Each shard gets its own validator group. Those groups rotate at the end of each epoch, which makes it expensive for an attacker to corrupt a single shard because they cannot predict or lock in a target set before the rotation happens. In MultiversX, Andromeda v1.9.6 set 400 validators per shard as part of this consensus model.</p>



<h3 class="wp-block-heading">How Cross-Shard Communication Works</h3>



<p>When a transaction touches accounts in two different shards, the protocol cannot process it entirely within one shard. Here is the sequence:</p>



<ol class="wp-block-list">
<li>The originating shard processes the sender&#x27;s side and creates a cross-shard message.</li>



<li>The metachain (a coordination layer above the execution shards) records and orders that message.</li>



<li>The destination shard picks up the message in the next round and finalizes the receiver&#x27;s side.</li>
</ol>



<p>Prior to Andromeda v1.9.6, this two-step process required a confirmation block, adding latency. The May 2025 upgrade removed that confirmation block. Cross-shard transactions now resolve faster, though they still require two rounds rather than one.</p>



<p>For developers, this means contracts that call across shards must account for the asynchronous nature of that second step. MultiversX&#x27;s tooling exposes async callbacks specifically for this pattern.</p>



<h2 class="wp-block-heading tablecontent">Why Do Projects Use Adaptive State Sharding?</h2>



<p>Projects choose adaptive state sharding for three concrete reasons.</p>



<p>First, throughput. Processing transactions in parallel across shards removes the single-chain bottleneck. A network with three execution shards can theoretically process three times the transactions per second of an equivalent single-shard design, assuming balanced load.</p>



<p>Second, lower fees. Higher throughput means less congestion. Less congestion means users pay less to get their transactions included.</p>



<p>Third, security without sacrificing scale. Validator rotation across epochs means an attacker cannot quietly accumulate influence in one shard over time. The target keeps moving.</p>



<p><a href="https://files.multiversx.com/MultiversX-State-of-the-Foundation-Report-2025.pdf" data-type="link" data-id="https://files.multiversx.com/MultiversX-State-of-the-Foundation-Report-2025.pdf">MultiversX&#x27;s 2025 State of the Foundation Report</a> documents the governance and validator-count changes tied to these security and performance goals.</p>



<h2 class="wp-block-heading tablecontent">MultiversX in Practice: What the Numbers Show</h2>



<p>MultiversX is the most documented live deployment of adaptive state sharding. Its upgrade history gives the clearest picture of what the architecture delivers in production.</p>



<p>Andromeda v1.9.6, released May 21, 2025 (MIP-25), made two structural changes. It removed confirmation blocks from the finality process, achieving what the protocol describes as instant finality. It also standardized validator consensus at 400 validators per shard.</p>



<p>The 2025 State of the Foundation Report notes finality improvements in the range of 12 seconds down to 6 seconds in earlier upgrade cycles, with Andromeda pushing further toward near-instant confirmation.</p>



<p>For a concrete sense of scale: a single-shard chain processes one queue of transactions. A three-shard adaptive system processes three queues simultaneously. If each shard handles 2,000 transactions per second, the combined throughput is 6,000 TPS. Adding a fourth shard scales that to 8,000 TPS without changing the base protocol logic.</p>



<p>These figures are theoretical maximums under balanced load. Real-world throughput depends on cross-shard traffic volume and validator latency.</p>



<h2 class="wp-block-heading tablecontent">What Are the Downsides?</h2>



<p>Adaptive state sharding solves real problems, but it introduces three that single-chain designs avoid.</p>



<p><strong>Cross-shard latency.</strong> Transactions touching two shards take two rounds instead of one. This is manageable for most use cases, but smart contracts that rely on synchronous cross-shard calls require redesign.</p>



<p><strong>State migration cost.</strong> When the protocol splits or merges shards to match demand, it must move account state between nodes. That migration adds overhead during resharding events and can briefly increase latency.</p>



<p><strong>Smaller per-shard security budget.</strong> Each shard has fewer validators than the full network. If a shard&#x27;s validator set is too small, the cost of a 33% Byzantine attack drops. MultiversX addresses this with epoch-based rotation and a minimum validator threshold per shard, but it remains the key security tradeoff to understand before building on any sharded L1.</p>



<h2 class="wp-block-heading tablecontent">The Future of Adaptive State Sharding</h2>



<p>The near-term direction for adaptive state sharding is clearer than most &#x27;future of blockchain&#x27; sections admit.</p>



<p>Governance-linked shard economics is the active frontier. MultiversX&#x27;s 2025 foundation report details how validator reward multipliers and shard counts are now governance-adjustable parameters. That means the community can vote to add shards or change validator incentives without a hard fork.</p>



<p>Cross-shard developer tooling is the other active area. The Andromeda upgrade specifically cited improvements to developer ergonomics around async cross-shard execution. As those tools mature, building applications that span multiple shards becomes less specialized work.</p>



<p>AI-driven shard rebalancing is discussed in research contexts but has no cited live deployment as of mid-2025. Treat it as a direction, not a near-term feature.</p>



<h2 class="wp-block-heading">Frequently Asked Questions</h2>



<h5 class="wp-block-heading">What makes Adaptive State Sharding different from regular sharding? </h5>



<p>Adaptive State Sharding dynamically adjusts the number of shards based on network activity, whereas traditional sharding has a fixed number of shards.</p>



<h5 class="wp-block-heading">Which blockchains use Adaptive State Sharding?</h5>



<p>MultiversX (formerly Elrond) is the primary live implementation. The network has run adaptive state sharding in production since its mainnet launch and has continued upgrading the architecture, most recently with Andromeda v1.9.6 in May 2025. No other major L1 implements all three sharding dimensions (state, transaction, network) in the same integrated design, though Near Protocol uses a related but distinct approach called Nightshade sharding.</p>



<h5 class="wp-block-heading">Does Adaptive State Sharding improve security? </h5>



<p>Yes, by regularly reassigning validators and dynamically adjusting shard structures, it reduces the risk of targeted attacks.</p>



<h5 class="wp-block-heading">Can Adaptive State Sharding reduce transaction fees? </h5>



<p>Yes, by increasing network efficiency and reducing congestion, users typically experience lower transaction costs.</p>



<h5 class="wp-block-heading">Is Adaptive State Sharding suitable for all blockchains? </h5>



<p>There’s no one-size-fits-all solution in blockchain. While Adaptive State Sharding works well for high-throughput networks, some blockchains simply aren’t suitable or don’t require that level of sophisticated network design.</p>

Ethereum L1 Fees Adoption Strategy: Why Cheap Blockspace Is the Point

Why Ethereum Still Leads Every Major Blockchain on User Retention

Decentralized Exchange (DEX)

Blockchain Consensus Mechanisms: How Each One Works

What Is Adaptive State Sharding?

Sharding splits a blockchain into parallel lanes for higher throughput. Learn how adaptive, fixed, and Nightshade sharding work and how each compares to rollups and DAG chains.