The three liquidity buckets
Every cross-border FX execution ultimately pulls from at least one of three sources:
- Banks and correspondent credit lines. Traditional FX liquidity. Deep in major pairs, slower in emerging markets, business-hour constrained.
- Dedicated market makers (MMs) and OTC desks. Specialist firms that stream quotes continuously, typically with tighter spreads than banks on deep pairs and the ability to price exotic pairs when banks cannot.
- On-chain pools. Automated market makers (Curve, Uniswap) and dedicated stablecoin pools, quoting 24/7 directly on-chain.
Good operations use all three. Great operations route each trade dynamically to whichever pool offers best execution at that moment.
Bucket 1: Banks
Strengths
- Deepest liquidity in USD-EUR, USD-JPY, USD-GBP at institutional size.
- Natural settlement into fiat accounts.
- Regulatory familiarity.
Weaknesses
- Business-hour windows reduce availability at peak volume times.
- Emerging-market pairs (BRL, MXN, INR, IDR) often thinly quoted outside local business hours.
- Spreads embed the bank's own cost of capital and correspondent network.
Bucket 2: Market makers and OTC desks
Strengths
- Continuous pricing, often 24/7.
- Willingness to quote exotic and mid-liquidity pairs.
- Competitive spreads driven by specialized technology and inventory.
- Increasing cross-venue market makers who quote both traditional and stablecoin pools.
Weaknesses
- Counterparty risk requires real due diligence and credit limits.
- Settlement integration is more varied (some settle via banks, some via stablecoin).
The best institutional MMs in stablecoin-era FX (Cumberland, Wintermute, Galaxy, B2C2, FalconX and others) now stream prices across dozens of pairs continuously.
Bucket 3: On-chain pools
Strengths
- 24/7 always-on pricing, no counterparty credit required for the execution itself.
- Instant atomic settlement for stablecoin-to-stablecoin swaps.
- Public, transparent pricing.
Weaknesses
- Slippage increases steeply on large tickets in all but the deepest pools.
- Smart contract risk must be managed with protocol whitelists.
- Some pools have MEV (maximum extractable value) exposure; institutional flows use private routing to mitigate.
Curve's 3pool and USDC-USDT stable pools are among the deepest on-chain sources of stablecoin-to-stablecoin liquidity in the world.
When each bucket is best
| Scenario | Best bucket | Why |
|---|---|---|
| USD-EUR, $10M+, business hours | Banks or MMs | Deepest natural liquidity |
| USD-BRL, $1M, after hours | MMs + on-chain | Banks partially closed |
| USDC-USDT swap, any size | On-chain pools | Deepest on-chain liquidity |
| BRL-MXN (exotic), $500k | MMs | Specialists price where banks do not |
| USD-IDR (thin), $5M | Aggregated MMs + banks | Single source insufficient |
Smart order routing: how aggregation actually works
A production liquidity aggregator does the following for each trade:
- Capture the required pair, size, direction, and urgency.
- Query streaming quotes from all connected banks, MMs, and on-chain pools.
- Normalize quotes to an all-in basis (including settlement costs, on-chain fees, FX spread).
- Evaluate slippage for the proposed size at each venue.
- Split the order across venues if aggregate cost is better than any single venue.
- Execute, settle, and reconcile within seconds.
Splitting a $5M USD-BRL order across, say, 40% to an MM, 30% to a correspondent bank, and 30% across two on-chain pools can outperform any single venue's pricing by 5 to 20 bps, depending on corridor and time.
Hedging and inventory management
Liquidity provision is balance-sheet-intensive. A provider streaming tight USD-BRL quotes must hold meaningful inventory of BRL (or BRZ), USD (or USDC), or both, and hedge that inventory against FX moves. Two recurring models:
- Net inventory. Hold position, hedge via FX forwards or on-chain perpetuals.
- Matched book. Offset every client trade with a back-to-back hedge immediately. Tighter risk, slightly thinner margin.
For an institution using a liquidity partner, the relevant question is not which model the partner uses, but whether their hedging discipline is strong enough to price consistently under volatility. Ask for historical spread behavior during stress events.
What thin corridors look like with and without aggregation
Without aggregation
A $1M BRL-MXN trade routed through a single regional bank in a thin corridor might:
- Execute at 300-500 bps spread over mid-market.
- Take hours to clear.
- Risk partial fill or delayed confirmation.
With aggregation
Same trade aggregated across two specialist MMs and one on-chain pool might:
- Execute at 100-180 bps spread.
- Clear in minutes.
- Provide firm pricing at quote.
The difference (often 100-300 bps) is pure execution quality improvement for the end client.
The single best way to evaluate an FX provider's liquidity sourcing is to request historical executed spreads on your top three corridors, at realistic size, across both peak and off-hours. A good provider supplies this monthly. A weak one obfuscates.
Bottom line
Tight FX is not a pricing claim, it is a sourcing discipline. The institutions that consistently deliver competitive execution are the ones that operate a real aggregation layer across banks, dedicated market makers, and on-chain pools, with smart routing, hedged inventory, and transparent reporting. That is the infrastructure underneath every credible stablecoin-era payments operation.