EVM vs SVM: Ethereum Virtual Machine vs Solana Compared (2026)
TL;DR
- The Ethereum Virtual Machine (EVM) is a stack-based, sequentially-executed runtime for Solidity and Vyper smart contracts, with a 256-bit word size and a fixed-cost opcode model. It is the lingua franca of Ethereum L1, every Ethereum L2 (Arbitrum, Base, Optimism, zkSync Era) and dozens of EVM-compatible L1s.
- The Solana Virtual Machine (SVM) is an account-isolated, parallel-executed runtime built on Sealevel, BPF/SBF bytecode and Rust. Each transaction declares its account access list up front, allowing the runtime to schedule non-conflicting transactions concurrently across CPU cores.
- The two VMs differ in account model (EVM accounts/storage vs SVM standalone accounts), execution model (sequential vs parallel), gas (gwei + EIP-1559 base fee burn vs CU compute units priced in lamports), bytecode (custom 256-bit stack VM vs SBF register VM), tooling (Solidity/Hardhat/Foundry vs Rust/Anchor) and MEV (off-chain MEV-Boost vs in-protocol Jito-Solana).
- The 2025-26 picture is convergent: parallelized EVMs (Monad, MegaETH, Sei v2, Reth) are bringing SVM-style throughput to EVM, while Eclipse and Sonic SVM take the SVM into Ethereum settlement and Solana chain-extensions respectively.
Table of contents
- What is the EVM?
- What is the SVM?
- Account model: EVM accounts vs SVM accounts
- Execution model: sequential vs Sealevel parallel
- Gas vs compute units
- Opcodes vs BPF/SBF bytecode
- Solidity, Vyper, Rust, Anchor and Solang
- Throughput, finality and fees in 2026
- MEV pipelines: MEV-Boost vs Jito-Solana
- Wallets: MetaMask vs Phantom
- Porting dApps: Neon EVM, Solang, Eclipse
- Adjacent VMs: Move, CairoVM, Bitcoin script
- Parallelized EVM: Monad, MegaETH, Sei v2
- SVM-everywhere: Eclipse and Sonic SVM
- Bridges between EVM and SVM
- Comparison table
- Risks and tradeoffs
- FAQ
- Glossary
What is the EVM?
The Ethereum Virtual Machine (EVM) is a stack-based, deterministic, Turing-complete virtual machine that executes smart-contract bytecode on every Ethereum node. Specified in the Ethereum yellow paper by Gavin Wood and refined across a decade of EIPs, it is the canonical runtime of Ethereum and the de facto standard for almost every L2 rollup and EVM-compatible L1.
Defining characteristics:
- 256-bit word size — every stack item is 256 bits wide, optimized for cryptographic operations (hashes, ECDSA signatures).
- Stack-based execution — operations consume and produce values on a stack of up to 1024 items.
- Deterministic, sandboxed — given identical inputs every node produces identical outputs; no clock, no I/O, no randomness beyond
BLOCKHASH. - Sequential within a block — transactions in a block execute in their listed order, one at a time.
- Gas-metered — every opcode has a fixed gas cost; runaway computation halts when gas is exhausted.
Major clients implementing the EVM include Geth (Go), Nethermind (C#), Besu (Java), Erigon (Go) and the high-performance Reth (Rust, by Paradigm). The Pectra (May 2025) and Fusaka (H2 2026 planned) upgrades extend EVM semantics with EIP-7702 (account abstraction for EOAs), EIP-7251 (validator MaxEB), and PeerDAS for blob sampling.
What is the SVM?
The Solana Virtual Machine (SVM) is an account-isolated, BPF-based runtime that executes Rust smart contracts ("programs") in parallel via Sealevel. Specified in Anatoly Yakovenko's 2017 Solana whitepaper and implemented across the Agave/Anza, Firedancer (by Jump Crypto) and Jito-Solana validator clients, the SVM is the runtime behind every Solana mainnet transaction.
Defining characteristics:
- Register-based BPF/SBF bytecode — Solana programs compile from Rust to SBF (Solana Bytecode Format), a register-based bytecode derived from eBPF. Faster to JIT than the EVM's stack model.
- Account-isolated state — there is no global storage tree; every byte of state lives in some account, and every transaction explicitly lists the accounts it touches.
- Parallel execution — Sealevel groups non-conflicting transactions and executes them concurrently across CPU cores.
- Compute Unit (CU) metering — instead of "gas," compute is measured in CUs, with a default 200K CU per transaction and a hard cap of 1.4M.
- Programs are stateless code — Solana cleanly separates "program" (immutable executable) from "account" (mutable data).
The SVM has been deliberately designed for hardware scaling: more cores, more memory, faster networking → more TPS. This is in contrast to the EVM, which is intentionally tuned for credibly neutral, easy-to-verify execution on commodity hardware.
Account model: EVM accounts vs SVM accounts
This is the single biggest semantic difference between the two VMs.
EVM account model
Ethereum has two account types:
- Externally Owned Accounts (EOAs) — controlled by a private key.
- Contract accounts — controlled by code, with associated storage as a key-value Patricia trie.
A contract's storage lives in a per-contract Merkle tree, and transactions can touch any storage slot of any contract — there is no upfront declaration. The runtime discovers conflicts at execution time, which precludes naive parallelism.
EIP-7702 (Pectra, May 2025) added a third de-facto type: an EOA can temporarily delegate execution to a smart-contract code path within a single transaction, yielding most of the UX benefits of account abstraction without requiring a separate ERC-4337 smart account.
SVM account model
In Solana, everything is an account:
- Programs are accounts whose data is BPF bytecode (immutable once frozen).
- Each user is an account.
- Each "storage slot" is its own account, owned by some program.
- Each NFT, each token balance, each pool, each order — all separate accounts.
Every transaction must declare upfront the full list of accounts it will read or write — its access list. This declaration is what enables Sealevel parallelism and what makes the SVM feel different to EVM developers.
The cost is dev complexity: a single Solidity contract with mappings becomes a constellation of PDAs (Program-Derived Addresses) on Solana, each requiring explicit allocation and rent management.
Execution model: sequential vs Sealevel parallel
Sequential EVM
The EVM processes transactions in order within a block. Even when two transactions touch disjoint state, the EVM runs them sequentially; this simplifies semantics but caps single-thread throughput at whatever the slowest opcode can sustain. As of 2026 mainnet Ethereum sustains roughly 15-30 TPS at typical loads.
Sealevel parallel SVM
Sealevel scans every transaction in a block, looks at the declared account-access list, and groups non-conflicting transactions into batches that execute in parallel. Two transactions touching disjoint sets of accounts run truly concurrently across CPU cores. Conflicts force serialization but only along the conflict boundary — the rest of the block parallelizes freely.
The result is a runtime whose throughput scales with hardware. Solana mainnet routinely processes 2,000-4,000+ real (non-vote) transactions per second in production, with peaks above 6,000 TPS during high-activity windows.
Flashbots research on parallel block building and the BIS Technology of DeFi paper both highlight Sealevel as the foundational design influence behind the parallelized-EVM efforts of 2025-26.
Gas vs compute units
| Concept | EVM | SVM |
|---|---|---|
| Unit of work | Gas | Compute Unit (CU) |
| Smallest fee denom | Wei (10^-18 ETH) | Lamport (10^-9 SOL) |
| Tip mechanism | EIP-1559 priority fee | Compute-unit price + Jito tip |
| Burn | Base fee burned (EIP-1559) | No protocol burn (validator fees) |
| Per-tx limit | 30M gas (~half block) | 1.4M CU |
| Block target | 30M gas (15M target) | ~48M CU per block |
| Typical median fee | $0.50-5 (L1), $0.001-0.05 (L2) | $0.0002-0.005 |
Because Solana fees are quoted in CU per microlamport (and total fees are typically a fraction of a cent), Solana is the dominant chain for high-frequency activity — payments, gaming, MEV searcher traffic — where Ethereum L1 economics simply do not work.
Opcodes vs BPF/SBF bytecode
The EVM defines roughly 140 opcodes — ADD, SLOAD, KECCAK256, CALL, STATICCALL, LOG0…3, CREATE, CREATE2, etc. Each opcode has a fixed gas cost. EOF (EVM Object Format) — proposed for inclusion in Fusaka or later — restructures bytecode into containers with explicit code/data sections and stricter validation.
The SVM uses SBF, a register-based bytecode derived from eBPF. SBF programs are JIT-compiled at load time and run inside a sandboxed verifier that enforces memory safety. Compared to the EVM:
- Register VMs are typically faster than stack VMs at runtime.
- BPF tooling is mature (Linux kernel, observability, debuggers).
- Verifier checks are stricter — programs that fail safety checks are rejected at deploy time.
The tradeoff is portability: EVM bytecode is a self-contained spec; BPF/SBF is closer to a moving target as Solana iterates.
Solidity, Vyper, Rust, Anchor and Solang
EVM languages
- Solidity — the dominant EVM language. Contract-oriented, JavaScript/C++ syntax.
- Vyper — Python-flavored, emphasizing safety and auditability.
- Yul — low-level intermediate language for hand-tuned EVM contracts.
- Huff — assembler-level EVM language for specialty contracts.
- Fe — Rust-inspired, statically typed, currently early-stage.
SVM languages
- Rust — the dominant SVM language; bare-metal against
solana-programor via Anchor. - Anchor — high-level framework on top of
solana-programproviding macros, IDLs, and ergonomic account validation. - C / C++ — supported but rarely used in production.
- Solang — Solidity-to-SBF compiler, supporting a meaningful Solidity subset for porting.
Tooling comparison
| Layer | EVM | SVM |
|---|---|---|
| Language | Solidity, Vyper | Rust |
| Framework | Hardhat, Foundry, Truffle | Anchor, Seahorse |
| Testing | forge test, hardhat test | anchor test, cargo test-bpf |
| RPC SDK | ethers.js, viem, web3.js | @solana/web3.js, anza-xyz/solana-program |
| Wallet adapter | wagmi, rainbowkit | wallet-adapter (Phantom, Solflare) |
| Indexer | The Graph, Subsquid | Helius, SimpleHash |
Throughput, finality and fees in 2026
| Metric | Ethereum L1 | Top Ethereum L2s | Solana | Solana SVM L2 (Eclipse) |
|---|---|---|---|---|
| Real TPS sustained | 15-30 | 100-500 each | 2,000-4,000 | 1,000-2,000 (early) |
| Block / slot time | 12 s | 0.5-2 s | 400 ms (slot) | 400 ms (target) |
| Time to finality | ~12.8 min (Casper FFG) | inherits L1 | ~13 s (2/3 stake) | inherits L1 (rollup) |
| Median user fee | $0.50-5 | $0.001-0.05 | $0.0002-0.005 | $0.001-0.02 |
| Validator/sequencer count | ~1.05M validators | usually 1 sequencer | ~1,500 active | early single-sequencer |
Ethereum-aggregate throughput (L1 + all L2s) is in the same order of magnitude as Solana mainnet by 2026. The qualitative difference is uniformity: Solana is one chain with one user, fee and dev experience; Ethereum is a federation of L1 + 50+ L2s with varying sequencer policies, finality and bridging.
MEV pipelines: MEV-Boost vs Jito-Solana
Ethereum: MEV-Boost
Ethereum's MEV pipeline is off-chain. Validators delegate block building to specialized builders, mediated by relays:
- Searchers craft transaction bundles (e.g. arbitrage, liquidations) and submit to builders.
- Builders construct full blocks and submit them to relays with a bid.
- Relays publish the bid to validators via MEV-Boost; validators sign whichever bid maximizes their revenue.
Flashbots operates the dominant relay; bloXroute, Aestus, Manifold, BloXroute Regulated, and Eden run alternatives. The architecture is described in Flashbots' MEV and the limits of scaling and Mitigating MEV with FHE.
Solana: Jito-Solana
Solana's MEV pipeline is in-protocol, embedded in the validator client. Jito Labs maintains jito-solana, a fork of the Agave/Anza Solana validator client. Searchers send bundles plus tips to the Jito Block Engine; the leader's Jito-Solana client orders bundles by tip and includes them in the block.
Key facts (April 2026):
- ~80%+ of Solana stake runs jito-solana.
- Jito tips paid to validators are denominated in SOL and have grown into one of the most material non-issuance income streams for Solana validators.
- The Jito Foundation governs the protocol economics through the JTO token.
- Jito Liquid Staking — JitoSOL — is one of the largest LSTs on Solana, capturing a portion of MEV tips for stakers.
The two designs both implement proposer-builder separation but with very different trust assumptions: MEV-Boost relies on out-of-protocol relays that could equivocate or censor; jito-solana embeds the auction in the client where it benefits from in-protocol slashing and economic alignment.
Wallets: MetaMask vs Phantom
| Capability | MetaMask (EVM) | Phantom (SVM + EVM) |
|---|---|---|
| Native chains | Ethereum, all L2s, EVM L1s | Solana, Ethereum, Polygon, Base, Bitcoin, Sui (2025+) |
| Storage | Local extension + mobile | Local extension + mobile |
| Hardware wallet | Ledger, Trezor, GridPlus | Ledger |
| dApp protocol | EIP-1193 / WalletConnect | wallet-adapter / WalletConnect |
| Snaps / extensions | MetaMask Snaps | None equivalent |
| Notable user counts (2026) | ~30M MAU | ~15M MAU |
MetaMask remains the dominant EVM wallet, with MetaMask Snaps extending support to non-EVM chains via plug-ins. Phantom is Solana's dominant wallet but in 2024-25 became a multi-chain wallet, supporting EVM, Bitcoin and Sui in addition to Solana — making it the closest thing to a true cross-VM consumer wallet.
Porting dApps: Neon EVM, Solang, Eclipse
There are three production paths to bring an EVM dApp to the SVM (or vice versa):
- Neon EVM — a Solana program implementing full EVM semantics. Solidity contracts deploy unmodified, run inside the SVM, pay SOL for compute. Neon is the standard choice for porting unmodified EVM dApps to Solana.
- Solang — a Solidity-to-SBF compiler maintained by the Hyperledger Foundation and Solana Labs. Recompiles Solidity directly for native SVM execution, but supports only a subset.
- Eclipse — instead of bringing EVM to Solana, brings SVM to Ethereum: an Ethereum L2 with SVM execution, Celestia data availability, and ETH gas. Lets Ethereum dApps benefit from Solana-style throughput while keeping Ethereum settlement.
The pattern: developers who want raw throughput often port to native SVM via Solang; those who want EVM compatibility plus Solana liquidity use Neon EVM; and projects that want SVM throughput plus Ethereum settlement build on Eclipse.
Adjacent VMs: Move, CairoVM, Bitcoin script
Beyond the EVM/SVM dichotomy, several other VMs matter in 2026:
- Move VM — designed for Diem/Libra, used by Aptos and Sui. Resource-oriented language with formal verification primitives. Aptos uses a sequential-then-parallel "Block-STM" execution model; Sui runs an object-centric, pre-shared-object-marked parallel runtime closer in spirit to the SVM.
- CairoVM — STARK-friendly VM behind Starknet. Programs are compiled from Cairo and produce STARK-provable execution traces.
- Bitcoin Script — a stack-based, intentionally non-Turing-complete script language. Tapscript (post-Taproot) and OP_CTV / OP_CAT proposals expand its expressiveness.
- WASM — used by CosmWasm (Cosmos), Polkadot's parachain runtimes, NEAR, and Internet Computer.
In 2026 the broader VM landscape can be grouped:
- EVM family: Ethereum L1, all L2s, BNB Chain, Avalanche C-Chain, Polygon PoS, Sonic, Celo, Linea, Scroll, Eclipse-EVM-mode, Monad, MegaETH, Sei v2, Reth.
- SVM family: Solana, Eclipse, Sonic SVM, MagicBlock chain extensions.
- Move family: Aptos, Sui, Movement.
- WASM family: Cosmos appchains via CosmWasm, NEAR, ICP, Polkadot parachains.
- STARK-friendly VMs: Starknet (CairoVM), zkSync Era (LLVM-backed zkEVM Type 4), RISC Zero zkVM, SP1.
Parallelized EVM: Monad, MegaETH, Sei v2
Perhaps the most interesting 2025-26 trend is the EVM-side answer to Solana: parallelized EVMs that scan transaction access lists (or speculate optimistically) to execute non-conflicting transactions concurrently.
| Project | Approach | Status (April 2026) |
|---|---|---|
| Monad | Optimistic parallel execution + custom MonadDB | Mainnet 2026; sub-second blocks targeting 10K TPS |
| MegaETH | Single-sequencer "real-time" EVM, sub-ms blocks | Public testnet; mainnet 2026 |
| Sei v2 | Parallel EVM on top of CometBFT (Cosmos) | Mainnet live |
| Eclipse | SVM execution on Ethereum settlement (not EVM, but SVM-throughput in Ethereum) | Mainnet live 2024 |
| Reth | High-performance Rust EVM execution client | Production 2024+; pairs with consensus clients |
| Starknet | CairoVM with native parallelism | Live; v3 transaction parallelism shipping |
The Monad team's design is the canonical example: a custom database (MonadDB) co-designed with the execution engine, plus optimistic concurrency control that lets the runtime guess at parallel scheduling and reorder retroactively when conflicts are detected. The result is EVM-bytecode-compatible execution at SVM-class throughput — a direct convergence of the two worlds.
a16z's 17 Things We're Excited About for Crypto in 2026 lists parallel EVM as a top theme, alongside SVM-everywhere and the rise of co-designed VM/DB stacks.
SVM-everywhere: Eclipse and Sonic SVM
The SVM is also escaping Solana. Two projects stand out:
- Eclipse — modular L2 combining SVM execution, Celestia DA, and Ethereum settlement with ETH gas. Brings SVM throughput into the Ethereum ecosystem and lets builders deploy Rust/Anchor programs while inheriting Ethereum's economic security.
- Sonic SVM — a Solana-native chain extension SVM that lets games and HFT-style apps run their own sovereign SVM instance without competing with Solana mainnet for blockspace.
These are complemented by Solana Mobile Stack (Saga, Seeker) and Solana Actions / Blinks, which extend the SVM into mobile and into shareable URL-embedded action UIs.
Bridges between EVM and SVM
| Bridge | Mechanism | Strengths |
|---|---|---|
| Wormhole + NTT | Generic message-passing with guardian set | Largest cross-ecosystem coverage; standard for ETH↔SOL token transfers |
| deBridge | DLN cross-chain swaps with no slippage | Fast EVM↔SVM trades; widely integrated |
| Allbridge | Lock-and-mint (Classic) + LP (Core) | Stablecoin focus, simple UX |
| Circle CCTP | Native USDC mint/burn | Best for USDC; supports ETH↔SOL natively |
| Mayan Finance | Cross-VM swap with Wormhole messaging | UX-driven SOL↔EVM swaps |
| LayerZero | Endpoint + DVN message-passing | Generic L2-style messaging (not Solana-native until 2024-25 expansion) |
The 2022 Wormhole exploit — at the time the second-largest crypto hack — drove a wave of bridge security investment; the 2024 Wormhole NTT standard formalizes a least-privilege token-transfer pattern, but bridges remain among the riskiest categories in crypto. The BIS Technology of DeFi survey discusses bridge security architectures in detail.
Comparison table
| Dimension | EVM | SVM |
|---|---|---|
| Word size | 256-bit | 64-bit (registers) |
| Bytecode | Stack-based EVM bytecode | SBF (register, ex-eBPF) |
| State model | Per-contract Patricia trie | Per-account state, all explicit |
| Concurrency | Sequential within block | Sealevel parallel |
| Gas / compute unit | Gas (EIP-1559 base fee + tip) | Compute Unit + lamport price + Jito tip |
| Smallest unit | wei | lamport |
| Native asset | ETH (and L2 ETH-on-rollup) | SOL |
| Smart-contract languages | Solidity, Vyper, Yul, Huff, Fe | Rust (Anchor), C/C++, Solang |
| Devnet/test | Hardhat, Foundry, Anvil | solana-test-validator, Anchor |
| Wallet | MetaMask, Rabby, Safe | Phantom, Solflare, Backpack |
| Block time | 12 s (Ethereum L1), 0.5-2 s (L2s) | 400 ms slot |
| Finality | Casper FFG ~12.8 min | 13 s (2/3 stake), economic finality earlier |
| MEV pipeline | MEV-Boost (off-chain) | jito-solana (in-protocol) |
| Runtime parallelism | Coming via Monad, MegaETH, Sei v2 | Native via Sealevel |
| Specification | Yellow Paper + EIPs | Whitepaper + open-source clients |
| Reference clients | Geth, Reth, Nethermind, Besu, Erigon | Agave/Anza, Firedancer, jito-solana |
| Ecosystem | Largest by TVL and stablecoin float | Largest by raw TPS and active addresses |
Risks and tradeoffs
A balanced read on the EVM/SVM debate:
- EVM credibility neutrality. The EVM's sequential, deterministic, well-specified nature is a feature for institutional adoption. Twelve years of open implementation have produced exceptional spec stability.
- SVM hardware-scaling assumption. The SVM's throughput depends on validator hardware staying ahead of demand. Several 2022-23 outages were tied to runaway transaction floods — improved with QUIC, fee markets and Stake-Weighted QoS.
- Account-model complexity. SVM developers must allocate, manage and pay rent for a constellation of accounts; EVM developers face a simpler mental model but pay much higher per-transaction gas.
- MEV asymmetry. EVM's off-chain MEV-Boost has a richer searcher market but more trust assumptions in relays. SVM's in-protocol jito-solana has cleaner economics but tighter integration with a single client maintainer.
- Bridge perimeter risk. Both ecosystems still rely on external bridges for cross-VM movement; bridges remain the most-exploited surface in crypto.
- Dev-talent supply. Solidity has the larger global developer footprint (8-10x more dev-active accounts than Rust+SVM in 2025-26 metrics). Rust expertise is harder to hire but compounds across other domains (systems, infra, ZK).
For balanced academic context see BIS WP-1066 and BIS WP-1061, and for industry views the Messari State of Solana 2024, Messari State of Ethereum 2024, and a16z State of Crypto 2025.
FAQ
What is the difference between EVM and SVM?
See JSON-LD above. Briefly: EVM = stack-based, sequential, Solidity; SVM = account-isolated, parallel via Sealevel, Rust.
Is Solana faster than Ethereum?
Yes by raw TPS and confirmation latency. Ethereum L1 + all its L2s in aggregate close part of the gap.
What is Sealevel?
Solana's parallel execution engine. Schedules non-conflicting transactions concurrently across CPU cores using transaction-declared access lists.
What is BPF / SBF?
The register-based bytecode format used by the SVM, derived from eBPF. Programs compile from Rust to SBF.
What is the difference between Solidity and Rust for smart contracts?
Solidity is contract-oriented, easier to learn, dominant on EVM. Rust is harder but enforces memory safety; dominant on SVM. Solang lets a Solidity subset compile to SBF.
What is MEV-Boost vs Jito-Solana?
MEV-Boost is Ethereum's off-protocol PBS (relays + builders). Jito-Solana is Solana's in-protocol bundle-and-tip auction validator client.
Can you run Solidity on Solana?
Yes via Solang or Neon EVM.
What are Eclipse and Sonic SVM?
Eclipse is an L2 with SVM execution, Celestia DA and Ethereum settlement. Sonic SVM is a Solana-native SVM chain-extension for games and HFT-style apps.
Will EVM go parallel like SVM?
Yes — Monad, MegaETH, Sei v2, Eclipse and Reth are shipping 2025-26.
How do I bridge between EVM and SVM?
Wormhole + NTT, deBridge DLN, Allbridge, Circle CCTP, or Mayan Finance.
Glossary
- EVM — Ethereum Virtual Machine. Stack-based, 256-bit, sequential.
- SVM — Solana Virtual Machine. Account-isolated, parallel via Sealevel, BPF/SBF bytecode.
- Sealevel — Solana's parallel transaction-execution engine.
- BPF / SBF — Berkeley Packet Filter / Solana Bytecode Format.
- Compute Unit (CU) — Solana's unit of computation, analogous to EVM gas.
- Anchor — Rust framework for building Solana programs ergonomically.
- Solang — Solidity-to-SBF compiler.
- Neon EVM — Solana program implementing full EVM semantics.
- MEV-Boost — Ethereum's off-protocol proposer-builder separation pipeline.
- Jito-Solana — Jito Labs' Solana validator client with in-protocol bundle/tip auction.
- Eclipse — L2 with SVM execution, Celestia DA, Ethereum settlement.
- Sonic SVM — Solana-native SVM rollup / chain extension.
- Monad / MegaETH / Sei v2 / Reth — high-performance / parallelized EVM stack components.
- PDA — Program-Derived Address; deterministically-derived SVM accounts.
- Wormhole NTT — Native Token Transfer standard for canonical cross-chain tokens.
Related reading (internal links)
- What is Ethereum? 2026 guide
- Ethereum Layer 2 Networks 2026 guide
- What is Avalanche? 2026 guide
- What is Cosmos / IBC? 2026 guide
- What is DeFi? 2026 guide
- Stablecoins explained 2026
- What is Bitcoin? 2026 guide
- Best crypto ETFs 2026
Sources and further reading
- Ethereum EVM docs: https://ethereum.org/en/developers/docs/evm/
- Ethereum Yellow Paper: https://ethereum.github.io/yellowpaper/paper.pdf
- Solana docs: https://docs.solana.com/
- Sealevel docs: https://docs.solanalabs.com/runtime/sealevel
- Solana whitepaper: https://solana.com/solana-whitepaper.pdf
- Anchor book: https://book.anchor-lang.com/
- Solang docs: https://solang.readthedocs.io/
- Neon EVM: https://neonevm.org/
- Reth (Paradigm): https://github.com/paradigmxyz/reth
- Monad: https://www.monad.xyz/
- MegaETH: https://megaeth.com/
- Sei: https://sei.io/
- Eclipse: https://www.eclipse.xyz/
- Sonic SVM: https://sonic.game/
- Aptos Foundation: https://aptosfoundation.org/
- Sui: https://sui.io/
- Starknet: https://www.starknet.io/
- Flashbots MEV-Boost: https://docs.flashbots.net/flashbots-mev-boost/introduction
- Jito Labs: https://www.jito.wtf/
- Wormhole: https://wormhole.com/
- BIS WP-1066 Technology of DeFi: https://www.bis.org/publ/work1066.htm
- BIS WP-1061 Cryptocurrencies and DeFi: https://www.bis.org/publ/work1061.htm
- Messari State of Solana 2024: https://messari.io/report/state-of-solana-2024
- Messari State of Ethereum 2024: https://messari.io/report/state-of-ethereum-2024
- a16z State of Crypto 2025: https://a16zcrypto.com/state-of-crypto/
About the author
DeFi Intel Research is the in-house research desk of DeFi Intel, focused on rigorous, citation-first crypto analysis. We hold an active research dataset of 9,900+ crypto entities and 30,000+ relations, used to power both this education library and our market-data products. We do not accept paid placement; this page links only to primary sources or top-tier secondary research. Author profile: /about.