Reserve & hold percentages by acquirer, 2026
- Median rolling reserve in 2026, across all verticals sampled, is 6.4%. In 2023 it was 4.1%.
- Stripe has no pre-agreed reserve structure — it imposes reserves reactively at 0-50% without notice. That is the worst-case for cash planning, even if the median is low.
- Specialist high-risk ISOs offer the widest reserve variance: 0% for clean supplement operators, 30% for peptide / SARMs / kratom newcomers.
- The difference between a "180-day rolling" reserve and a "capped at 6 months" reserve is roughly 1.8% of annualized GMV in operator cash-flow terms.
- Every reserve in our sample was negotiable after 6 months of clean processing. Operators who asked got a median 3-percentage-point reduction.
On this page
Reserves are the single most under-discussed line item in payments. Operators obsess over discount rates and bury the reserve number in the contract, where it quietly compounds into a six-figure cash-flow drag nobody modeled. This report is the 2026 benchmark for what every major US acquirer is actually charging in reserves, how those reserves are structured, and — critically — how they get negotiated down.
Methodology: we compiled reserve data from 287 operators across 11 verticals between April 2025 and March 2026. For each, we captured: (a) the percentage, (b) the structure (upfront vs. rolling), (c) the duration, (d) the trigger (vertical, chargebacks, new merchant, volume spike), and (e) whether it had been renegotiated successfully. We excluded operators who had not seen a single payout cycle (they had no real-world data on whether the reserve actually held).
Reserve types, plain English
Before the numbers, terminology. Reserves come in four flavors and every acquirer mixes them differently:
- Upfront reserve — A fixed dollar amount held at contract signing. Usually 1-4 months of projected monthly volume. Released at contract end or milestone.
- Rolling reserve — A percentage of each day's processed volume held for N days, then released. Classic structure: 5% of daily volume, released 180 days later. At steady state, this equates to 5% × (180/30) = 30% of one month's volume tied up permanently as long as you process.
- Capped reserve — A rolling reserve that stops growing once a dollar cap is hit. Operator-friendly: the reserve peaks and releases against future volume on a one-for-one basis.
- Dynamic / situational reserve — Stripe's special flavor. No pre-agreed structure. Acquirer flips on a reserve at any time, at any percentage, for any duration.
Reserve percentages by acquirer, across all verticals
| Acquirer / platform | Typical reserve range | Structure | Duration |
|---|---|---|---|
| Stripe (standard) | 0% default, 0-50% triggered | Dynamic | Indefinite once applied |
| Stripe Connect (platform) | 0% default, 0-30% triggered | Dynamic | Indefinite once applied |
| Adyen | 0-5% | Rolling | 90-180 days |
| Worldpay / FIS | 0-10% | Rolling or capped | 180 days standard |
| Fiserv / First Data | 0-8% | Rolling | 180 days |
| Chase Paymentech | 0-10% | Rolling or upfront | 180 days |
| Global Payments | 0-10% | Rolling | 180 days |
| TSYS (Merchant Lynx, etc.) | 2-12% | Rolling | 180-270 days |
| EasyPayDirect ISO | 5-15% | Rolling | 180 days |
| Durango | 5-15% | Rolling | 180 days |
| Corepay | 5-18% | Rolling | 180 days |
| PayKings | 7-20% | Rolling or upfront + rolling | 180 days |
| Soar Payments | 5-15% | Rolling | 180 days |
| High Risk Pay | 10-20% | Rolling | 180 days |
| Authorize.net via nutra ISO | 5-15% | Rolling | 180 days |
| NMI via specialist ISO | 7-20% | Rolling | 180 days |
| BlueSnap | 0-7% | Rolling | 90-180 days |
| PayPal / Braintree | 0-25% dynamic | Dynamic rolling | Indefinite |
| Square | 0-30% dynamic | Dynamic | Indefinite once applied |
| multiflow orchestration | 5-10% parent-level | Capped rolling | 180 days, cap negotiated |
Reserve percentages by vertical, median operator
| Vertical | Low end (clean, established) | Median (new, typical) | High end (first-time, VAMP-exposed) |
|---|---|---|---|
| General DTC / apparel | 0% | 2% | 8% |
| B2B SaaS | 0% | 0% | 3% |
| Supplements (mainstream) | 2% | 5% | 12% |
| Nutra (high-claim) | 5% | 10% | 20% |
| CBD | 5% | 11% | 25% |
| Kratom | 8% | 15% | 30% |
| Peptides (research) | 8% | 15% | 30% |
| SARMs | 10% | 18% | 35% |
| Telehealth / TRT | 4% | 9% | 20% |
| Subscription boxes | 0% | 4% | 10% |
| Marketplaces (3P) | 3% | 7% | 15% |
These numbers are the benchmark. If you are at "median" for your vertical and being charged "high end," you either (a) have a real chargeback problem your acquirer is reacting to, or (b) have been assigned default terms and never renegotiated.
The cash-flow impact of reserves, properly computed
Operators frequently underestimate reserves because they compute them as a one-time percentage rather than an ongoing cash lockup. Correct math: the steady-state cash tied up in a rolling reserve equals the reserve percentage × the reserve duration × average daily volume.
Example, 10% reserve on 180-day rolling with $500k/mo volume: $500k ÷ 30 × 180 × 0.10 = $300,000 in perpetual reserve balance. That is cash you do not have, and — critically — it is cash your acquirer is earning interest on, not you.
At 5% interest rates, the opportunity cost alone is $15,000/year on that $300k reserve. That is before the reserve percentage itself, which is a headline-rate compounding factor. This is why we built the free reserve cost calculator.
Stripe's dynamic reserve is the worst-case scenario
Stripe's "0% default" reserve looks attractive on paper, and for 1-2 brand low-risk shops it is. But the dynamic reserve model means Stripe can flip a 30% reserve on at any time, in response to any risk signal, for any duration. Our dataset contains 94 cases of Stripe applying a reserve reactively. Distribution:
| Triggered reserve level | % of cases | Median duration observed |
|---|---|---|
| 5-10% | 18% | 60 days |
| 10-20% | 34% | 180 days |
| 20-30% | 24% | 240 days |
| 30-50% | 16% | 360+ days |
| 100% (full hold) | 8% | Open-ended |
For cash-flow planning, a 0% baseline with an 8% chance of going to 100% is worse than a pre-agreed 10% rolling reserve. Multi-brand operators need predictable reserves, not optimistic ones. See what happens after a Stripe freeze for the full cash-flow math.
The acquirer-type decomposition
Reserves cluster by acquirer type, not vertical, more than operators realize. Three clusters in our data:
| Acquirer type | Median reserve (across verticals) | Reserve structure | Negotiability |
|---|---|---|---|
| Enterprise direct (Adyen, Chase, Worldpay) | 2% | Capped rolling | High |
| Big-tech aggregator (Stripe, Square, PayPal) | 0% baseline / dynamic | Dynamic | Low |
| High-risk ISO (EasyPayDirect, Durango, Corepay, PayKings) | 10% | Rolling 180-day | Medium (after 6 months) |
Operators often pay high-risk ISO reserves for low-risk business because they were onboarded that way and never moved the easier brands elsewhere. This is a common multi-brand drag: your general-retail brand is sitting on a 10% reserve because it was onboarded alongside a peptide SKU.
Negotiation — what actually works
Of 287 operators in our dataset, 113 attempted a reserve renegotiation in the last 12 months. Outcomes:
| Outcome | % of negotiation attempts | Median reduction achieved |
|---|---|---|
| Reserve reduced | 58% | -3.2 percentage points |
| Reserve capped (converted rolling to capped) | 19% | Cap at 2× monthly volume |
| Duration shortened | 11% | 180 to 120 days |
| No change | 9% | — |
| Reserve increased (adverse) | 3% | +5 percentage points |
What worked, in operator-reported order of effectiveness:
- 6+ months of clean processing — the minimum threshold at which most acquirers will even entertain a reduction.
- A concrete second-acquirer bid in writing — asking for a reduction without a competitive quote gets polite no's. Attaching a competitive bid flips conversion rates.
- Chargeback ratio well under vertical median — below p50 in the chargeback benchmark table.
- A representment win rate above 40% — signals operator discipline.
- A documented volume projection — acquirers want sticky merchants; showing growth trajectory buys concessions.
What did not work: long emails about fairness, threats to leave, or requesting reductions during an active risk review. The risk review closes before the reserve conversation opens — never combine them.
The VAMP effect on reserves
In 2024, we saw median reserves in VAMP-exposed verticals hold steady year-over-year. In 2025 and 2026, they jumped because acquirers are passing VAMP fine-exposure risk through to merchants in the form of higher upfront reserves. Expect another 50-100 basis point creep through 2026 until Visa's calibration settles. See VAMP explained for operators.
Multi-brand reserve structures
Multi-brand operators can structurally win on reserves, but only if they negotiate at the parent level. Common multi-brand reserve architectures:
| Structure | Reserve behavior | Best for |
|---|---|---|
| N independent MIDs, N independent reserves | Reserves do not share, no netting | Small portfolios with diverse verticals |
| Parent account + sub-MIDs (Adyen MarketPay style) | Parent-level reserve; sub-MIDs share | 8+ brand portfolios, same operator |
| Orchestration layer + routed reserves | Brand-level reserves at each routed acquirer, but visible in one console | Multi-acquirer high-risk portfolios |
| Aggregation (PayFac) | Aggregator holds reserve; sub-brands invisible to acquirer | Very small sub-brands only (regulatory risk) |
The "parent + sub-MID" structure consistently produced the lowest effective reserve in our multi-brand dataset — roughly 40% lower than the same operator running N independent ISO relationships. The reason: a single reserve negotiation at the parent level beats N separate ones, and netting low-risk brand volume against the reserve base lowers the effective drag. See reserve holds for high-risk verticals and the reserve cost calculator.
How to audit your own reserve today
- Pull your current reserve percentage, structure (rolling/capped/upfront), and duration from your acquirer.
- Compute your steady-state cash lockup: % × (duration/30) × average monthly volume.
- Compare the percentage to the vertical benchmark table above. Are you at low end / median / high end?
- If you're at median and have 6+ months of clean processing, you have negotiation room.
- If you're at high end, your acquirer is treating you as risky — you need a chargeback and representment audit, not a reserve negotiation.
How to cite this report
Data window: April 2025 to March 2026. Respondent count: 287 operators. Reserve data pulled directly from acquirer statements where respondents shared them; self-reported in remaining cases with cross-check against processed volume consistency.
Cite: "multiflow, Reserve & Hold Percentages by Acquirer 2026." Link: multi-flow.pro/blog/reserve-hold-percentages-by-acquirer-2026/.