The chain choice for an institutional tokenisation product is a strategic decision before it is a technical one. The chain dictates which custodians can hold the asset, which counterparties can transact in it, which regulators have already touched the surface, and how cleanly the product composes with the rest of the on-chain institutional stack. Ethereum mainnet sits at one end of that spectrum: highest cost per transaction, slowest finality, and by some distance the deepest pool of institutional infrastructure, regulated counterparties, and tokenised-asset gravity. This part covers what Ethereum mainnet actually buys an issuer, what it costs them, and why MONY (and BUIDL, BENJI, OUSG before it) chose mainnet despite cheaper alternatives existing.
What Ethereum mainnet actually offers
The largest economic security budget in smart-contract land. Ethereum's validator set is the deepest, the staked ETH backing it is the largest, and the operational track record across multiple market cycles is the longest. For an issuer holding regulated client assets, that depth is not a vanity metric. It is the answer to the diligence question every institutional allocator eventually asks: what is the chain itself worth attacking, and what does an attacker need to commit to do so. On mainnet, the answer is "more than any plausible attacker will commit." On a frontier L1, the answer is "less, by orders of magnitude."
The deepest institutional infrastructure surface. Every institutional custodian that supports a single chain supports Ethereum first. Every transfer agent that has built a tokenisation rail has built it on Ethereum first. Every regulated venue, every qualified custodian, every MPC (multi-party computation) custody provider, every analytics shop, every block explorer. When BlackRock wanted Securitize to tokenise a money-market fund, the question of which chain was barely a question. Going anywhere else would have meant rebuilding a stack that already existed on mainnet.
Composability with the existing tokenised-asset universe. BUIDL is on mainnet. BENJI (Franklin Templeton's tokenised MMF) is on mainnet. OUSG (Ondo Finance) is on mainnet. Apollo's ACRED is on mainnet. The dollar stablecoins (USDC, USDT, the institutional variants) all have their canonical issuance on mainnet. For a new tokenised MMF such as MONY, being on mainnet means an existing institutional holder can rebalance between BUIDL, BENJI, and MONY inside a single custody domain, against a single set of policy rules. That is composability in the only sense that matters at the institutional layer: rebalancing without bridging.
What it costs
Gas. Specifically, gas variance. A simple ERC-20 transfer on mainnet costs single-digit dollars on a quiet day and tens of dollars during congestion. A more involved transfer through a permissioned token (identity registry lookup, compliance check, transfer agent record update) is multiples of that. For an institutional MMF subscribing in eight-figure tickets, gas is rounding error. For a product with retail-style transaction frequency, mainnet is cost-prohibitive.
Finality. Specifically, the gap between block inclusion and economic finality. A block on mainnet is produced every twelve seconds. Probabilistic finality (the point at which reorg risk is negligible) takes multiple blocks; deterministic finality under the post-Merge consensus takes roughly thirteen minutes. For most institutional flows, that lag is acceptable; for a real-time payments rail, it is not. This is one of the reasons JPMD did not launch on mainnet (covered in Part 2).
Throughput ceiling. Mainnet processes roughly fifteen transactions per second at base layer. The L2 ecosystem exists in part because mainnet alone cannot serve consumer-grade transaction frequency. For an institutional MMF with low transaction velocity, this is fine. For a payments product, it is not.
Why MONY chose mainnet
JPMAM launched MONY on Ethereum mainnet for the same set of reasons every institutional MMF tokenisation has chosen mainnet: institutional brand recognition, the broadest infrastructure support, and a deliberate decision to land where the existing tokenised-asset gravity already sits.
The composability argument is the load-bearing one. An allocator who already holds BUIDL and BENJI in a Fireblocks or Anchorage wallet can add MONY into the same wallet, the same policy framework, the same reporting pipeline, without touching any new chain integration. The marginal cost of adopting MONY is close to zero. If MONY had launched on a different chain, every prospective holder would need to evaluate the chain integration as a separate exercise, which materially raises the friction for an institutional product whose entire point is operational ease.
There is also a regulatory-posture point. JPMAM is a regulated MMF issuer in the US. Choosing the chain with the longest regulatory track record, the broadest custody coverage by federally chartered crypto banks, and the most established institutional infrastructure is the conservative choice. A frontier chain, even one with better technical properties, would have been the wrong shape for a first institutional MMF tokenisation under JPMAM's name.
What this sets up for the rest of the chapter
Ethereum mainnet is the right answer when institutional gravity, brand, and composability dominate the trade-off. It is the wrong answer when transaction frequency, fee predictability, or sub-transaction privacy dominate. The next four parts walk through the chains that solve for those alternative constraints: Base for the predictable-fee payments case (Part 2), Canton Network for the privacy and institutional-securities case (Part 3), Solana for the high-throughput crypto-native partner case (Part 4), and finally the multi-chain-by-design synthesis (Part 5) where each chain does what it is best at and the orchestration layer connects them.