Rakshith Aloori
How and why of StablecoinsAugust 9, 2025 · ChatGPT

> Tell me about the company bridge and stablecoins

short version: bridge is a (now stripe-owned) stablecoin infra company: apis to accept/send stablecoins, custody wallets, issue white-label stablecoins (and its own “usdb”), and even spin up visa cards that spend from a stablecoin balance. it’s basically “stripe, but with stablecoins as the rail.” (Stripe, bridge.xyz, apidocs.bridge.xyz)

details, bc you probably want the receipts:

  • who/what: founded 2022 by zach abrams + sean yu (ex square/coinbase/airbnb/brex). stripe closed the acquisition on feb 4, 2025. branding is still “bridge,” but it’s a stripe company rn. (TechCrunch, bridge.xyz, Stripe)

  • what it ships: “orchestration” apis to move/store/accept stablecoins; “issuance” apis to create your own branded stablecoin or use bridge’s usdb; custodial wallets; and “cards” to issue visa cards that draw on stablecoin balances. (bridge.xyz)

  • cards x visa: on apr 30, 2025 visa + bridge launched stablecoin-linked card issuing so devs can programmatically issue cards in multiple countries via one api. (Visa Investor Relations, Stripe)

  • usdb (their coin): dollar-pegged, marketed as 1:1 backed by cash + short-duration money market funds at blackrock; pitch includes sharing the reserve yield with builders/users (vs issuer keeping it all). afaict, that’s the differentiator vs usdc/usdt. (bridge.xyz, Ledger Insights)

  • live usage: e.g., neobank slash just launched “usdsl” on bridge for global b2b payments (goal: faster settlement, lower fx). (CoinDesk, Yahoo Finance)

context on stablecoins (super quick):

  • fiat-backed (usdt, usdc, pyusd, etc.) = tokens redeemable for dollars held as cash/treasuries; used for cross-border and 24/7 settlement. bridge plugs into this stack and also issues its own. (Circle)

  • why anyone cares rn: costs + speed vs switf/ach, and global coverage. even tradfi is leaning in (visa partnership; stripe rolling stablecoin accounts to 100+ countries). (Stripe, Bloomberg.com)

caveats / risk notes (bc be real):

  • “1:1 backing” and “yield-sharing” are claims—read the docs + legal fine print; reserve structure, segregation, and who actually custodies the funds matter. bridge says reserves are in segregated, bankruptcy-remote accounts with regulated institutions, but disclosures aren’t as battle-tested as usdc’s transparency stack. (apidocs.bridge.xyz, Circle)

  • regulation is in flux country-by-country; product availability and compliance obligations will vary (money transmission, securities if yield is shared, etc.). idk your jurisdiction, so confirm before you ship.

tl;dr: if you’re building with stablecoins, bridge gives you a one-stop api: on/off-ramps, wallets, issuance (incl. usdb), and spend via visa—now with stripe’s distribution. if you want chains/fees/regions they support rn, say the word and i’ll pull the exact matrix. (bridge.xyz, Stripe)

> How are stablecoins different from Fiat? Is stablecoins another layer, something like compilers that match c code to specific machine code? The local Fiat will be c code, the stablecoins will be assembly and the machine code will be a Fiat of different currency?

short answer: stablecoins aren’t new money; they’re a new RAIL + FORMAT for old money. they wrap fiat (usually usd) into a blockchain-native IOU so it can move/settle 24/7 with code.

what that means, concretely:

  • issuer: fiat (cash/base money) = central bank liability; bank deposit = commercial bank liability; stablecoin = private issuer (or protocol) liability. different legal umbrellas, different recourse.

  • settlement: cash = physical handoff; bank deposit = private bank ledgers + rtgs/ach/swift; stablecoin = public chains (eth/solana/etc.). same dollars, different plumbing.

  • convertibility: bank deposits ↔ cash at par (modulo bank failure); stablecoin ↔ fiat at par only if issuer/redeemers actually honor it. pegs are claims, not laws.

  • risk stack: fiat (sovereign, inflation); deposits (bank/insurer risk); stablecoins (issuer reserve quality, redemption gates, chain/bridge risk, smart-contract bugs, regs). this is where things break first (see ust/terra; overcollateralized ones survive better; fully-reserved fiat ones act like e-money).

  • features: programmability + composability (contracts can hold/spend them), global reach, finality mins not days; privacy is pseudonymous not cash-anonymous; compliance varies by venue.

your compiler analogy (cute) but off by a layer:

  • think in three axes: SEMANTICS (economic unit: usd/eur/inr), ENCODING (how the claim is represented: cash note, bank ledger entry, blockchain token), TRANSPORT (which rails move it: rtgs/ach/swift vs ethereum/solana).

  • stablecoins change ENCODING + TRANSPORT, not SEMANTICS. a usdc is still “usd semantics,” just token-encoded and chain-transported.

  • fx (usd→inr) is a SEPARATE translation step done by a market maker/exchange; the chain doesn’t magically compile usd into inr. you still need a counterparty and a price.

if you must force a cs metaphor:

  • semantics = c the language (usd as the unit of account)

  • encoding = object file format (bank ledger vs token)

  • transport = linker/loader + bus (ach/swift vs blockchain)

  • fx = transpiler to a different language (usd→inr via a market/exchange)
    stablecoins aren’t “assembly” for fiat; they’re more like “the same program packaged as a wasm module so it runs anywhere.”

types, bc taxonomy matters:

  • fiat-backed custodial (usdc, pyusd, etc.): reserves = cash/treasuries; redeemable if kyc’d; lowest volatility; issuer/blacklist risk.

  • crypto-backed overcollateralized (dai in practice): on-chain collateral >100%; peg via incentives; carries on-chain liquidity risk.

  • algorithmic/seigniorage (terra/ust et al.): peg via reflexivity; historically fragile. proceed with… lmao no.

tl;dr: stablecoins = tokenized fiat claims with public, programmable settlement. not new currency, just a different wrapper and pipe. useful for speed/composability; still subject to issuer + market + chain risk.

> Give me an example of a transaction done through stablecoins. Add as much detail as possible and mention where the real “value” of the money resides at each step

here’s a concrete, end-to-end example with maximal guts. scenario: you (payer) in india pay a freelancer in nigeria 1,000 usdc on solana, then they off-ramp to ngn. i’ll track “who owes what to whom” at every hop (aka where the real value sits).

0) prelude: what a fiat-backed stablecoin actually is

  • object: usdc = a digital IOU from the issuer for 1 usd per token.

  • backing: issuer holds fiat reserves (cash/treasuries) at custodians; the tokens are its liability, reserves are its asset.

  • key idea: chains record who holds the IOU; the IOU’s value lives in the issuer’s reserves.

1) fund the on-ramp

  • action: you move inr from your bank to a regulated crypto on-ramp/exchange.

  • ledger moves:

    • your bank deposit (inr) ↓; your bank’s liability to you shrinks.

    • on-ramp’s bank deposit (inr) ↑; the on-ramp now owes you a balance in its system.

  • where the value is rn: at the on-ramp’s bank (a commercial bank deposit asset the on-ramp holds).

2) buy 1,000 usdc

  • action: inside the on-ramp, you trade inr → usdc at the quoted inr/usdc rate (fx + spread + fee).

  • what actually happens under the hood (typical paths):

    • simplest: the on-ramp already has inventory of usdc in a hot wallet. they credit your exchange account 1,000 usdc and debit their inventory 1,000.

    • or: a market maker sells you usdc and later redeems or mints with the issuer to balance their book.

  • where the value is rn:

    • your “1,000 usdc” is now an issuer liability owed to the on-ramp’s omnibus wallet on-chain.

    • the economic backing sits in the issuer’s reserves (cash/treasuries), not at the on-ramp. the on-ramp just controls the claim for you.

3) withdraw to your self-custody wallet

  • action: you provide your solana address; the on-ramp submits a token transfer: on-ramp-wallet → your-wallet, 1,000 usdc. network fee paid in a dust amount of sol. finality in seconds.

  • chain effect: ownership of those exact token units flips to keys you control.

  • where the value is rn: still the issuer’s reserves. only the pointer (who can redeem the IOU) changed.

4) you pay the freelancer

  • action: you push 1,000 usdc on solana to their address. they see it confirmed in a few secs.

  • settlement semantics: this is final at the chain level (subject to chain guarantees); there’s no correspondent bank, no cut-off windows.

  • where the value is rn: still in issuer reserves; now the freelancer controls the redemption claim via their keys.

5) freelancer decides to keep usdc (optional path)

  • if they hold usdc, nothing redeems. the IOU persists, backed by issuer reserves.

  • value rn: issuer reserves; the risk stack = issuer/asset risk + chain blacklisting/smart-contract risk.

6) or they off-ramp to ngn (more common)

  • action (custodial off-ramp): freelancer deposits 1,000 usdc to a local exchange in nigeria.

    1. chain: their wallet → exchange wallet (1,000 usdc).

    2. kyc/aml checks happen here.

  • the exchange sells usdc → ngn:

    1. internally, either a market maker buys the 1,000 usdc (paying ngn) or the exchange itself takes the other side.

    2. to square books, the buyer can redeem with the issuer (burn 1,000 usdc, receive ~$1,000 usd to their bank) or keep the usdc as inventory for future sellers.

  • payout: exchange credits freelancer’s ngn balance, then sends a local bank transfer.

  • where the value is rn (stepwise):

    1. when exchange receives tokens: issuer reserves back the IOU; exchange controls it.

    2. if redemption happens: issuer destroys 1,000 usdc and wires usd to the redeemer’s bank → now the value sits as a commercial bank usd deposit owned by the market maker/exchange.

    3. fx usd→ngn occurs via banking/forex rails; the counterparty’s ngn bank deposit funds the freelancer’s payout.

    4. after local transfer clears, the freelancer ends with a commercial bank deposit (ngn) at their bank.

  • net: the tokenized claim is extinguished (if redeemed); the “value” has migrated from issuer reserves → redeemer’s usd deposit → ngn deposit at the freelancer’s bank.

7) fees + slippage (illustrative math)

assume:

  • on-ramp fee 0.50% on fiat buy, spread 0.20% (effective 0.70%).

  • solana transfer fees ≈ negligible for this size.

  • off-ramp fee 1.00% + fx spread 0.40% on usd→ngn.

if you intend to deliver exactly 1,000 usdc to the freelancer, you’d buy a bit more to cover on-chain dust, but ignoring dust:

  • you pay: inr for 1,007.07 usdc gross, the on-ramp clips 0.70% → 1,000 usdc net to your wallet.

  • you send: 1,000 usdc to freelancer (fee ~$0.00x).

  • they off-ramp: 1,000 × (1 − 1.00% − 0.40%) = 985.9 usd-equiv worth of ngn hits their bank (at whatever fx rate).

  • moral: most friction is on/off ramps + fx, not the chain.

8) where things can break (and where value “pauses”)

  • your bank/on-ramp: transfer limits, blocked deposits → value stuck as your bank deposit or on-ramp deposit.

  • issuer: freeze/blacklist or redemption gates → value still in issuer reserves, but your ability to realize it is impaired.

  • chain: outages/bugs don’t erase reserves, but delay control changes. value remains in reserves while ownership can’t update.

  • off-ramp/fx: local controls, illiquid books → value trapped as usdc at exchange or as usd at a correspondent bank waiting for fx.

9) tl;dr money map (assets vs records)

  • unit of account: usd (semantic layer) never changed.

  • encoding/transport:

    1. bank world: commercial bank deposits (inr/usd/ngn) on private ledgers.

    2. crypto world: usdc tokens on a public ledger.

  • “real value” at each step:

    1. before buy → your bank deposit (inr)

    2. after buy, before withdraw → issuer reserves back the usdc; exchange controls the claim

    3. after withdraw / after you pay → issuer reserves, with you/freelancer controlling the claim

    4. after off-ramp redemption → usd bank deposit of the redeemer

    5. after fx + payout → freelancer’s ngn bank deposit

stablecoins don’t mint new value; they let you re-encode and teleport a fiat claim across an open settlement rail, then decode it back into local deposits. everything else is fees, risk, and latency.

> Where in this flow are transactions cheaper or being made easy compared to Fiat to Fiat transactions? Because ultimately it is a Fiat to Fiat transaction

short version: the savings aren’t at the endpoints (fiat→on-ramp and off-ramp→fiat); they’re in the MIDDLE (the “teleport” leg) and in OPS (automation, 24/7, fewer hops). it’s still fiat→fiat overall, but the center of the pipeline gets way cheaper/faster and that often dominates.

where it’s cheaper/easier vs bank-only rails:

  1. on-chain settlement (the middle hop)
  • cost: cents to fractions of a cent per tx on fast chains vs $10–$50 wire fees or ~2–3% card rails.

  • speed: seconds/minutes, 24/7/365, no cutoffs or “banking hours.”

  • fewer intermediaries: no correspondent chain, so fewer tolls and fewer failure modes (fewer returns/compliance holds).

  • push, final-ish: no chargebacks. for b2b/creator payouts, that replaces card interchange + chargeback risk with near-zero network fees.

  1. working-capital + treasury effects (quiet but real)
  • no pre-funding nostro/vostro, no “float stuck over the weekend.”

  • example opportunity cost: $100k delayed 2 business days at 5% annual cost of capital ≈ $27 saved per transfer simply by shaving latency. tiny per tx, material at scale.

  1. ops + reconciliation (aka “time = money”)
  • programmable: webhooks/scripts settle + post to your ledger instantly; no csv wrangling or opaque bank refs.

  • composability: escrow, split-payments, streaming pay, conditional release — all “one tx,” which subs out legal/escrow/admin fees you’d otherwise pay.

  1. corridor arbitrage (sometimes)
  • in ill-served fx corridors, crypto market makers quote tighter spreads than retail banks/remitters. when the off-ramp is competitive, total all-in (on-ramp + chain + off-ramp + fx) can beat bank-only by 1–2% points. not universal; very corridor-dependent.

where it’s NOT cheaper (be real):

  • great domestic rails (upi in india, sepa instant in eu): bank transfer is ~free and instant; stablecoins add on/off-ramp fees for no gain.

  • weak/expensive off-ramps or illiquid local fx: you’ll eat 1–3% on the way out; then the advantage of a near-free middle hop shrinks or flips.

  • compliance friction: if your counterparties can’t/won’t kyc at exchanges, you’ll pay an aggregator to mediate (fees creep back in).

how to think about total cost (rough mental model):
total = on-ramp fee + chain fee + off-ramp fee + fx spread ± slippage + your time/ops cost − working-capital saved

  • best case (both sides willing to hold/spend stablecoin; no off-ramp): ~0.0x% + seconds.

  • typical cross-border with healthy ramps: ~1–3% all-in and same-day, vs wires/remitters that land ~2–5% and 1–3 biz days (or cards at 2–4% + fx).

  • worst case (bad off-ramp, tight regs): >3–5% and lots of friction; just use bank rails.

mapping to the earlier india→nigeria example:

  • cheaper/easier: step 3–4 (on-chain transfer + finality). that’s where you win on fees/speed and ops.

  • break-even/depends: step 2 + 6 (on/off-ramps + fx). if those ramps are competitive, you still net ahead; if not, they eat the savings.

  • “where the value lives” doesn’t change this: it sits in the issuer’s reserves during the middle hop; you’re optimizing the transport/ops, not conjuring new value.

tl;dr: stablecoins don’t change the unit (usd/inr/etc.); they change the ENCODING + TRANSPORT. savings come from a near-free, 24/7 middle leg and programmable ops. if your ramps/fx are decent or you can stay on-chain end-to-end, you win; if your ramps are trash, you don’t. afaict that’s the honest cut.

WritingsAnthologyFavorite PoemsLLM ChatsMissing SemestersMapsRSS FeedsBookshelfReach Out