21 Web3 Mandates: Where the Gas Costs Actually Went, and What We Would Chain-Pick Today

Across 21+ Web3 mandates Wavect has shipped, roughly 60-75% of cumulative gas cost went to write transactions on user-facing flows, around 15-20% to contract deploys (one-time but painful on mainnet), about 5-10% to read calls that should have been off-chain, and the rest to retries, failed transactions and migrations. If we picked chains today with what we know now, almost no greenfield product would launch on Ethereum mainnet first. Base, Arbitrum or Solana would handle 80% of use-cases.

Numbers are aggregate across mandates, framed as ranges from Wavect's engagement history, not exact per-project audits.

Where does gas actually accrue in a typical Web3 product?

Founders worry about deploy cost. Deploy is a one-time hit. The real cost compounds in user-facing write transactions over months. Here is how we see it break down across mandates.

• Contract deploys (15-20%). One-time per contract. Painful on Ethereum mainnet (often four to five figures EUR for non-trivial logic). Trivial on L2s and Solana.
• Write transactions (60-75%). Mints, transfers, swaps, position updates, claim flows. Every user action. This is the budget that compounds.
• Read calls (5-10%). Should be free via standard RPC. Becomes a cost when teams use paid providers (Alchemy, Infura paid tiers) or do reads on-chain that should have been a subgraph.
• Off-chain compute (variable). Indexers, oracles, keepers, signers. Not "gas" strictly, but in the same operational budget. Often underestimated.
• Retries and failed txs (3-7%). Slippage failures, nonce conflicts, gas underestimation. Eats more budget than founders expect.

How do the chains actually compare for typical workloads?

Comparison across the chains we have shipped on. Costs are typical-case order of magnitude, not exact pricing snapshots (gas markets move). Use as decision framework, not as a quote.

The pattern is obvious. Anything EVM-L2 is roughly 50-200x cheaper than mainnet per transaction. Solana is roughly another 5-20x cheaper than L2s for simple ops. For high-frequency consumer use-cases, mainnet is not competitive on cost.

What did we actually ship on which chain?

Some specifics from products Wavect built. Treat this as engagement context, not as a chain leaderboard.

• Scramble Pay: cross-chain payment flows, multi-chain EVM by design. The chain selection was a product decision, not a cost decision.
• Quivr: built around Solana for transaction cost and confirmation speed on consumer flows. Different product, different chain, same logic.
• Account abstraction work: EVM L2 by default. Mainnet AA exists, the economics rarely justify it for consumer flows.
• Lightbridge: cross-chain messaging, by definition multi-chain. Chain selection here is "wherever the user already is", not "wherever is cheapest".
• MetaMask Snap work: EVM ecosystem by design. The Snap is the wallet extension layer, the chain depends on what the user interacts with.

"Chain selection is a product decision, not a technology decision. Pick the chain where your users already are, and where your unit economics survive at the transaction frequency your product needs."

What would we chain-pick today, by use-case?

• Consumer app, mass-market, low value per tx. Solana, or Base if you must stay EVM. Mainnet is wrong. Polygon is acceptable but losing relative share.
• DeFi product, sophisticated users, higher value per tx. Arbitrum or Base. Mainnet is acceptable if your average tx size justifies $20+ gas.
• NFT-heavy product. Solana for cost, Base for distribution into Coinbase-adjacent users, Polygon if you have an existing brand pipeline there.
• Cross-chain or wallet infrastructure. EVM-first plus selective Solana support. Both ecosystems are too big to ignore in 2026.
• Account abstraction product. EVM L2 default. The AA tooling on L2s is more mature than alternatives for most teams.
• Regulated / institutional use-case. Ethereum mainnet or a regulated L2. Compliance and audit ecosystem still concentrates there. Cost is not the primary criterion.

What did we get wrong on chain selection?

Honestly: too much default-to-Ethereum-mainnet in 2021 and 2022. At the time, L2s were not as mature and the bridge UX was rough. With hindsight, several mandates would have been better served by launching directly on Arbitrum or Polygon, even with the slightly less liquid token ecosystem at the time. The product cost of being on mainnet was higher than the distribution benefit for the user segment.

We also under-weighted Solana for consumer-facing use-cases for too long. The chain has tradeoffs (different toolchain, smaller EVM-skilled hiring pool), but for sub-dollar transactions on consumer flows the economics are not close.

How should a founder pick a chain in 2026?

1. Estimate transactions per user per month at scale. Not at MVP. At scale.
2. Multiply by 12 months times target user count. That is your annualized gas footprint at unit cost X.
3. If that number on Ethereum mainnet is more than ~5% of expected per-user revenue, mainnet is wrong.
4. Then decide between EVM L2s (developer pool, tooling maturity) and Solana (cost, speed) based on the rest of your product surface.
5. Avoid launching multi-chain on day one. Pick one. Add chains when you have users and reasons.