66 posts tagged with "Scalability"

Blast's total value locked collapsed 97%—from $2.2 billion to $67 million. Kinto shut down entirely. Loopring closed its wallet service. And that's just the beginning. As 2026 unfolds, Ethereum's Layer 2 ecosystem is witnessing a mass extinction event that's reshaping the entire blockchain scaling landscape.

While more than 50 Layer 2 networks compete for attention, 21Shares' latest State of Crypto report delivers a sobering verdict: most won't survive past 2026. Three networks—Base, Arbitrum, and Optimism—now process nearly 90% of all L2 transactions, with Base alone commanding over 60% market share. The rest? They're becoming "zombie chains," ghost networks with usage down 61% since mid-2025, drained of liquidity, users, and any meaningful future.

The Three Horsemen of L2 Dominance​

The consolidation numbers tell a stark story. Base captured 62% of total L2 revenue year-to-date in 2025, generating $75.4 million of the ecosystem's $120.7 million. Arbitrum and Optimism follow, but the gap is widening rather than closing.

What separates the winners from the walking dead?

Distribution advantage: Base's primary weapon is direct access to Coinbase's 9.3 million monthly active users—a built-in distribution channel that no other L2 can replicate. When Coinbase users applied for $866.3 million in loans through Morpho, 90% of that activity happened on Base. Morpho's TVL on Base exploded 1,906% year-to-date, from $48.2 million to $966.4 million.

Transaction volume: Base handled nearly 40 million transactions in the last 30 days. Compare that to Arbitrum's 6.21 million and Polygon's 29.3 million. Base boasts 15 million unique active wallets versus Arbitrum's 1.12 million and Polygon's 3.69 million.

Profitability: Here's the killer metric—Base was the only L2 that turned a profit in 2025, earning approximately $55 million. Every other rollup operated at a loss after Ethereum's Dencun upgrade slashed data fees by 90%, triggering aggressive fee wars that most networks couldn't win.

The Dencun Aftermath: When Lower Fees Became a Death Sentence​

Ethereum's Dencun upgrade was supposed to be a gift to Layer 2 networks. By reducing data posting costs by roughly 90%, it would make rollups cheaper to operate and more attractive to users. Instead, it triggered a race to the bottom that exposed the fundamental weakness of undifferentiated L2s.

When everyone can offer cheap transactions, nobody has pricing power. The result was a fee war that pushed most rollups into loss-making territory. Without a unique value proposition—whether that's a built-in user base like Base, a mature DeFi ecosystem like Arbitrum, or a network of enterprise chains like Optimism's Superchain—there's no sustainable path forward.

The economic reality is brutal: competitive pressure intensified to the point where only networks with massive scale or strategic backing can survive. That leaves dozens of L2s running on fumes, hoping for a turnaround that likely isn't coming.

Anatomy of a Zombie Chain: The Blast Case Study​

Blast's trajectory offers a masterclass in how quickly an L2 can go from hype to hospice. At its peak, Blast commanded $2.2 billion in TVL and 77,000 daily active users. Today? TVL sits at $55-67 million—a 97% collapse—with just 3,500 daily active users.

The warning signs were there for anyone watching:

Airdrop-driven growth: Like many L2s, Blast's initial traction came from points-fueled speculation rather than organic demand. Users piled in to farm the airdrop, then fled the moment tokens hit wallets.

Disappointing token launch: The BLAST token airdrop failed to retain users, triggering an immediate exodus to rivals like Base and Arbitrum with established ecosystems and deeper liquidity.

Developer abandonment: The official Blast account on X has been inactive since May 2025. The founder's page shows no posts in months. When core teams go silent, the community follows.

Protocol retreat: Even major DeFi protocols like Aave and Synthetix scaled back their Blast deployments, citing poor liquidity and limited returns. When blue-chip DeFi abandons your network, retail isn't far behind.

Blast isn't alone. Many emerging L2s have followed similar trajectories: heavy, incentive-driven activity ahead of a token generation event, a points-fueled surge in usage, then rapid post-TGE decline as liquidity and users migrate elsewhere.

The Rise of Enterprise Rollups​

While zombie chains wither, 2025 marked the rise of a new category: the enterprise rollup. Major institutions began launching or adopting L2 infrastructure, often standardizing on the OP Stack framework:

• Kraken's Ink: The exchange launched its own L2, recently announcing the Ink Foundation and plans for an INK token to power a liquidity protocol built with Aave.
• Uniswap's UniChain: The dominant DEX now has its own chain, capturing value that previously leaked to other networks.
• Sony's Soneium: Targeting gaming and media distribution, Sony's L2 represents traditional entertainment's blockchain ambitions.
• Robinhood's Arbitrum integration: The trading platform uses Arbitrum for quasi-L2 settlement rails for brokerage clients.

These networks bring something most indie L2s lack: captive user bases, brand recognition, and the resources to sustain operations through lean periods. The Optimism Superchain now comprises 34 OP Chains live on mainnet, with Base and OP Mainnet as the most active, followed by World, Soneium, Unichain, Ink, BOB, and Celo.

The consolidation around OP Stack isn't just technical preference—it's economic survival. Shared security, interoperability, and network effects make going alone increasingly untenable.

What Survives the Extinction?​

21Shares expects a "leaner, more resilient" set of networks to define Ethereum's scaling layer by end of 2026. The firm sees the landscape coalescing around three pillars:

1. Ethereum-aligned designs: Networks like Linea route value back to the main chain, aligning their success with Ethereum's ecosystem health rather than competing with it.

2. High-performance contenders: MegaETH and similar projects target near real-time execution, differentiating through speed rather than price. When everyone's cheap, being fast becomes the moat.

3. Exchange-backed networks: Base, BNB Chain, Mantle, and Ink leverage their parent exchanges' user bases and capital reserves to weather market downturns that would kill independent chains.

The DeFi TVL hierarchy reinforces this prediction. Base (46.58%) and Arbitrum (30.86%) dominate Layer 2 DeFi, with total value secured showing a similar concentration—together representing over 75% of the category.

The 2026 Roadmaps: Survivors Building for the Future​

The winning L2s aren't resting on their dominance. Their 2026 roadmaps reveal aggressive expansion plans:

Base: Coinbase's L2 is pivoting toward the creator economy via the "Base App"—a super app integrating messaging, wallet, and mini-apps. The potential total market size approaches $500 billion. Base is also exploring token issuance, though specifics on allocation, utility, and launch date remain unannounced.

Arbitrum: The $215M Gaming Catalyst Program deploys capital through 2026 to fund game studios and infrastructure, targeting SDKs for Unity/Unreal Engine integration. First funded titles launch Q3 2026. The ArbOS Dia Upgrade (Q1 2026) enhances fee predictability and throughput, while Orbit Ecosystem Expansion enables custom chain deployments across industries.

Optimism: The foundation announced plans to dedicate 50% of incoming Superchain revenue to monthly OP token buybacks starting February 2026—a move that transforms OP from pure governance token to one directly aligned with ecosystem growth. The Interop Layer Launch in early 2026 enables cross-chain messaging and shared security across Superchain networks.

The Implications for Builders and Users​

If you're building on a smaller L2, the writing is on the wall. The 61% usage decline across weaker networks since June 2025 isn't a temporary setback—it's the new normal. Smart teams are already migrating to networks with sustainable economics and proven traction.

For users, the consolidation actually brings benefits:

• Deeper liquidity: Concentrated activity means better trading conditions, tighter spreads, and more efficient markets.
• Better tooling: Developer resources naturally flow to dominant platforms, meaning superior wallet support, analytics, and application ecosystems.
• Network effects: The more users and applications concentrate on winning L2s, the more valuable those networks become.

The tradeoff is reduced decentralization and increased dependence on a handful of players. Base's dominance, in particular, raises questions about whether the L2 ecosystem is simply recreating Web2's platform concentration under a blockchain wrapper.

The Bottom Line​

Ethereum's Layer 2 landscape is entering its final form—not the diverse, competitive ecosystem many hoped for, but a tight oligopoly where three networks control nearly everything that matters. The zombie chains will linger for years, running on minimal activity while their teams pivot to other projects or slowly wind down.

For the winners, 2026 represents an opportunity to cement dominance and expand into adjacent markets. For everyone else, the question isn't whether to compete with Base, Arbitrum, and Optimism—it's how to coexist in a world they dominate.

The L2 extinction event isn't coming. It's already here.

---

Building on Ethereum L2s requires reliable infrastructure that scales with your success. BlockEden.xyz provides enterprise-grade RPC endpoints for leading Layer 2 networks including Arbitrum, Optimism, and Base. Explore our API marketplace to power your applications on the platforms that matter.

What if the bottleneck holding back blockchain wasn't software at all, but hardware? That's the premise behind Solayer's audacious new infrastructure play: a $35 million ecosystem fund backing applications built on infiniSVM, the first blockchain to leverage RDMA and InfiniBand networking technology borrowed from supercomputers and high-frequency trading floors.

The announcement, made on January 20, 2026, marks a pivotal moment in the ongoing race for blockchain scalability. While competitors inch toward 10,000 TPS with clever software optimizations, Solayer claims to have already achieved 330,000 TPS with sub-400ms finality on mainnet alpha, with a theoretical ceiling of one million transactions per second.

But raw speed alone doesn't build ecosystems. The real question is whether Solayer can attract the developers and use cases that make such extreme performance necessary.

The Hardware Revolution: RDMA and InfiniBand in Blockchain​

Traditional blockchains are constrained by networking protocols designed for general-purpose computing. TCP/IP stacks, operating system overhead, and CPU-mediated data transfers create latency that compounds across distributed networks. InfiniSVM takes a different approach entirely.

At its core, infiniSVM employs Remote Direct Memory Access (RDMA) technology, which allows nodes to read and write directly to each other's memory without involving the CPU or operating system kernel. Combined with InfiniBand networking, which is the backbone of the world's fastest supercomputers, infiniSVM achieves what Solayer calls "zero-copy data movement."

The technical architecture involves multiple execution clusters connected via Software-Defined Networking (SDN), enabling horizontal scaling that maintains atomic state consistency. This is the same infrastructure powering high-frequency trading operations, where microseconds determine profit or loss.

The numbers are staggering: 100+ Gbps network throughput, sub-50ms devnet finality (approximately 400ms on mainnet alpha), and sustained throughput of 300,000+ TPS. For context, Solana mainnet processes around 4,000 TPS under normal conditions, and Visa handles approximately 24,000 TPS globally.

The $35 Million Ecosystem Play​

Capital allocation tells you where smart money sees opportunity. Solayer's ecosystem fund, backed by Solayer Labs and the Solayer Foundation, is explicitly targeting four verticals:

DeFi Applications: High-frequency trading, perpetual exchanges, and market-making operations that have historically been impossible on-chain due to latency constraints. The fund is backing projects like DoxX, a hardware-accelerated MetaDEX featuring dual-engine architecture designed for institutional-grade, deterministic trade execution.

AI-Driven Systems: Perhaps most intriguingly, Solayer is investing in autonomous AI agents that execute blockchain transactions in real-time. Through their Accel accelerator program, they're backing buff.trade, a platform where AI agents execute tokenized trading strategies. The real-world performance of each agent directly influences the value of its associated token, creating a tight feedback loop between execution quality and on-chain economics.

Tokenized Real-World Assets: Spout Finance is building infrastructure for tokenizing traditional financial assets like U.S. Treasuries on infiniSVM. The combination of high throughput and fast finality makes on-chain treasury operations practical for institutional use cases.

Payments Infrastructure: The fund is positioning infiniSVM as backbone infrastructure for real-time payment processing, where the difference between 400ms and 12-second finality determines whether blockchain can compete with traditional payment rails.

Why Solana Compatibility Matters​

InfiniSVM maintains full compatibility with the Solana Virtual Machine, meaning existing Solana applications can deploy with minimal modification. This is a calculated strategic decision. Rather than building an ecosystem from scratch, Solayer is betting that performance-hungry Solana developers will migrate to infrastructure that removes their current bottlenecks.

The SVM itself is fundamentally different from the Ethereum Virtual Machine. While EVM processes transactions sequentially, SVM was designed around parallel execution using a runtime called Sealevel. Smart contracts on SVM declare their state dependencies upfront, allowing the system to identify which transactions can execute simultaneously across CPU cores.

InfiniSVM takes this parallelism to its logical extreme. By offloading network coordination to specialized hardware and eliminating traditional Ethernet-based node communication, Solayer removes constraints that limit even Solana's native performance.

The LAYER token uses SOL for gas, further reducing friction for Solana developers considering the platform.

The Institutional Finance Angle​

Solayer's timing coincides with a broader shift in institutional blockchain requirements. Traditional finance operates on millisecond timescales. When JPMorgan's Canton Network processes securities settlements, or when BlackRock's BUIDL fund manages tokenized treasuries, latency directly impacts the viability of blockchain integration.

The 300,000 TPS mainnet milestone, achieved in December 2025, represents the first sustained performance at this level on a public network. For institutional use cases requiring deterministic execution, this is table stakes rather than a nice-to-have feature.

The fund's focus on revenue-generating applications over speculative token projects reflects a maturing approach to ecosystem development. Projects must demonstrate clear business models and "strong fundamentals" to receive backing. This is a notable departure from the 2021-era playbook of subsidizing user acquisition through token emissions.

The Competitive Landscape​

Solayer isn't operating in a vacuum. The broader SVM ecosystem includes Eclipse (SVM on Ethereum), Nitro (Cosmos-based SVM), and Solana's own Firedancer validator client from Jump Crypto, which promises significant performance improvements.

Ethereum's roadmap toward parallel execution through sharding and danksharding represents a different philosophical approach: achieving scale through many chains rather than one extremely fast chain.

Meanwhile, chains like Monad and Sei are pursuing their own high-performance EVM strategies, betting that Ethereum compatibility outweighs the technical advantages of SVM.

Solayer's differentiation lies in hardware acceleration. While competitors optimize software, Solayer is optimizing the physical layer. This approach has precedent in traditional finance, where co-location services and FPGA-based trading systems provide edges measured in microseconds.

The risk is that hardware acceleration requires specialized infrastructure that limits decentralization. Solayer's documentation acknowledges this tradeoff, positioning infiniSVM for use cases where performance requirements outweigh maximal decentralization.

What This Means for Blockchain Development​

The $35 million fund signals a hypothesis about where blockchain infrastructure is heading: toward specialized, high-performance networks optimized for specific use cases rather than general-purpose chains trying to serve everyone.

For developers building applications that require real-time execution, whether high-frequency trading, AI agent coordination, or institutional settlement, infiniSVM represents a new category of infrastructure. The SVM compatibility layer reduces migration costs while hardware acceleration unlocks previously impossible application architectures.

For the broader ecosystem, Solayer's success or failure will inform debates about the scalability trilemma. Can hardware-accelerated infrastructure maintain sufficient decentralization while achieving throughput that matches centralized alternatives? The market will ultimately decide.

Looking Ahead​

Solayer's Q1 2026 mainnet launch represents the next major milestone. The transition from mainnet alpha to full production will test whether the 330,000 TPS figures hold under real-world load conditions with diverse application workloads.

The projects emerging from Solayer Accel, particularly the AI agent trading platforms and tokenized treasury infrastructure, will serve as proof points for whether extreme performance translates into genuine product-market fit.

With $35 million in ecosystem capital deployed, Solayer is making one of the more interesting bets in the 2026 infrastructure wars: that the future of blockchain scaling isn't in software optimization alone, but in rethinking the hardware layer entirely.

---

BlockEden.xyz provides high-performance RPC and API infrastructure for SVM-compatible blockchains including Solana. As the ecosystem expands to high-throughput networks like infiniSVM, our infrastructure scales alongside developer needs. Explore our API marketplace for enterprise-grade blockchain connectivity.

---

Sources​

• Solayer and Hack VC Unveil the Future of Real-Time Blockchains with InfiniSVM
• Solayer Launches $35 Million Ecosystem Fund
• Solayer unveils $35M fund for real-time DeFi, AI and tokenization apps - CoinDesk
• Solayer Mainnet Alpha Goes Live - Benzinga
• Solayer Official Documentation
• What is the Solana Virtual Machine (SVM)? - Helius

What if blockchain transactions were as instant as pressing a button in a video game? That's the audacious promise of MegaETH, the Vitalik Buterin-backed Layer 2 that's launching its mainnet and token this January 2026. With claims of 100,000+ transactions per second and 10-millisecond block times—compared to Ethereum's 15 seconds and Base's 1.78 seconds—MegaETH isn't just iterating on existing L2 technology. It's attempting to redefine what "real-time" means for blockchain.

After raising $450 million in its public sale (from $1.39 billion in total bids) and securing backing from Ethereum's co-creator himself, MegaETH has become one of the most anticipated launches of 2026. But can it deliver on promises that sound more like science fiction than blockchain engineering?

Ethereum validators currently process transactions the way a grocery store checkout works with a single lane: one item at a time, in order, no matter how long the line stretches. The Glamsterdam upgrade, scheduled for mid-2026, fundamentally changes this architecture. By introducing Block Access Lists (BAL) and enshrined Proposer-Builder Separation (ePBS), Ethereum is preparing to scale from roughly 21 transactions per second to 10,000 TPS—a 476x improvement that could reshape DeFi, NFTs, and on-chain applications.

When Ethereum L2s paid $3.83 per megabyte to post data using blobs, Eclipse was paying Celestia $0.07 for the same megabyte. That's not a typo—55 times cheaper, enabling Eclipse to post over 83 GB of data without bankrupting its treasury. This cost differential isn't a temporary market anomaly. It's the structural advantage of purpose-built infrastructure.

Celestia has now processed over 160 GB of rollup data, generates daily blob fees that have grown 10x since late 2024, and commands roughly 50% market share in the data availability sector. The question isn't whether modular data availability works—it's whether Celestia can maintain its lead as EigenDA, Avail, and Ethereum's native blobs compete for the same rollup customers.

Understanding Blob Economics: The Foundation​

Before analyzing Celestia's numbers, it's worth understanding what makes data availability economically distinct from other blockchain services.

What Rollups Actually Pay For​

When a rollup processes transactions, it produces state changes that need to be verifiable. Rather than trust the rollup operator, users can verify by re-executing transactions against the original data. This requires that transaction data remains available—not forever, but long enough for challenges and verification.

Traditional rollups posted this data directly to Ethereum calldata, paying premium prices for permanent storage on the world's most secure ledger. But most rollup data only needs availability for a challenge window (typically 7-14 days), not eternity. This mismatch created the opening for specialized data availability layers.

Celestia's PayForBlob Model​

Celestia's fee model is straightforward: rollups pay per blob based on size and current gas prices. Unlike execution layers where computation costs dominate, data availability is fundamentally about bandwidth and storage—resources that scale more predictably with hardware improvements.

The economics create a flywheel: lower DA costs enable more rollups, more rollups generate more fee revenue, and increased usage justifies infrastructure investment for even greater capacity. Celestia's current throughput of approximately 1.33 MB/s (8 MB blocks every 6 seconds) represents early-stage capacity with a clear path to 100x improvement.

The 160 GB Reality: Who's Using Celestia​

The aggregate numbers tell a story of rapid adoption. Over 160 GB of data has been published to Celestia since mainnet launch, with daily data volume averaging around 2.5 GB. But the composition of this data reveals more interesting patterns.

Eclipse: The Volume Leader​

Eclipse—a Layer 2 combining Solana's virtual machine with Ethereum settlement—has published over 83 GB of data to Celestia, more than half of all network volume. Eclipse uses Celestia for data availability while settling to Ethereum, demonstrating the modular architecture in practice.

The volume isn't surprising given Eclipse's design choices. Solana Virtual Machine execution generates more data than EVM equivalents, and Eclipse's focus on high-throughput applications (gaming, DeFi, social) means transaction volumes that would be cost-prohibitive on Ethereum DA.

The Enterprise Cohort​

Beyond Eclipse, the rollup ecosystem includes:

• Manta Pacific: Over 7 GB posted, an OP Stack rollup focused on ZK applications with Universal Circuits technology
• Plume Network: RWA-specialized L2 using Celestia for tokenized asset transaction data
• Derive: On-chain options and structured products trading
• Aevo: Decentralized derivatives exchange processing high-frequency trading data
• Orderly Network: Cross-chain orderbook infrastructure

Twenty-six rollups now build on Celestia, with major frameworks—Arbitrum Orbit, OP Stack, Polygon CDK—all offering Celestia as a DA option. Rollups-as-a-Service platforms like Conduit and Caldera have made Celestia integration a standard offering.

Fee Revenue Growth​

At the end of 2024, Celestia generated approximately $225 per day in blob fees. That number has grown nearly 10x, reflecting both increased usage and the network's ability to capture value as demand rises. The fee market remains early-stage—capacity utilization is low relative to tested limits—but the growth trajectory validates the economic model.

Cost Comparison: Celestia vs. The Competition​

Data availability has become a competitive market. Understanding the cost structures helps explain rollup decisions.

Celestia vs. Ethereum Blobs​

Ethereum's EIP-4844 (Dencun upgrade) introduced blob transactions, reducing DA costs by 90%+ compared to calldata. But Celestia remains significantly cheaper:

Metric	Ethereum Blobs	Celestia
Cost per MB	~$3.83	~$0.07
Cost advantage	Baseline	55x cheaper
Capacity	Limited blob space	8 MB blocks (scaling to 1 GB)

For high-volume rollups like Eclipse, this difference is existential. At Ethereum blob prices, Eclipse's 83 GB of data would have cost over $300,000. On Celestia, it cost approximately $6,000.

Celestia vs. EigenDA​

EigenDA offers a different value proposition: Ethereum-aligned security through restaking, with claimed throughput of 100 MB/s. The tradeoffs:

Aspect	Celestia	EigenDA
Security model	Independent validator set	Ethereum restaking
Throughput	1.33 MB/s (8 MB blocks)	100 MB/s claimed
Architecture	Blockchain-based	Data Availability Committee
Decentralization	Public verification	Trust assumptions

EigenDA's DAC architecture enables higher throughput but introduces trust assumptions that fully blockchain-based solutions avoid. For teams deeply embedded in Ethereum's ecosystem, EigenDA's restaking integration may outweigh Celestia's independence.

Celestia vs. Avail​

Avail positions as the most flexible option for multichain applications:

Aspect	Celestia	Avail
Cost per MB	Higher	Lower
Economic security	Higher	Lower
Mainnet capacity	8 MB blocks	4 MB blocks
Test capacity	128 MB proven	128 MB proven

Avail's lower costs come with lower economic security—a reasonable tradeoff for applications where the marginal cost savings matter more than maximum security guarantees.

The Scaling Roadmap: From 1 MB/s to 1 GB/s​

Celestia's current capacity—approximately 1.33 MB/s—is intentionally conservative. The network has demonstrated dramatically higher throughput in controlled testing, providing a clear upgrade path.

Mammoth Testing Results​

In October 2024, the Mammoth Mini devnet achieved 88 MB blocks with 3-second block times, delivering approximately 27 MB/s throughput—over 20x current mainnet capacity.

In April 2025, the mamo-1 testnet pushed further: 128 MB blocks with 6-second block times, achieving 21.33 MB/s sustained throughput. This represented 16x current mainnet capacity while incorporating new propagation algorithms like Vacuum! designed for efficient large-block data movement.

Mainnet Upgrade Progress​

The scaling is happening incrementally:

• Ginger Upgrade (December 2024): Reduced block times from 12 seconds to 6 seconds
• 8 MB Block Increase (January 2025): Doubled block size via on-chain governance
• Matcha Upgrade (January 2026): Enabled 128 MB blocks through improved propagation mechanics, reducing node storage requirements by 77%
• Lotus Upgrade (July 2025): V4 mainnet release with further TIA holder improvements

The roadmap targets gigabyte-scale blocks by 2030, representing a 1,000x increase from current capacity. Whether market demand grows to justify this capacity remains uncertain, but the technical path is clear.

TIA Tokenomics: How Value Accrues​

Understanding Celestia's economics requires understanding TIA's role in the system.

Token Utility​

TIA serves three functions:

1. Blob fees: Rollups pay TIA for data availability
2. Staking: Validators stake TIA to secure the network and earn rewards
3. Governance: Token holders vote on network parameters and upgrades

The fee mechanism creates direct linkage between network usage and token demand. As blob submissions increase, TIA is purchased and spent, creating buy pressure proportional to network utility.

Supply Dynamics​

TIA launched with 1 billion genesis tokens. Initial inflation was set at 8% annually, decreasing over time toward 1.5% terminal inflation.

The January 2026 Matcha upgrade introduced Proof-of-Governance (PoG), slashing annual token issuance from 5% to 0.25%. This structural change:

• Reduces sell pressure from inflation
• Aligns rewards with governance participation
• Strengthens value capture as network usage grows

Additionally, the Celestia Foundation announced a $62.5 million TIA buyback program in 2025, further reducing circulating supply.

Validator Economics​

Effective January 2026, maximum validator commission increased from 10% to 20%. This addresses validators' rising operational expenses—particularly as block sizes grow—while maintaining competitive staking yields.

The Competitive Moat: First-Mover or Sustainable Advantage?​

Celestia's 50% DA market share and 160+ GB of posted data represent clear traction. But moats in infrastructure can erode quickly.

Advantages​

Framework Integration: Every major rollup framework—Arbitrum Orbit, OP Stack, Polygon CDK—supports Celestia as a DA option. This integration creates switching costs and reduces friction for new rollups.

Proven Scale: The 128 MB block testing provides confidence in future capacity that competitors haven't demonstrated at the same level.

Economic Alignment: The Proof-of-Governance tokenomics and buyback programs create stronger value capture than alternative models.

Challenges​

EigenDA's Ethereum Alignment: For teams prioritizing Ethereum-native security, EigenDA's restaking model may be more attractive despite architectural trade-offs.

Avail's Cost Advantage: For cost-sensitive applications, Avail's lower fees may outweigh security differences.

Ethereum's Native Improvement: If Ethereum expands blob capacity significantly (as proposed in various roadmap discussions), the cost differential shrinks.

The Ecosystem Lock-in Question​

Celestia's real moat may be ecosystem lock-in. Eclipse's 83+ GB of data creates path dependency—migrating to a different DA layer would require significant infrastructure changes. As more rollups accumulate history on Celestia, switching costs increase.

What the Data Tells Us​

Celestia's blob economics validate the modular thesis: specialized infrastructure for data availability can be dramatically cheaper than general-purpose L1 solutions. The 55x cost advantage over Ethereum blobs isn't magic—it's the result of purpose-built architecture optimized for a specific function.

The 160+ GB of posted data proves market demand exists. The 10x growth in fee revenue demonstrates value capture. The scaling roadmap provides confidence in future capacity.

For rollup developers, the calculus is straightforward: Celestia offers the best-tested, most integrated DA solution with a clear path to gigabyte-scale capacity. EigenDA makes sense for Ethereum-native projects willing to accept DAC trust assumptions. Avail serves multichain applications prioritizing flexibility over maximum security.

The data availability market has room for multiple winners serving different segments. But Celestia's combination of proven scale, deep integrations, and improving tokenomics positions it well for the coming wave of rollup expansion.

---

Building rollups that need reliable data availability infrastructure? BlockEden.xyz provides RPC endpoints across 30+ networks including major L2s built on Celestia DA. Explore our API marketplace to access the infrastructure your modular stack needs.

What if the blockchain performance debates of 2021-2023 already feel ancient? In 2025, the industry quietly crossed a threshold that venture capitalists and skeptics alike thought was years away: multiple mainnets now routinely process thousands of transactions per second while keeping fees below a single cent. The era of "blockchain can't scale" has officially ended.

This isn't about theoretical benchmarks or testnet claims. Real users, real applications, and real money are flowing through networks that would have been science fiction just two years ago. Let's examine the hard numbers behind blockchain's performance revolution.

The New TPS Leaders: No Longer a Two-Horse Race​

The performance landscape has fundamentally shifted. While Bitcoin and Ethereum dominated blockchain conversations for years, 2025 established a new generation of speed champions.

Solana set the headline-grabbing record on August 17, 2025, processing 107,664 transactions per second on its mainnet—not in a laboratory, but under real-world conditions. This wasn't a one-off spike; the network demonstrated sustained high throughput that validates years of architectural decisions prioritizing performance.

But Solana's achievement is just one data point in a broader revolution:

• Aptos has demonstrated 13,367 TPS on mainnet without failures, delays, or gas fee spikes. Their Block-STM parallel execution engine theoretically supports up to 160,000 TPS.
• Sui has proven 297,000 TPS in controlled testing, with mainnet peaks reaching 822 TPS under typical usage and the Mysticeti v2 consensus achieving just 390ms latency.
• BNB Chain consistently delivers around 2,200 TPS in production, with the Lorentz and Maxwell hard forks delivering 4x faster block times.
• Avalanche processes 4,500 TPS through its unique subnet architecture, enabling horizontal scaling across specialized chains.

These numbers represent a 10x to 100x improvement over what the same networks achieved in 2023. More importantly, they're not theoretical maximums—they're observed, verifiable performance under actual usage conditions.

Firedancer: The Million-TPS Client That Changed Everything​

The most significant technical breakthrough of 2025 wasn't a new blockchain—it was Firedancer, Jump Crypto's complete reimplementation of the Solana validator client. After three years of development, Firedancer went live on mainnet on December 12, 2025.

The numbers are staggering. In demonstrations at Breakpoint 2024, Jump's Chief Scientist Kevin Bowers showed Firedancer processing over 1 million transactions per second on commodity hardware. Benchmarks consistently showed 600,000 to 1,000,000 TPS in controlled tests—20x higher than the previous Agave client's demonstrated throughput.

What makes Firedancer different? Architecture. Unlike Agave's monolithic design, Firedancer uses a modular, tile-based architecture that splits validator tasks to run in parallel. Written in C rather than Rust, every component was optimized for raw performance from the ground up.

The adoption trajectory tells its own story. Frankendancer, a hybrid implementation combining Firedancer's networking stack with Agave's runtime, now runs on 207 validators representing 20.9% of all staked SOL—up from just 8% in June 2025. This isn't experimental software anymore; it's infrastructure that secures billions of dollars.

Solana's Alpenglow upgrade in September 2025 added another layer, replacing the original Proof of History and TowerBFT mechanisms with new Votor and Rotor systems. The result: 150ms block finality and support for multiple concurrent leaders enabling parallel execution.

Sub-Penny Fees: EIP-4844's Quiet Revolution​

While TPS numbers grab headlines, the fee revolution is equally transformative. Ethereum's EIP-4844 upgrade in March 2024 fundamentally restructured how Layer 2 networks pay for data availability, and by 2025, the effects became impossible to ignore.

The mechanism is elegant: blob transactions provide temporary data storage for rollups at a fraction of previous costs. Where Layer 2s previously competed for expensive calldata space, blobs offer 18-day temporary storage that rollups actually need.

The impact on fees was immediate and dramatic:

• Arbitrum fees dropped from $0.37 to $0.012 per transaction
• Optimism fell from $0.32 to $0.009
• Base achieved fees as low as $0.01

These aren't promotional rates or subsidized transactions—they're sustainable operating costs enabled by architectural improvement. Ethereum now effectively provides 10-100x cheaper data storage for Layer 2 solutions.

The activity surge followed predictably. Base saw a 319.3% increase in daily transactions post-upgrade, Arbitrum increased 45.7%, and Optimism rose 29.8%. Users and developers responded exactly as economics predicted: when transactions become cheap enough, usage explodes.

The May 2025 Pectra upgrade pushed further, expanding blob throughput from 6 to 9 blobs per block and raising the gas limit to 37.3 million. Ethereum's effective TPS through Layer 2s now exceeds 100,000, with average transaction costs dropping to $0.08 on L2 networks.

The Real-World Performance Gap​

Here's what the benchmarks don't tell you: theoretical TPS and observed TPS remain very different numbers. This gap reveals important truths about blockchain maturity.

Consider Avalanche. While the network supports 4,500 TPS theoretically, observed activity averages around 18 TPS, with the C-Chain closer to 3-4 TPS. Sui demonstrates 297,000 TPS in testing but peaks at 822 TPS on mainnet.

This isn't failure—it's proof of headroom. These networks can handle massive demand spikes without degradation. When the next NFT frenzy or DeFi summer arrives, the infrastructure won't buckle.

The practical implications matter enormously for builders:

• Gaming applications need consistent low latency more than peak TPS
• DeFi protocols require predictable fees during volatility
• Payment systems demand reliable throughput during holiday shopping spikes
• Enterprise applications need guaranteed SLAs regardless of network conditions

Networks with significant headroom can offer these guarantees. Those operating near capacity cannot.

Move VM Chains: The Performance Architecture Advantage​

A pattern emerges when examining 2025's top performers: the Move programming language shows up repeatedly. Both Sui and Aptos, built by teams with Facebook/Diem heritage, leverage Move's object-centric data model for parallelization advantages impossible in account-model blockchains.

Aptos's Block-STM engine demonstrates this clearly. By processing transactions simultaneously rather than sequentially, the network achieved 326 million successful transactions in a single day during peak periods—while maintaining approximately $0.002 average fees.

Sui's approach differs but follows similar principles. The Mysticeti consensus protocol achieves 390ms latency by treating objects rather than accounts as the fundamental unit. Transactions that don't touch the same objects execute in parallel automatically.

Both networks attracted significant capital in 2025. BlackRock's BUIDL fund added $500 million in tokenized assets to Aptos in October, making it the second-largest BUIDL chain. Aptos also powered the official digital wallet for Expo 2025 in Osaka, processing 558,000+ transactions and onboarding 133,000+ users—real-world validation at scale.

What High TPS Actually Enables​

Beyond bragging rights, what do thousands of TPS unlock?

Institutional-grade settlement: When processing 2,000+ TPS with sub-second finality, blockchains compete directly with traditional payment rails. BNB Chain's Lorentz and Maxwell upgrades specifically targeted "Nasdaq-scale settlement" for institutional DeFi.

Microtransaction viability: At $0.01 per transaction, business models impossible at $5 fees become profitable. Streaming payments, per-API-call billing, and granular royalty distribution all require sub-penny economics.

Game state synchronization: Blockchain gaming requires updating player states hundreds of times per session. 2025's performance levels finally enable genuine on-chain gaming rather than the settlement-only models of previous years.

IoT and sensor networks: When devices can transact for fractions of a cent, supply chain tracking, environmental monitoring, and machine-to-machine payments become economically viable.

The common thread: 2025's performance improvements didn't just make existing applications faster—they enabled entirely new categories of blockchain usage.

The Decentralization Trade-off Debate​

Critics correctly note that raw TPS often correlates with reduced decentralization. Solana runs fewer validators than Ethereum. Aptos and Sui require more expensive hardware. These trade-offs are real.

But 2025 also demonstrated that the binary choice between speed and decentralization is false. Ethereum's Layer 2 ecosystem delivers 100,000+ effective TPS while inheriting Ethereum's security guarantees. Firedancer improves Solana's throughput without reducing validator counts.

The industry is learning to specialize: settlement layers optimize for security, execution layers optimize for speed, and proper bridging connects them. This modular approach—data availability from Celestia, execution from rollups, settlement on Ethereum—achieves speed, security, and decentralization through composition rather than compromise.

Looking Forward: The Million-TPS Mainnet​

If 2025 established high-TPS mainnets as reality rather than promise, what comes next?

Ethereum's Fusaka upgrade will introduce full danksharding via PeerDAS, potentially enabling millions of TPS across rollups. Firedancer's production deployment should push Solana toward its tested 1 million TPS capacity. New entrants continue emerging with novel architectures.

More importantly, the developer experience has matured. Building applications that require thousands of TPS is no longer a research project—it's standard practice. The tooling, documentation, and infrastructure supporting high-performance blockchain development in 2025 would be unrecognizable to a 2021 developer.

The question is no longer whether blockchain can scale. The question is what we'll build now that it has.

---

BlockEden.xyz provides enterprise-grade RPC and API access for high-performance chains including Sui, Aptos, and Solana. When your application demands the throughput and reliability that 2025's performance revolution enables, explore our infrastructure designed for production-grade blockchain development.

What if you could verify a 500-page book exists without reading a single page? That's essentially what Ethereum just learned to do with PeerDAS—and it's quietly reshaping how blockchains can scale without sacrificing decentralization.

On December 3, 2025, Ethereum activated its Fusaka upgrade, introducing PeerDAS (Peer Data Availability Sampling) as the headline feature. While most headlines focused on the 40-60% fee reductions for Layer 2 networks, the underlying mechanism represents something far more significant: a fundamental shift in how blockchain nodes prove data exists without actually storing all of it.

In April 2025, Vitalik Buterin proposed something that would have seemed heretical a year earlier: replacing Ethereum's EVM with RISC-V. The suggestion sparked immediate debate. But what most commentators missed was that Polkadot had already been building exactly this architecture for over a year—and was months away from deploying it to production.

Polkadot's JAM (Join-Accumulate Machine) isn't just another blockchain upgrade. It represents a fundamental rethinking of what a "blockchain" even means. Where Ethereum's worldview centers on a global virtual machine processing transactions, JAM eliminates the transaction concept entirely at its core layer, replacing it with a computation model that promises 850 MB/s data availability—42 times Polkadot's previous capacity and 650 times Ethereum's 1.3 MB/s.

The implications extend far beyond performance benchmarks. JAM may be the clearest articulation yet of a post-Ethereum paradigm for blockchain architecture.

The Gray Paper: Gavin Wood's Third Act​

Dr. Gavin Wood wrote the Ethereum Yellow Paper in 2014, providing the formal specification that made Ethereum possible. He followed with the Polkadot White Paper in 2016, introducing heterogeneous sharding and shared security. In April 2024, he released the JAM Gray Paper at Token2049 in Dubai—completing a trilogy that spans the entire history of programmable blockchains.

The Gray Paper describes JAM as "a global singleton permissionless object environment—akin to Ethereum's smart-contract environment—paired with secure sideband computation parallelized over a scalable node network." But this undersells the conceptual shift.

JAM doesn't just improve on existing blockchain designs. It asks: what if we stopped thinking about blockchains as virtual machines entirely?

The Transaction Problem​

Traditional blockchains—Ethereum included—are fundamentally transaction-processing systems. Users submit transactions, validators order and execute them, and the blockchain records state changes. This model has served well but carries inherent limitations:

• Sequential bottlenecks: Transactions must be ordered, creating throughput constraints
• Global state contention: Every transaction potentially touches shared state
• Execution coupling: Consensus and computation are tightly bound

JAM decouples these concerns through what Wood calls the "Refine-Accumulate" paradigm. The system operates in two phases:

Refine: Computation happens in parallel across the network. Work is divided into independent units that can execute simultaneously without coordination.

Accumulate: Results are collected and merged into global state. Only this phase requires consensus on ordering.

The result is a "transactionless" core protocol. JAM itself doesn't process transactions—applications built on JAM do. This separation allows the base layer to focus purely on secure, parallel computation.

PolkaVM: Why RISC-V Matters​

At the heart of JAM sits PolkaVM, a purpose-built virtual machine based on the RISC-V instruction set. This choice has profound implications for blockchain computation.

The EVM's Architectural Debt​

Ethereum's EVM was designed in 2013-2014, before many modern assumptions about blockchain execution were understood. Its architecture reflects that era:

• Stack-based execution: Operations push and pop values from an unbounded stack, requiring complex tracking
• 256-bit word size: Chosen for cryptographic convenience but wasteful for most operations
• Single-dimensional gas: One metric attempts to price vastly different computational resources
• Interpretation-only: EVM bytecode cannot be compiled to native code efficiently

These design decisions made sense as initial choices but create ongoing performance penalties.

RISC-V's Advantages​

PolkaVM takes a fundamentally different approach:

Register-based architecture: Like modern CPUs, PolkaVM uses a finite set of registers for argument passing. This aligns with actual hardware, enabling efficient translation to native instruction sets.

64-bit word size: Modern processors are 64-bit. Using a matching word size eliminates the overhead of emulating 256-bit operations for the vast majority of computations.

Multi-dimensional gas: Different resources (computation, storage, bandwidth) are priced independently, better reflecting true costs and preventing mispricing attacks.

Dual execution modes: Code can be interpreted for immediate execution or JIT-compiled for optimized performance. The system chooses the appropriate mode based on workload characteristics.

Performance Impact​

The architectural differences translate to real performance gains. Benchmarks show PolkaVM achieving 10x+ improvements over WebAssembly for arithmetic-intensive contracts—and the EVM is slower still. For complex, multi-contract interactions, the gap widens further as JIT compilation amortizes setup costs.

Perhaps more importantly, PolkaVM supports any language that compiles to RISC-V. While EVM developers are limited to Solidity, Vyper, and a handful of specialized languages, PolkaVM opens the door to Rust, C++, and eventually any LLVM-supported language. This dramatically expands the potential developer pool.

Maintaining Developer Experience​

Despite the architectural overhaul, PolkaVM maintains compatibility with existing workflows. The Revive compiler provides complete Solidity support, including inline assembler. Developers can continue using Hardhat, Remix, and MetaMask without changing their processes.

The Papermoon team demonstrated this compatibility by successfully migrating Uniswap V2's contract code to the PolkaVM testnet—proving that even complex, battle-tested DeFi code can transition without rewrites.

JAM's Performance Targets​

The numbers Wood projects for JAM are staggering by current blockchain standards.

Data Availability​

JAM targets 850 MB/s of data availability—roughly 42 times the vanilla Polkadot capacity before recent optimizations and 650 times Ethereum's 1.3 MB/s. For context, this approaches the throughput of enterprise database systems.

Computational Throughput​

The Gray Paper estimates JAM can achieve approximately 150 billion gas per second at full capacity. Translating gas to transactions is imprecise, but theoretical maximum throughput reaches 3.4+ million TPS based on the data availability target.

Real-World Validation​

These aren't purely theoretical numbers. Stress tests have validated the architecture:

• Kusama (August 2025): Achieved 143,000 TPS at only 23% load capacity
• Polkadot "Spammening" (2024): Reached 623,000 TPS in controlled testing

These figures represent genuine transaction throughput, not optimistic projections or testnet conditions that don't reflect production environments.

Development Status and Timeline​

JAM development follows a structured milestone system, with 43 implementation teams competing for a prize pool exceeding $60 million (10 million DOT + 100,000 KSM).

Current Progress (Late 2025)​

The ecosystem has reached several critical milestones:

• Multiple teams have achieved 100% conformance with Web3 Foundation test vectors
• Development has progressed through Gray Paper versions 0.6.2 through 0.8.0, approaching v1.0
• The JAM Experience conference in Lisbon (May 2025) brought together implementation teams for deep technical collaboration
• University tours reached over 1,300 attendees across nine global locations, including Cambridge, Peking University, and Fudan University

Milestone Structure​

Teams progress through a series of milestones:

1. IMPORTER (M1): Passing state-transitioning conformance tests and importing blocks
2. AUTHORER (M2): Full conformance including block production, networking, and off-chain components
3. HALF-SPEED (M3): Achieving Kusama-level performance, with access to JAM Toaster for full-scale testing
4. FULL-SPEED (M4): Polkadot mainnet-level performance with professional security audits

Multiple teams have completed M1, with several progressing toward M2.

Timeline to Mainnet​

• Late 2025: Final Gray Paper revisions, continued milestone submissions, expanded testnet participation
• Q1 2026: JAM mainnet upgrade on Polkadot following governance approval via OpenGov referendum
• 2026: CoreChain Phase 1 deployment, official public JAM testnet, full network transition

The governance process has already shown strong community support. A near-unanimous DOT holder vote approved the upgrade direction in May 2024.

JAM vs. Ethereum: Complementary or Competitive?​

The question of whether JAM represents an "Ethereum killer" misses the architectural nuance.

Different Design Philosophies​

Ethereum builds outward from a monolithic foundation. The EVM provides a global execution environment, and scaling solutions—L2s, rollups, sharding—layer on top. This approach has created an enormous ecosystem but also accumulated technical debt.

JAM starts with modularity at its core. The separation of Refine and Accumulate phases, the domain-specific optimization for rollup handling, and the transactionless base layer all reflect a ground-up design for scalability.

Convergent Technical Choices​

Despite different starting points, the projects are converging on similar conclusions. Vitalik's April 2025 RISC-V proposal acknowledged that the EVM's architecture limits long-term performance. Polkadot had already deployed RISC-V support to testnet months earlier.

This convergence validates both projects' technical judgment while highlighting the execution gap: Polkadot is shipping what Ethereum is proposing.

Ecosystem Realities​

Technical superiority doesn't automatically translate to ecosystem dominance. Ethereum's developer community, application diversity, and liquidity depth represent substantial network effects that can't be replicated overnight.

The more likely outcome isn't replacement but specialization. JAM's architecture is optimized for certain workloads—particularly high-throughput applications and rollup infrastructure—while Ethereum retains advantages in ecosystem maturity and capital formation.

In 2026, they look less like competitors and more like complementary layers of a multi-chain internet.

What JAM Means for Blockchain Architecture​

JAM's significance extends beyond Polkadot. It represents the clearest articulation of a post-EVM paradigm that other projects will study and selectively adopt.

Key Principles​

Computation separation: Decoupling execution from consensus enables parallel processing at the base layer, not as an afterthought.

Domain-specific optimization: Rather than building a general-purpose VM and hoping it scales, JAM is architected specifically for the workloads blockchains actually run.

Hardware alignment: Using RISC-V and 64-bit words aligns virtual machine architecture with physical hardware, eliminating emulation overhead.

Transaction abstraction: Moving transaction handling to the application layer allows the protocol to focus on computation and state management.

Industry Impact​

Whether JAM succeeds or fails commercially, these architectural choices will influence blockchain design for the next decade. The Gray Paper provides a formal specification that other projects can study, critique, and selectively implement.

Ethereum's RISC-V proposal already demonstrates this influence. The question isn't whether these ideas will spread, but how quickly and in what form.

The Road Ahead​

JAM represents Gavin Wood's most ambitious technical vision since Polkadot itself. The stakes match the ambition: success would validate an entirely different approach to blockchain architecture, while failure would leave Polkadot competing with newer L1s without a differentiated technical narrative.

The next 18 months will determine whether JAM's theoretical advantages translate to production reality. With 43 implementation teams, a nine-figure prize pool, and a clear roadmap to mainnet, the project has resources and momentum. What remains to be seen is whether the complexity of the Refine-Accumulate paradigm can deliver on Wood's vision of a "distributed computer that can run almost any kind of task."

For developers and projects evaluating blockchain infrastructure, JAM merits serious attention—not as hype, but as a technically rigorous attempt to solve problems that every major blockchain faces. The blockchain-as-virtual-machine paradigm served the industry well for a decade. JAM bets that the next decade requires something fundamentally different.

---

Building on next-generation blockchain infrastructure? BlockEden.xyz provides high-performance RPC endpoints across the Polkadot ecosystem and 30+ other networks. Explore our API marketplace to access enterprise-grade infrastructure for your applications.

In 2022, Vitalik Buterin posed a simple question that would define the next four years of Ethereum scaling: how much Ethereum compatibility are you willing to sacrifice for faster zero-knowledge proofs? His answer came in the form of a five-type classification system for zkEVMs that has since become the industry standard for evaluating these critical scaling solutions.

Fast forward to 2026, and the answer isn't so simple anymore. Proving times have collapsed from 16 minutes to 16 seconds. Costs have dropped 45x. Multiple teams have demonstrated real-time proof generation faster than Ethereum's 12-second block times. Yet the fundamental trade-off Vitalik identified remains—and understanding it is essential for any developer or project choosing where to build.

The Vitalik Classification: Types 1 Through 4​

Vitalik's framework categorizes zkEVMs along a spectrum from perfect Ethereum equivalence to maximum proving efficiency. Higher type numbers mean faster proofs but less compatibility with existing Ethereum infrastructure.

Type 1: Fully Ethereum-Equivalent​

Type 1 zkEVMs don't change anything about Ethereum. They prove the exact same execution environment that Ethereum L1 uses—same opcodes, same data structures, same everything.

The upside: Perfect compatibility. Ethereum execution clients work as-is. Every tool, every contract, every piece of infrastructure transfers directly. This is ultimately what Ethereum needs to make L1 itself more scalable.

The downside: Ethereum wasn't designed for zero-knowledge proofs. The EVM's stack-based architecture is notoriously inefficient for ZK proof generation. Early Type 1 implementations required hours to generate a single proof.

Leading project: Taiko aims for Type 1 equivalence as a based rollup using Ethereum's validators for sequencing, enabling synchronous composability with other based rollups.

Type 2: Fully EVM-Equivalent​

Type 2 zkEVMs maintain full EVM compatibility but change internal representations—how state is stored, how data structures are organized—to improve proof generation.

The upside: Contracts written for Ethereum run without modification. The developer experience remains identical. Migration friction approaches zero.

The downside: Block explorers and debugging tools may need modifications. State proofs work differently than on Ethereum L1.

Leading projects: Scroll and Linea target Type 2 compatibility, achieving near-perfect EVM equivalence at the VM level without transpilers or custom compilers.

Type 2.5: EVM-Equivalent with Gas Cost Changes​

Type 2.5 is a pragmatic middle ground. The zkEVM remains EVM-compatible but increases gas costs for operations that are particularly expensive to prove in zero-knowledge.

The trade-off: Since Ethereum has a gas limit per block, increasing gas costs for specific opcodes means fewer of those opcodes can execute per block. Applications work, but certain computational patterns become prohibitively expensive.

Type 3: Almost EVM-Equivalent​

Type 3 zkEVMs sacrifice specific EVM features—often related to precompiles, memory handling, or how contract code is treated—to dramatically improve proof generation.

The upside: Faster proofs, lower costs, better performance.

The downside: Some Ethereum applications won't work without modification. Developers may need to rewrite contracts that rely on unsupported features.

Reality check: No team actually wants to stay at Type 3. It's understood as a transitional stage while teams work on adding the complex precompile support needed to reach Type 2.5 or Type 2. Both Scroll and Polygon zkEVM operated as Type 3 before advancing up the compatibility ladder.

Type 4: High-Level Language Compatible​

Type 4 systems abandon EVM compatibility entirely at the bytecode level. Instead, they compile Solidity or Vyper to a custom VM designed specifically for efficient ZK proofs.

The upside: Fastest proof generation. Lowest costs. Maximum performance.

The downside: Contracts may behave differently. Addresses might not match Ethereum deployments. Debugging tools need complete rewrites. Migration requires careful testing.

Leading projects: zkSync Era and StarkNet represent the Type 4 approach. zkSync transpiles Solidity to custom bytecode optimized for ZK. StarkNet uses Cairo, an entirely new language designed for provability.

Performance Benchmarks: Where We Stand in 2026​

The numbers have transformed dramatically since Vitalik's original post. What was theoretical in 2022 is production reality in 2026.

Proving Times​

Early zkEVMs required approximately 16 minutes to generate proofs. Current implementations complete the same process in roughly 16 seconds—a 60x improvement. Several teams have demonstrated proof generation in under 2 seconds, faster than Ethereum's 12-second block times.

The Ethereum Foundation has set an ambitious target: proving 99% of mainnet blocks in under 10 seconds using less than $100,000 in hardware and 10kW of power consumption. Multiple teams have already demonstrated capability close to this target.

Transaction Costs​

The Dencun upgrade in March 2024 (EIP-4844 introducing "blobs") reduced L2 fees by 75-90%, making all rollups dramatically more cost-effective. Current benchmarks show:

Platform	Transaction Cost	Notes
Polygon zkEVM	$0.00275	Per transaction for full batches
zkSync Era	$0.00378	Median transaction cost
Linea	$0.05-0.15	Average transaction

Throughput​

Real-world performance varies significantly based on transaction complexity:

Platform	TPS (Complex DeFi)	Notes
Polygon zkEVM	5.4 tx/s	AMM swap benchmark
zkSync Era	71 TPS	Complex DeFi swaps
Theoretical (Linea)	100,000 TPS	With advanced sharding

These numbers will continue improving as hardware acceleration, parallelization, and algorithmic optimizations mature.

Market Adoption: TVL and Developer Traction​

The zkEVM landscape has consolidated around several clear leaders, each representing different points on the type spectrum:

Current TVL Rankings (2025)​

• Scroll: $748 million TVL, largest pure zkEVM
• StarkNet: $826 million TVS
• zkSync Era: $569 million TVL, 270+ deployed dApps
• Linea: ~$963 million TVS, 400%+ growth in daily active addresses

The overall Layer 2 ecosystem has reached $70 billion in TVL, with ZK rollups capturing increasing market share as proving costs continue declining.

Developer Adoption Signals​

• Over 65% of new smart contracts in 2025 deployed on Layer 2 networks
• zkSync Era attracted approximately $1.9 billion in tokenized real-world assets, capturing ~25% of on-chain RWA market share
• Layer 2 networks handled an estimated 1.9 million daily transactions in 2025

The Compatibility-Performance Trade-off in Practice​

Understanding the theoretical types is useful, but the practical implications for developers are what matter.

Type 1-2: Zero Migration Friction​

For Scroll and Linea (Type 2), migration means literally zero code changes for most applications. Deploy the same Solidity bytecode, use the same tools (MetaMask, Hardhat, Remix), expect the same behavior.

Best for: Existing Ethereum applications prioritizing seamless migration; projects where proven, audited code must remain unchanged; teams without resources for extensive testing and modification.

Type 3: Careful Testing Required​

For Polygon zkEVM and similar Type 3 implementations, most applications work but edge cases exist. Certain precompiles may behave differently or be unsupported.

Best for: Teams with resources for thorough testnet validation; projects not relying on exotic EVM features; applications prioritizing cost efficiency over perfect compatibility.

Type 4: Different Mental Model​

For zkSync Era and StarkNet, the development experience differs meaningfully from Ethereum:

zkSync Era supports Solidity but transpiles it to custom bytecode. Contracts compile and run, but behavior may differ in subtle ways. Addresses aren't guaranteed to match Ethereum deployments.

StarkNet uses Cairo, requiring developers to learn an entirely new language—though one specifically designed for provable computation.

Best for: Greenfield projects not constrained by existing code; applications prioritizing maximum performance; teams willing to invest in specialized tooling and testing.

Security: The Non-Negotiable Constraint​

The Ethereum Foundation introduced clear cryptographic security requirements for zkEVM developers in 2025:

• 100-bit provable security by May 2026
• 128-bit security by end of 2026

These requirements reflect the reality that faster proofs mean nothing if the underlying cryptography isn't bulletproof. Teams are expected to meet these thresholds regardless of their type classification.

The security focus has slowed some performance improvements—the Ethereum Foundation explicitly chose security over speed through 2026—but ensures the foundation for mainstream adoption remains solid.

Choosing Your zkEVM: A Decision Framework​

Choose Type 1-2 (Taiko, Scroll, Linea) if:​

• You're migrating existing battle-tested contracts
• Audit costs are a concern (no reaudit needed)
• Your team is Ethereum-native without ZK expertise
• Composability with Ethereum L1 matters
• You need synchronous interoperability with other based rollups

Choose Type 3 (Polygon zkEVM) if:​

• You want a balance of compatibility and performance
• You can invest in thorough testnet validation
• Cost efficiency is a priority
• You don't rely on exotic EVM precompiles

Choose Type 4 (zkSync Era, StarkNet) if:​

• You're building from scratch without migration constraints
• Maximum performance justifies tooling investment
• Your use case benefits from ZK-native design patterns
• You have resources for specialized development

What Comes Next​

The type classifications won't remain static. Vitalik noted that zkEVM projects can "easily start at higher-numbered types and jump to lower-numbered types over time." We're seeing this in practice—projects that launched as Type 3 are advancing toward Type 2 as they complete precompile implementations.

More intriguingly, if Ethereum L1 adopts modifications to become more ZK-friendly, Type 2 and Type 3 implementations could become Type 1 without changing their own code.

The endgame appears increasingly clear: proving times will continue compressing, costs will continue declining, and the distinction between types will blur as hardware acceleration and algorithmic improvements close the performance gap. The question isn't which type will win—it's how quickly the entire spectrum converges toward practical equivalence.

For now, the framework remains valuable. Understanding where a zkEVM sits on the compatibility-performance spectrum tells you what to expect during development, deployment, and operation. That knowledge is essential for any team building on Ethereum's ZK-powered future.

---

Building on zkEVM infrastructure? BlockEden.xyz provides high-performance RPC endpoints across multiple zkEVM chains including Polygon zkEVM, Scroll, and Linea. Explore our API marketplace to access the infrastructure layer your ZK applications need.