Fermi swap in four stages, from quote to treasury settlement

Fermi swap moves from a price quote to settlement by checking its own inventory, executing an Ethereum transaction and updating the balances that back future trades. That differs from a conventional automated market maker, where a swap draws against a shared pool governed by a formula. The distinction shifts attention from pool depth and price impact to the maker’s available inventory, the quote’s freshness and whether the transaction lands at the expected output.

The four stages are easier to follow as one trade: a user or router asks for a price, compares it with other venues, submits the swap, then accounts for the tokens received and spent. The fermi swap model is described as a proprietary market maker using its own token inventory rather than shared liquidity pools. That changes who supplies the assets, but not the basic onchain requirement: tokens move only when the transaction executes successfully.

How does a Fermi swap move from quote to settlement?

A Fermi swap starts with a quote for a particular token pair and trade size; the quote estimates the output available from the venue’s current state. A quote is not a completed trade, and a price for one size does not establish the price for another. The route from quote to settlement has four steps:

• Quote: A wallet, interface or router requests an output estimate for a specified input. The estimate depends on the pair, amount and inventory available to the market maker at that moment.
• Compare: A router can check that estimate against other venues, including pool based exchanges. It chooses a route based on expected output and any execution limits set for the trade.
• Execute: The user signs a transaction that transfers the input token and calls the swap. The contract checks the trade against its onchain conditions. If execution fails, the transaction reverts instead of completing only part of the exchange.
• Settle: On success, the input and output token balances change onchain. The inventory vault has one less unit of the token sold and more of the token received, giving the market maker a changed position for later quotes.

This sequence is useful because it separates pricing from settlement. A quote can be requested without moving tokens; a swap commits them. Between those moments, the market can move or the available inventory can change. That gap is why a minimum output and deadline matter: they let a transaction reject an execution whose delivered amount falls below the user’s chosen floor or arrives too late.

How does its inventory model compare with a shared pool?

A shared pool prices trades from the reserves supplied to that pool, often using a formula that changes the exchange rate as the reserves shift. A proprietary market maker instead fills from inventory it controls and sets prices for the size and direction it is willing to trade. The first model makes pool reserves and price impact central. The second makes the market maker’s quote and inventory central.

That can make a useful difference for a large order if the maker is willing to quote enough size at a competitive price. It can also mean less capacity than a router’s headline quote implies: a venue cannot deliver tokens it does not hold or choose to sell. In a shared pool, liquidity providers bear exposure to both assets in the pool. In the proprietary model, inventory management and pricing risk sit with the maker. Neither arrangement guarantees a better fill for every user or every market condition.

For traders, the comparison should use the same input amount and include the likely output, transaction cost and execution conditions. A small trade may show little difference between a formula based pool and an inventory quote. For a larger trade, compare whether the quote covers the full amount and whether another route can deliver more after fees and gas. Protocol wide trading activity does not answer that question for a specific order.

What can change between a quote and the final fill?

The final fill depends on the state the transaction encounters, so a displayed quote should be read as conditional on the trade executing against a compatible state. Ethereum transactions are ordered into blocks; other trades can affect prices or balances before yours is included. The market maker may also no longer have the same inventory available. A router may then use another venue, or the transaction may revert, depending on its route and the protections configured.

That is the central trade-off against a familiar shared pool. A direct inventory quote can be attractive when its offered price and size beat alternatives. But a quote does not guarantee execution, and an available route does not guarantee it remains the best route at inclusion. A minimum output protects against receiving less than the specified threshold; setting it too loosely accepts more price movement, while setting it too tightly can cause a failed transaction and wasted gas.

Before signing, check the token addresses, input amount, minimum output, network and approval request. Token approvals authorize spending up to the approved amount, so avoid approving more than the transaction needs when the wallet gives that choice. These checks apply to swaps generally; they do not establish that any particular token or contract is safe.

Which signals matter for Fermi swap next?

The useful signals are order specific. Watch whether a fresh quote covers the intended size, how its output compares with pool based routes, whether transactions settle at quoted terms, and how inventory depth changes across market conditions. These reveal more about execution quality than a single venue wide volume figure.

The model is clearest when treated as a competing liquidity source inside a broader routing decision. It replaces shared reserves with market maker inventory, while leaving the user to judge price, size, timing and transaction cost. The next comparison is whether that inventory remains available at competitive terms as orders arrive, and whether the quote-to-fill gap stays narrow enough to make the route dependable.