CBDC speed claims are obsolete. Central banks tout sub-second finality as revolutionary, but this benchmark was surpassed years ago by optimistic and ZK rollups like Arbitrum and zkSync. Their performance is not a future roadmap item; it is live on mainnet today.
history-of-money-and-the-crypto-thesis
Blog
Why Layer 2 Scaling Solutions Make CBDC Speed Claims Obsolete
A technical analysis demonstrating that permissionless L2 networks like Arbitrum and Base have already surpassed the theoretical performance ceilings of centralized CBDC designs, invalidating their core efficiency premise.
introduction
THE SPEED GAP
Introduction
Layer 2 scaling solutions have already achieved transaction speeds that render traditional CBDC performance targets irrelevant.
The bottleneck is institutional design, not technology. A CBDC's latency stems from governance and compliance layers, not technical limits. Permissionless L2s like Base and Starknet process transactions in milliseconds because their stack eliminates centralized intermediaries.
Evidence: Arbitrum Nova handles over 2 million transactions per day with sub-second user-experience finality, a throughput volume and speed no proposed wholesale CBDC architecture currently matches in production.
key-insights
WHY L2S RENDER CBDC SPEED CLAIMS MOOT
Executive Summary
Central banks touting 'instant' CBDC settlements are competing against a benchmark that no longer exists in the real world.
01
The Latency Illusion
CBDCs target ~2-5 second finality, a legacy of batch processing. Modern L2s like Arbitrum and Optimism achieve sub-second pre-confirmations with ~12 second full finality. The user experience benchmark is now milliseconds, not seconds.
~500ms
L2 Pre-Confirm
2-5s
CBDC Target
02
Cost Per Transaction is a Red Herring
CBDC architectures focus on marginal cost reduction. L2s have already solved this with <$0.01 fees, making the cost debate irrelevant. The real competition is in programmability and composability—features CBDCs structurally lack.
- Base processes 50+ TPS at near-zero cost
- CBDCs cannot natively interact with DeFi or smart contracts
<$0.01
L2 Tx Cost
50+
Sustained TPS
03
Settlement Finality vs. Economic Finality
CBDCs promise 'immediate settlement' on a single ledger. L2s like zkSync and Starknet provide cryptographically proven state transitions to Ethereum L1, creating a stronger finality guarantee than any permissioned ledger. The security model is decentralized, not federated.
- Ethereum L1 as the universal settlement layer
- ZK-proofs provide instant, verifiable finality
ZK-Proof
Finality Type
~12s
To L1
04
The Interoperability Chasm
A CBDC is a walled garden. An L2 is a gateway. Polygon, Arbitrum, and Optimism are natively interoperable with a $50B+ DeFi ecosystem via shared bridges and standards. CBDC interoperability requires complex, slow legal agreements, not code.
- LayerZero and Axelar enable cross-chain asset flows
- CBDCs lack a native composability primitive
$50B+
DeFi TVL
Native
Composability
05
Throughput is a Solved Problem
Central banks design for thousands of TPS as a pinnacle. Parallelized EVMs like Monad and Sei, and modular data layers like Celestia, are architecting for 100,000+ TPS. The scaling trajectory of L2s is exponential, while CBDC infrastructure is linear and bureaucratic.
100k+
L2 Target TPS
Parallel EVM
Architecture
06
The Privacy Paradox
CBDCs promise privacy but are inherently surveillable by the issuer. L2s like Aztec offer programmable privacy with zero-knowledge proofs. The market is choosing user-controlled privacy over state-granted confidentiality.
- ZK-SNARKs enable private transactions on public ledgers
- CBDC privacy is a policy, not a cryptographic guarantee
ZK-SNARK
Tech Stack
Programmable
Privacy Model
thesis-statement
THE THROUGHPUT REALITY
The Core Argument: Permissionless Innovation Outpaces Permissioned Design
CBDC speed benchmarks are rendered irrelevant by the exponential scaling of permissionless Layer 2 networks.
Permissionless competition drives exponential scaling. Centralized CBDC projects optimize for a single, controlled ledger. In contrast, Ethereum L2s like Arbitrum and Optimism compete on raw throughput, forcing rapid iteration on data availability and execution environments.
CBDC TPS is a vanity metric. A central bank touting 100k TPS for domestic transfers ignores the composability tax of walled gardens. Permissionless L2s achieve similar speeds while enabling seamless, trust-minimized value movement via bridges like Across and Stargate.
The scaling roadmap is public. CBDC roadmaps are political documents. The Ethereum rollup roadmap—with danksharding and parallelized EVMs via Monad—projects a concrete path to 100k+ TPS for global, permissionless finance, making centralized targets obsolete.
market-context
THE REALITY CHECK
The State of Play: L2s Are Production-Grade, CBDCs Are PowerPoints
Layer 2 scaling solutions have already achieved the performance benchmarks that central banks are still theorizing about.
CBDCs are chasing benchmarks that modern L2s like Arbitrum and Base already exceed. Central bank pilots tout 100k TPS as a future goal, but Arbitrum Nova's data availability layer already handles this scale today.
The innovation gap is structural. Public L2s operate in a permissionless, competitive market where rollup sequencers and validium proofs are stress-tested by billions in real capital. Centralized CBDC labs lack this pressure.
Real-world throughput is proven. Polygon zkEVM processes transactions for Uniswap and Aave with sub-second finality. This is a live, economic network, not a sandbox simulation.
The infrastructure is commoditized. Developers deploy ZK-stack chains and OP Stack rollups in days using standardized tooling from Polygon and Optimism. CBDC architectures remain bespoke and closed.
deep-dive
THE THROUGHPUT GAP
Architectural Asymmetry: Why L2s Will Always Be Faster
Layer 2 scaling solutions operate on a fundamentally different, more efficient architectural plane than monolithic CBDC ledgers, guaranteeing a permanent performance advantage.
Decoupled execution from consensus defines the L2 advantage. L2s like Arbitrum and Optimism batch thousands of transactions into a single, verifiable proof on Ethereum. This parallel processing model is architecturally impossible for a single-threaded CBDC ledger, which must serialize every transaction.
Sovereign data availability layers like Celestia and EigenDA provide L2s with dedicated, high-bandwidth data channels. This specialization creates a performance flywheel where throughput scales independently of the base layer's constraints, a concept foreign to integrated CBDC systems.
Optimistic and ZK-Rollups employ cryptographic proofs to compress state transitions. A single validity proof from a ZK-rollup like zkSync can represent millions of computations, a verification efficiency that monolithic ledgers cannot replicate without sacrificing decentralization or security.
Evidence: The Arbitrum Nova chain, leveraging Ethereum for security and a Data Availability Committee for speed, consistently processes over 100,000 TPS for off-chain games. No proposed CBDC architecture matches this scale without centralizing critical components.
protocol-spotlight
WHY L2S RENDER CBDC SPEED CLAIMS OBSOLETE
The Vanguard: Networks Proving the Point
Central bank digital currencies tout 'fast' settlement, but leading Layer 2 networks already operate at a scale and speed that makes CBDC benchmarks look like dial-up.
01
Arbitrum: The Liquidity Superhighway
The Problem: CBDCs operate in isolated, permissioned silos with limited finality guarantees.\nThe Solution: A general-purpose L2 with >$18B TVL and ~250ms block times, enabling sub-2 second transaction finality for DeFi, NFTs, and payments. Its Nitro stack proves high-throughput execution is a solved problem.
~250ms
Block Time
$18B+
Secured Value
02
Base: Mainstream UX at Scale
The Problem: CBDC user experience is often clunky, requiring custodial wallets and bank intermediaries.\nThe Solution: Coinbase's L2 demonstrates sub-second confirmations for millions of users, processing 50+ TPS sustained with peaks over 200 TPS. Its EIP-4844 blob integration reduces fees to <$0.01, making micro-payments viable.
200+
Peak TPS
<$0.01
Avg. Cost
03
Starknet: Proving Programmable Privacy
The Problem: CBDCs offer a binary choice: total transparency to the state or cumbersome privacy layers.\nThe Solution: A ZK-Rollup using Cairo VM for native account abstraction and validium mode for private, high-throughput transactions. Proves ~1000 TPS is achievable with cryptographic privacy for compliant use cases, far beyond CBDC designs.
~1000
Theoretical TPS
ZK
Proof System
04
Solana: The Throughput Baseline
The Problem: CBDC architects debate 1,000 TPS as a futuristic goal.\nThe Solution: Solana's monolithic L1, a benchmark for L2s, consistently handles ~3,000 TPS with 400ms block times. Its local fee markets and Sealevel VM demonstrate that high throughput with low latency is an architectural choice, not a theoretical limit.
~3,000
Sustained TPS
400ms
Block Time
05
The Interoperability Mandate
The Problem: A CBDC is a single, closed ledger. Real-world finance requires cross-chain asset movement.\nThe Solution: Networks like Arbitrum, Base, and zkSync are natively connected via bridges and shared sequencing layers. Protocols like Across and LayerZero enable <1 minute cross-chain settlements, creating a financial internet where CBDCs would be an isolated island.
<1 min
Bridge Finality
Multi-Chain
Native Design
06
Finality vs. Settlement: The Critical Distinction
The Problem: CBDCs often conflate database commit ('settlement') with cryptographic finality.\nThe Solution: L2s like Optimism and Arbitrum provide economic finality in seconds, with cryptographic finality inherited from Ethereum in ~12 minutes. This layered security model is more robust than a central bank's permissioned ledger, which can be reversed by fiat.
~12 min
L1 Finality
Seconds
User Finality
counter-argument
THE THROUGHPUT FALLACY
Steelman: The CBDC Rebuttal (And Why It's Wrong)
Central bank digital currency speed claims are rendered obsolete by the scaling trajectory of permissionless Layer 2 networks.
CBDCs tout centralized speed as their primary advantage, but this is a static comparison against legacy blockchains. They ignore the exponential scaling of optimistic and ZK-rollups like Arbitrum and zkSync, which batch thousands of transactions off-chain. The relevant benchmark is not a single chain, but the aggregate capacity of the entire L2 ecosystem.
Settlement finality is the real bottleneck. A CBDC transaction is only final when its central ledger updates. An L2 transaction achieves cryptographic finality on Ethereum within minutes, inheriting its security. This is slower than a database commit but is trust-minimized and globally verifiable, a trade-off legacy finance cannot replicate.
The interoperability advantage is decisive. A CBDC exists in a walled garden. A user on Arbitrum can atomically swap assets via Uniswap, bridge to Polygon via Hop Protocol, and use a loan on Aave. This composable financial stack, enabled by shared settlement on Ethereum, creates utility that raw transaction speed cannot match.
Evidence: Arbitrum One currently processes over 200,000 transactions daily. The upcoming Ethereum danksharding upgrade will increase L2 data availability, enabling theoretical throughput exceeding 100,000 TPS across all rollups. No proposed CBDC architecture scales to this level of decentralized, programmable capacity.
future-outlook
THE PERFORMANCE CHASM
The Inevitable Gap: Where This Goes Next
Centralized CBDC architectures are structurally incapable of matching the performance trajectory of decentralized L2 scaling solutions.
CBDCs are architecturally obsolete. Their speed claims rely on permissioned, centralized databases, which hit hard scalability ceilings. Decentralized L2s like Arbitrum and Optimism treat this as a starting point, not a goal.
The scaling roadmap diverges. CBDC development focuses on incremental database optimization. The L2 stack pursues exponential scaling via ZK-rollups, parallel execution engines, and modular data availability layers like Celestia and EigenDA.
Evidence: Current CBDC pilots target ~1,700 TPS. The Starknet L2, using its Cairo VM, has demonstrated a theoretical throughput exceeding 300,000 TPS for specific computations. The gap widens annually.
takeaways
WHY L2S RENDER CBDC SPEED MOOT
TL;DR for Architects
Central bank digital currency pilots touting 1-2 second finality are competing against a benchmark that no longer exists in private crypto infrastructure.
01
The Throughput Ceiling Problem
CBDC architectures like wholesale DLTs (e.g., Project Helvetia, mBridge) are bottlenecked by consensus among a handful of permissioned nodes, capping throughput at ~1k-10k TPS.\n- L2 Solution: Optimistic Rollups (Arbitrum, OP Mainnet) batch thousands of transactions off-chain, achieving effective throughput >100k TPS.\n- Key Benefit: Settlement latency is decoupled from execution speed, making raw TPS claims irrelevant.
>100k
Effective TPS
10-100x
Throughput Multiplier
02
The Finality Latency Illusion
CBDCs market "instant" settlement, but this is only within their closed system. Cross-border or interbank settlement still takes days.\n- L2 Solution: ZK-Rollups (zkSync Era, Starknet) provide cryptographically proven finality in ~10 minutes on Ethereum, with validity proofs ensuring absolute security.\n- Key Benefit: Universal finality to the most secure settlement layer (Ethereum) is more valuable than fast finality in a silo.
~10 min
Proven Finality
100%
Security Guarantee
03
The Cost & Programmability Gap
CBDCs are fundamentally simple payment rails with high operational costs for intermediaries and no native composability.\n- L2 Solution: General-Purpose L2s enable sub-cent transaction fees and host a global, permissionless ecosystem of DeFi (Uniswap, Aave) and smart contracts.\n- Key Benefit: Speed is worthless without utility. L2s provide a programmable financial layer CBDCs can't match, making them the actual infrastructure for future finance.
<$0.01
Avg. TX Cost
$20B+
DeFi TVL
ENQUIRY
Get In Touch
today.
Our experts will offer a free quote and a 30min call to discuss your project.
NDA Protected
Confidentiality24h Response
Time to ValueDirectly to Engineering Team
No Sales Fluff10+
Protocols Shipped
$20M+
TVL Overall