iGaming PSP Comparison 2026: Card vs Crypto-Native Payment Gateways

The iGaming payment stack split into two rails in 2026, and operators picking one are leaving money on the table. Card acquiring still moves the licensed deposit volume that regulators expect, but crypto-native gateways now handle the volumes a card stack would have rejected, repriced punitively, or held in reserve for six months. The question this guide answers is not which single card acquirer to pick, or which single crypto rail. It is which combination of the two matches your volume tier, your license, and your tolerance for held capital.

Ten providers in our database serve iGaming books at production volume. Five card-side. Five crypto-native. None of them is a "best gateway" in isolation. The sections below shortlist each set, then collapse the choice into a volume-tiered decision rule.

Why operators run two rails in 2026

The April 2026 VAMP threshold drop is the proximate cause. Visa moved the excessive merchant ratio from 2.2 percent to 1.5 percent on April 1 in North America, Europe, and Asia-Pacific (other regions stay at 2.2 percent for now), with $8 per disputed transaction and no warning tier, per the Visa fact sheet and Cside's 2026 playbook. iGaming dispute rates run higher than mainstream e-commerce because of buyer's-remorse-after-loss disputes and tighter issuer scrutiny on MCC 7995, so a tighter threshold lands harder on iGaming books than on retail.

Crypto-native gateways do not face the VAMP problem at all on wallet-paid volume. A deposit paid from the player's own crypto wallet never touches Visa or Mastercard, so there is no card transaction to dispute and no chargeback ratio to manage. Where a player buys crypto with a card first, that purchase runs through a third-party on-ramp (the Simplex and MoonPay class), and the on-ramp is the merchant of record for the card leg, not the gateway and not the operator. Either way, the operator's own MID carries none of the dispute burden. That architectural difference is why operators are layering crypto-native rails alongside card stacks rather than replacing one with the other.

The second reason is structural cash flow. A card acquirer at $2 million monthly volume holds 5 to 10 percent rolling reserve for six months, which means $600,000 to $1.2 million of working capital sits at the acquirer at any given time. A crypto-native rail settles in USDC or USDT to a wallet with no rolling-reserve mechanic; custody risk is a separate question, covered in the negotiation section below. Operators report the cash-drag math is what moved their boards, not the dispute math.

The third reason is segment fit. Part of the player base already holds crypto and prefers depositing in it. A card-only stack either loses those players or forces them through the approval rates MCC 7995 card traffic gets, while a crypto rail picks them up at instant settlement and fees around 1 percent.

Five card acquirers worth a shortlist

These five all underwrite iGaming MIDs at production volume in 2026 and have actual operator references in regulated markets. The first comparison sets the dimensions; the paragraphs below it explain the fit.

Worldpay's gaming practice carries 200+ gaming operators, with heritage references like Ladbrokes and Coral, and its underwriting handles UKGC- and MGA-licensed books at production volume. Settlement is T+7 or longer, deposit fees blend 1.5 to 3.5 percent. Reserves run 8 to 15 percent — among the heaviest of the card acquirers here, and a negotiable starting point rather than a floor. The trade-off operators report is slower settlement and a six-month migration cost to switch out once embedded.

Where Nuvei wins is iGaming-specific tooling. Network tokenization, 3DS exemption flows, and a published VAMP playbook addressing the April 2026 threshold change. Pricing is custom in the same 1.5 to 3.5 percent band as Worldpay; settlement is T+7 or longer. The differentiator operators reference is approval-rate uplift on MCC 7995 versus generic acquirers, sourced from Nuvei's own benchmarks rather than independent measurement.

Adyen is the strictest gate of the tier-1 names here. It holds banking licenses in the EU, UK, and US and runs acquiring, processing, and settlement on one platform, so there is no orchestration layer in the middle to negotiate with. Operators with a clean MGA or UKGC stack get tier-1 acquiring on interchange-plus pricing; operators with thin compliance documentation get rejected at onboarding rather than at volume review.

Paysafe operates 30+ local payment methods alongside card acquiring, plus the digital-wallet brands Skrill, Neteller, and paysafecard under one roof. That matters for operators whose player base in regulated EU markets already uses those wallets. Settlement is T+2 to T+3, faster than Worldpay or Nuvei direct. Fees blend 1 to 2.9 percent on custom contracts.

IXOPAY's value is portfolio diversification. The product sits in front of multiple acquirers and routes per transaction by issuer BIN, decline patterns, or fee. The orchestration fee is 0.1 to 0.5 percent on top of the underlying acquirer cost. The reason to pick IXOPAY is failover: if one acquirer rate-jacks or pulls out of a vertical, the orchestration layer routes around it without a six-month migration project.

Five crypto-native rails worth a shortlist

These five all take deposits straight from the player's crypto wallet, settle to stablecoin or fiat with no rolling-reserve mechanic, and target the operators that either cannot get card-only approval or want a parallel rail to reduce acquirer dependency.

BitPay opened in 2011 as the first crypto payment processor of any scale and remains the regulated US-friendly option in 2026. It holds state money transmitter licenses where the law requires one, including the New York BitLicense, and converts to fiat at T+1 when the merchant wants off-ramp. Fees are tiered: 2 percent at low volume, dropping to 1 percent above $1 million monthly. Chain support is conservative (Bitcoin and Ethereum first, others added selectively), which fits operators that want a regulated counterparty more than maximum coin breadth.

CoinGate holds a MiCA CASP licence from the Bank of Lithuania (December 2025) and runs Chainalysis-grade KYT on inbound wallets. Instant settlement, around 1 percent fees, broader chain support than BitPay. A reasonable default for European-licensed operators who want a crypto rail without taking on offshore counterparty risk.

CoinsPaid is the iGaming specialist of the set, and it is a custodial gateway: player crypto lands in CoinsPaid-controlled wallets, converts there, and settles to the merchant in EUR or stablecoin. The custody is the trade-off to price in. Its July 2023 hot-wallet breach, attributed to the Lazarus Group, drained more than $30 million from exactly those pooled wallets. What the custody buys is the deepest iGaming-specific feature set in the segment, around 0.8 percent fees, and instant settlement. Independent reporting puts 1xBit, Bitcasino, and Sportsbet.io on its customer roster.

Where NOWPayments competes is breadth. 300+ cryptocurrencies supported, 0.5 to 1 percent fees, instant settlement. Non-custodial by default: inbound payments forward to the merchant's own wallet as they clear, with custodial conversion as an optional add-on. That is the mirror image of CoinsPaid's model, and the structural reason NOWPayments carries less freeze risk at the gateway. The fit is operators serving global player bases who need long-tail altcoin support alongside USDT and USDC.

CoinPayments is the documentation-light option, not the breadth play it used to be. After the platform revamp its supported-coins list sits around 40 assets (the legacy 2,000-coin era is over), with 0.5 to 1 percent fees, instant settlement, and lighter compliance requirements than the four above. Production fit is gray-market books running primarily on USDT volume. Less appropriate as the only rail for a regulated EU operator, and if long-tail coin support is the requirement, NOWPayments is the breadth leader of this set.

The decision rule per volume tier

The "best" combination is the one whose constraints match yours. Run the rule against your monthly card volume and your license stack; the answer comes out in one row.

1. Under $500k monthly card volume, regardless of license: orchestrator (IXOPAY) plus a crypto-native fallback (CoinGate or NOWPayments). Direct acquirer relationships are overkill at this scale, and the orchestration fee is cheaper than maintaining two acquirer accounts.
2. $500k to $2 million monthly with a UKGC, MGA, or DGOJ license: Worldpay or Paysafe direct as primary, with IXOPAY layered for failover, plus CoinGate or CoinsPaid as a crypto rail. The card stack carries the regulator-facing volume; the crypto rail picks up segments where card approval is poor.
3. $2 million to $10 million monthly with a tier-1 license: Nuvei or Worldpay direct, IXOPAY for redundancy, plus a crypto rail picked on custody posture: NOWPayments if funds should forward to your own wallet, CoinsPaid if the iGaming feature set is worth custodial settlement. At this volume the rolling-reserve drag is a board-level number, so the crypto rail's structure matters more than coin breadth.
4. Above $10 million monthly: dual card acquirers (Nuvei and Worldpay) on IXOPAY orchestration, plus two crypto rails for redundancy (NOWPayments for non-custodial volume, CoinsPaid for the iGaming-specialist stack). Concentration risk is the constraint, not cost.
5. Offshore-licensed operators (Curacao GCB, Anjouan, Costa Rica): card acquiring is mostly closed. Run a crypto-first stack: NOWPayments for non-custodial settlement and coin breadth, CoinsPaid or CoinGate for the specialist iGaming feature set, CoinPayments where documentation-light USDT volume is the book. Add Paysafe selectively if you can clear their post-2024 Curacao reform underwriting.

What to negotiate before signing

The rate sheet is not the contract. The clauses below decide whether month-nine economics match the day-one quote.

Card-side levers

• Push on the rolling reserve percentage and hold period. Ten percent for six months is the negotiable starting point, not the floor. Get the release schedule in writing on a fixed monthly cadence, not "at acquirer discretion".
• Cap chargeback fees at $25 to $35 per dispute. Some acquirers quote $40 plus, and the math compounds quickly on iGaming dispute volumes.
• Clarify who carries MID concentration risk if the acquirer pulls out of iGaming mid-contract. The answer decides whether you need a second acquirer in parallel or whether the orchestrator covers the gap.
• Lock in the network-token rollout timeline. Approval-rate uplift from tokens runs in the 3 to 4.6 percent band per Visa's published tokenization data, and it only materializes after the rollout actually ships. Vague timelines are a tell.

Crypto-side levers

• Decide on custody upfront. Custodial gateways carry the same freeze-risk shape as card acquirers; non-custodial gateways do not. Ask what specifically triggers a freeze in the custodial model. Few providers will commit the trigger conditions to writing.
• Confirm fiat off-ramp restrictions per geography before signing. The wallet-side coin support is rarely the constraint; the fiat-out side is.
• Confirm KYT coverage on inbound wallets, especially on USDT volume. The FinCEN and OFAC rules implementing the GENIUS Act's AML requirements are still a proposed rule (NPRM, April 2026) and target permitted stablecoin issuers rather than gateways, but they are the direction of travel for dollar-stablecoin compliance.

The two-rail pattern is not a hedge. It is the structural answer to a 2026 stack where card acquirers handle the volume their underwriting wants and crypto rails handle the rest. Pick one card acquirer or orchestrator from the shortlist above. Pick one crypto-native gateway whose compliance posture matches your jurisdiction. Run both in production for ninety days, measure approval rates, settlement velocity, and held-capital cost. The combination your numbers identify is the right one for your book.