Written by | GaryMa, Wu Blockchain

The WuBlockchain summarizes key developments in the blockchain technology space for October:

Bitcoin

Bitcoin Core officially released v32.1. Version 32.0 had been tagged but was never fully released after last-minute issues were discovered. Alongside the broader v32 improvements, including parallelized block validation and mempool-based fee estimation, v32.1 fixes a potential IP-address leak affecting transactions broadcast through sendrawtransaction under -privatebroadcast=1. Transaction relay rate limiting was also redesigned around a global backlog to reduce CPU and memory pressure during bursts in transaction volume. Bitcoin Core also updated its security disclosure policy: vulnerabilities that are easy to rediscover, including with LLM-based tooling, will now be backported across maintained releases and disclosed more quickly rather than waiting for the traditional long embargo window.

Coinbase published a technical proposal for Bitcoin’s eventual post-quantum migration, arguing that such a transition could take years and should be prepared for well in advance. It suggested exploring extensible signing paths such as BIP-360 (P2MR) and P2TRv2, with a staged transition in which conventional and post-quantum signatures coexist. Coinbase also highlighted the need to evaluate the impact of new signatures on block space, verification performance, and institutional MPC custody. BTC that is not migrated in time, including coins whose keys are lost or have remained dormant for long periods, would need to be handled through a separate policy discussion. The proposal is an institutional technical recommendation, not a Bitcoin network consensus decision.

Ethereum

Glamsterdam upgrade: L1 scaling, parallel execution, and MEV architecture redesign. Progress: Glamsterdam activated on the Sepolia testnet on October 6, testing core changes including ePBS, BALs, and Gas repricing while raising Sepolia’s per-block Gas Limit to 200M to stress client and validator hardware under much heavier loads. Developers then planned a Hoodi activation for October 26 with the same 200M Gas Limit, but postponed it on October 22 to investigate compatibility between EIP-8037’s state-creation Gas changes and existing mainnet smart contracts. No replacement Hoodi date or mainnet date has been set. Testing also surfaced a deposit-handling flaw involving Builder and Validator deposits sharing a public key, which could theoretically allow a malicious operator to have ETH intended for new validators credited to a Builder account. A fix was merged on October 26, and no actual fund losses have been reported. Glamsterdam has therefore moved from basic feature validation into a phase centered on high-Gas stress testing, mainnet compatibility, and security boundaries, with increased risk of schedule slippage.

Hegotá upgrade: censorship resistance, native account abstraction, and post-quantum readiness. Progress: Scope continued to narrow in October. FOCIL (EIP-7805) and Frame Transactions (EIP-8141) remain the core scheduled directions, while developers formally removed EIP-8363, Tapered Issuance Burn, and EIP-8333 from the Hegotá scope. EIP-8363 proposed burning a growing share of validator issuance as the staking ratio rises in order to discourage excessively high staking participation. Developers concluded that ETH issuance policy affects the entire staking ecosystem and the long-term security budget and should not be settled through a single fork-scoping process. The topic is expected to move into a broader issuance-policy discussion.

Ethereum continued advancing execution-proof research. Developers are experimenting with EIP-8025 on Sepolia using OpenVM and 16 RTX 5090 GPUs, split between ethrex and Reth proving workloads. Individual block execution proofs are being generated in roughly 6–7 seconds, with more than 36,000 proofs produced so far. The long-term goal is to let nodes verify execution results through proofs rather than requiring every validator to re-execute every transaction. The system remains experimental, and validators may not yet rely on execution proofs for fork choice or voting.

Ethereum’s post-quantum roadmap became more concrete in October. Vitalik Buterin backed an EIP-8288-style recursive STARK and aggregation path that could support arbitrary signature schemes, allowing users to sign locally with systems such as Falcon or ML-DSA and have those signatures verified through a common STARK layer instead of continually adding new precompiles. Separately, developers submitted an updated EIP-8450 proposal to simplify post-quantum key registration for validators, removing the withdrawal-address authorization path and instead using existing BLS keys to authorize registration and rotation of post-quantum keys, with rotation delays and replay protection. Both remain proposal or research-stage designs.

Ethereum L2s

ZKsync introduced Airbender Prime, a real-time Ethereum prover. In ZKsync’s tests, a single NVIDIA RTX PRO 6000 Blackwell GPU proved 7,200 consecutive Ethereum mainnet blocks, with 99% completed in under 10 seconds and all blocks completed within 12 seconds, including witness preparation. Airbender Prime is based on RISC-V and hash-based STARKs, combining components such as GKR, WHIR, and BLAKE2s with the goal of real-time proving, independent verification, and post-quantum security. Vitalik later argued that zkEVM proving speed is no longer the primary bottleneck and that the next phase should focus on formal verification, proof-system security, and hard bounds between Gas consumption and proving cost.

Starknet advanced both privacy and architecture work in October. It launched Shadow Accounts, allowing users to interact with public DeFi through separate addresses while hiding the relationship between those addresses and their main wallets. Starknet then held a non-binding poll on whether to further research becoming an independent post-quantum L1; 93.52% of voting power supported continuing the research. The vote does not approve a departure from Ethereum. It gives StarkWare and the Starknet Foundation a mandate to study the technical requirements, trade-offs, and migration implications of an independent L1. The motivation is that Starknet itself can move relatively quickly toward post-quantum accounts and STARK-based security, while its bridge, data availability, and other L2 security surfaces remain tied to Ethereum L1’s roadmap.

The Optimism Foundation and Uniswap Labs announced the first planned mainnet set for native OP Stack Interop, consisting of OP Mainnet and Unichain. The initial release is expected to focus on native ETH transfers between the two chains, with cross-chain message validity checked at the protocol level by participating chains rather than by an additional third-party bridge validator set. Upgrade 20 has already moved fault proofs from single-chain Output Roots toward multi-chain-capable Super Roots, while Upgrade 21 is expected to prepare shared ETHLockbox and settlement infrastructure. Interop itself will still require a separate governance proposal and a coordinated L2 hard fork.

Polygon introduced CDK Besu, allowing banks and financial institutions to launch permissioned private chains using Hyperledger Besu while connecting them to external chains and liquidity through Agglayer. The design uses QBFT, lets institutions control validator membership and transaction permissions, and keeps private-chain transaction data inside institutional infrastructure. Cross-chain asset interoperability is handled through Agglayer’s ZK-based security model. Polygon says the product is production-ready and can also connect existing Besu networks.

Solana

Solana completed the final stage of SIMD-0525 in October, reducing the mainnet target slot time from 400ms to 200ms. The final step took effect at epoch 1053 on October 9, increasing target block production from roughly 2.5 to 5 slots per second. Solana Foundation data indicates that skipped-slot rates remained broadly stable through the staged reductions, although shorter slots place greater demands on cross-region networking, validator hardware, and voting latency. The shorter leader window is also framed as a censorship-resistance improvement because each leader controls block construction for less time.

Solana’s largest consensus change since 2020, Alpenglow, activated on mainnet at epoch 1064 on October 21, replacing TowerBFT with the Votor voting algorithm. Alpenglow removes validator votes from on-chain transactions and instead has validators exchange votes directly and aggregate them into certificates. Its target is to reduce deterministic finality from roughly 12.8 seconds to around 150ms. Solana explicitly describes 150ms as a target finality figure rather than a guarantee that every mainnet transaction will always finalize at that latency. The upgrade also changes validator economics by eliminating continuous vote-transaction fees and introducing a per-epoch Validator Admission Ticket.

Solana Foundation released the open-source Solana DvP settlement infrastructure for institutions, with JPMorgan providing design input on securities settlement workflows and institutional requirements. Solana DvP uses a single atomic transaction to settle the asset and cash legs simultaneously, ensuring that both legs settle together or neither does. It supports SPL Token and Token-2022 and has completed an external security audit. Privacy-preserving settlement is planned for future versions.

BNB Chain

BNB Smart Chain activated the Jenner Hard Fork on the Chapel testnet in October. Jenner introduces three major capabilities: BEP-703 Payment Lane, which reserves 5% of block Gas by default for qualifying payment transactions so they cannot be fully crowded out during congestion; BEP-702, which introduces the protocol-native CAS20 standard for stablecoins and tokenized RWAs with optional allowlists, blocklists, pausing, and separate mint, burn, and seizure roles; and BEP-706, which exposes millisecond-precision block timestamps to smart contracts. Jenner is currently active only on testnet, with no confirmed mainnet activation date.

Aptos

Aptos governance proposal #206, Enable ChunkyDKG V1, was executed on October 6, enabling full ChunkyDKG V1 mode together with encrypted_transactions. The system uses distributed key generation to keep transaction contents encrypted before consensus ordering, after which validators cooperate to decrypt them, reducing front-running and MEV exposure caused by transaction contents becoming visible too early. Aptos had previously initialized the encrypted-mempool framework and enabled ChunkyDKG in shadow mode in August; the October change moves encrypted transactions into active functionality.

Aptos Labs announced a new execution engine called MonoMove. Internal benchmarks for the Decibel fully on-chain trading workload showed up to 55x faster collateral withdrawals, 22x faster order placement, and up to 11x higher end-to-end sequential transaction throughput. MonoMove is targeted for feature completion by the end of 2026 and, subject to governance approval, a 2027 mainnet rollout. These numbers are internal benchmark results and should not be treated as guaranteed future mainnet performance.

Celestia

Celestia completed the Mainnet Beta rollout of v10 / Fibre in October. v10 is a breaking network upgrade centered on CIP-51 Fibre. Under Fibre, the chain mainly records Blob Commitments while Blob Payloads are erasure-coded, propagated, and stored through validator-operated Fibre Servers. A publisher collects signatures representing at least two-thirds of bonded voting power before submitting settlement on-chain. This further separates data transport from consensus commitment compared with the traditional path in which all blob data travels through block propagation. Celestia says Fibre went live on mainnet on October 19.

Celestia also published a large-scale Fibre benchmark in October. Across 120 validators in an end-to-end environment, the full 143-second load window averaged 3.07 Tb/s, with a best 60-second period of 3.69 Tb/s. The benchmark used special parameters, including 2 GiB blobs, one-second blocks, and up to 2,000 Fibre submissions per block, which differ from initial mainnet settings, so 3 Tb/s should not be interpreted as the current production network’s sustained capacity. On October 29, Celestia also introduced Voyager, an experimental execution architecture for AI-agent payments, reporting a sustained benchmark of roughly 21 million TPS across 100 million accounts over five hours. Voyager remains early-stage and is still moving toward production.

Security

Binance disclosed a major security incident on October 30 with an estimated impact of approximately $6.3 billion. Binance’s preliminary investigation says the entry point may have involved a compromised DevOps engineer’s laptop, which allowed an attacker to access a service container and submit unauthorized withdrawal transactions. As of the announcement, Binance said it had found no evidence that wallet private keys were leaked. The company reported that roughly $3 billion in affected assets had been frozen on-chain and said it would cover losses using SAFU and corporate funds, while withdrawals were temporarily suspended during security checks. The incident is still under investigation, so the complete attack path, net losses, and root cause remain subject to a future technical post-mortem.

Succinct disclosed a high-severity soundness vulnerability in the SP1 V6 native Rust verifier affecting versions 6.0.0 through 6.8.0. During deserialization, the verifier did not ensure that a Tensor’s declared dimensions matched the actual number of stored elements. A malicious prover could exploit the mismatch so that different verification stages processed different parts of the same Tensor, allowing an invalid execution proof to be accepted. The known affected deployment included Succinct’s mainnet Prover Network. Version 6.8.1 added length and physical-shape checks and has been deployed to the network. Groth16 and Plonk wrapped-proof paths are not affected.

Qubic suffered a wallet-randomness incident in October in which 1,404 wallets lost roughly 2.94 trillion QU, worth approximately $1.3–$1.4 million. Security analysis found that some Qubic Wallet versions used the non-cryptographic Math.random() function to generate account seeds while also exposing outputs from the same pseudorandom sequence through the dejaVu field in network packets. Under certain conditions, observing a small number of outputs could allow recovery of the PRNG’s internal state and ultimately the account seed. The root cause was in wallet key generation rather than the Qubic consensus protocol.

XRP Ledger publicly disclosed in October a payment-engine integer overflow bug that had existed for roughly a decade. An attacker could theoretically construct many extreme order-book offers and use one Payment transaction to overflow a 64-bit sum, creating spendable XRP from nothing. The existing “no XRP created” invariant could overflow in a similar way and fail to detect the mint. The issue was emergency-fixed in xrpld 3.4.1 on September 25 and fully disclosed on October 9. No evidence has been found that it was exploited on a public network. Because of the severity, the fix was shipped without waiting for the normal amendment activation process.

Others

Zcash continued converging on the scope and activation height of its next network upgrade, NU7, expected around November 5. NU7 is planned to reduce the block interval from 75 seconds to 25 seconds and shut down the legacy Sprout privacy system. In parallel, Zcash developers are researching hash-based post-quantum signatures, initially for transparent addresses, and more granular viewing-key designs for cross-chain swaps to reduce future quantum-related privacy leakage. The post-quantum features do not yet have a confirmed mainnet activation date.

Sui Devnet completed a native ML-DSA-65 post-quantum signature transaction. Mysten Labs’ devnet-v1.82.0 added ML-DSA-65 verification and support across ExecuteTransaction, VerifySignature, multisig, GraphQL, and the Rust SDK. Sui plans to move the feature toward testnet after independent security review, with a target of native post-quantum account authentication on mainnet in Q1 2027.

NEAR said its account architecture can already support post-quantum schemes such as ML-DSA, while Near One is advancing NEP-655, which derives implicit accounts from the hash of initial account state to reduce premature exposure of public keys or their hashes on-chain. NEP-655 is on testnet, while the team is also designing extensible signature support for NEAR Intents and a Q-day recovery-registration mechanism for chains that do not yet support post-quantum keys.

LayerZero, a16z crypto research, Carnegie Mellon University, and Imperial College London researchers introduced Aerie, a post-quantum signature aggregation system based on the Akita lattice-based polynomial commitment scheme. Official benchmarks report that on an M4 Max using 16 CPU cores and no GPU, Aerie can aggregate 32,768 Falcon-512 signatures in around 0.65 seconds, with aggregate-proof verification roughly 10x faster than verifying the signatures individually on a single core. The work aims to reduce the bandwidth and verification cost of post-quantum consensus.

Espresso published post-upgrade mainnet benchmark results showing roughly 181 MB/s of data throughput in a 100-node environment with simulated global network latency, with average finality around 778ms under peak load. Based on a 180-byte stablecoin-transfer assumption, that corresponds to theoretical throughput above one million transfers per second. The architecture moves block building into validator nodes, uses a fully connected encrypted network, XOR erasure coding, and BLAKE3, and removes the separate data availability committee in favor of consensus votes proving data availability. These are benchmark results rather than sustained production business throughput.

Paradigm completed an audit of evm2, a new Rust EVM implementation with built-in JIT compilation that is expected to be adopted across Foundry, Reth, Tempo, and related projects. The codebase claims execution performance of up to roughly 2x revm while allowing developers to customize opcodes, Gas rules, transaction types, and precompiles without modifying the core. The 2x figure is a project benchmark and should not be interpreted as a guaranteed end-to-end Ethereum mainnet node performance gain.

Follow us

Twitter: https://twitter.com/WuBlockchain

Telegram: https://t.me/wublockchainenglish