Historical results: six merchants and two networks
Under the filtered measure that excludes user-side causes and platform-policy blocks, all six overseas platform merchants this period (2026-10; the figures are taken from transactions between 2026-08-01 and 2026-08-31, observation has continued to 2026-10-27, and this period was not recomputed) — covering AI subscriptions, ad campaigns, app stores and more — posted approval rates of at least 97.1% on both Visa and Mastercard, with OpenAI / ChatGPT, Apple and Cursor at 100% on both networks.
- The finding is not "your card will never be declined": this measure answers "can the card itself pay this merchant", and neither user-side causes nor platform-policy blocks count in the numerator or the denominator.
- The two card networks are close: the differences this period mainly appear with Anthropic / Claude (Visa 97.6% vs. Mastercard 98.0%) and the Google family (Visa 97.1% vs. Mastercard 100%), all within 3 percentage points.
- The one cell to read with care: Meta / Facebook Ads shows 100% on Visa, but the sample tier is "sparse" — this results from our own product recommendation (we primarily recommend Mastercard for the Facebook Ads scenario), not from merchant behavior.
Approval-rate matrix: merchants × card networks
Dark blue: Visa; light blue: Mastercard. Tiers are qualitative and 100% does not mean no declines across all attempts. Values share the matrix and JSON source; no samples are added.
OpenAI / ChatGPT
Anthropic / Claude
Google (Play / Ads / One)
Apple
Meta / Facebook Ads
Cursor
| Merchant | Visa approval rate | Sample | Mastercard approval rate | Sample |
|---|---|---|---|---|
| OpenAI / ChatGPT | 100% | Adequate | 100% | Adequate |
| Anthropic / Claude | 97.6% | Adequate | 98.0% | Moderate |
| Google (Play / Ads / One) | 97.1% | Moderate | 100% | Adequate |
| Apple | 100% | Moderate | 100% | Moderate |
| Meta / Facebook Ads | 100% | Sparse | 98.9% | Adequate |
| Cursor | 100% | Sparse | 100% | Sparse |
Sample sufficiency is a qualitative tier (adequate / moderate / sparse); tier thresholds and transaction counts are not disclosed, and "sparse"-tier values fluctuate more within a single period, so read trends rather than single-period figures. On 100%: if all of a cell's declines fell into the excluded categories, that cell also shows 100% — 100% does not mean no decline occurred during the period. See the formal measure in the methodology below.
3DS / AVS observations
Within the observation window, in the full decline-reason records (pre-exclusion measure) of the merchants in the table above, no declines attributable to 3DS verification challenges or AVS address-mismatch were found. Note that 3DS is a card-network-level specification (see the EMVCo 3-D Secure specification), and different merchants apply different enablement policies and exemption conditions.
Methodology (fully public)
Data source
Real clearing records from the PIKABAO platform (all authorization records), aggregated by normalized merchant name. No third-party or industry averages are cited; the data-selection criteria and merchant-normalization rules are archived internally to keep each period reproducible.
Time window and update cadence
Observation has run continuously since 2026-08-01, and the observation window advances with the underlying data — currently 2026-08-01 to 2026-10-27 (period 2026-10, dataset revised 2026-10-29, next update 2026-11-30). The 12 approval-rate figures are taken from transactions between 2026-08-01 and 2026-08-31 — a different thing from the observation window above: observation is still running, but this period was not recomputed, so the figures still reflect the August basis. Data from the later part of the window is consolidated when a later period recalculates the rates. The data on Overseas payment compatibility report is a historical period archive, not a statement about the current situation; please cite the period and the approval-rate window when referencing it.
The approval-rate measure
Approval rate = (total authorizations − all declines) ÷ (total authorizations − excluded declines). Don't treat the figures on Overseas payment compatibility report as overall success rates: the overall success rate (including user-side causes and platform-policy blocks) is lower than the filtered approval rate shown here. Two categories are excluded: first, user-side causes — insufficient card balance / limit, a wrong billing address, incorrect card details (CVV / expiry), or a card that is frozen / suspended / not activated; second, platform-policy blocks — transactions actively blocked by restricted merchant categories (MCC) and risk-control rules. Neither category reflects whether the card itself can pay that merchant; the exclusions.
Residual declines
All declines outside the exclusions — including upstream transient errors and unknown-reason items — are counted as they are. Not every cell shows 100%, and this is why.
Known limitations
- Single-platform view: only reflects the performance of PIKABAO platform cards and billing-address configuration, not industry averages.
- Self-reported data: not audited by a third party.
- Samples expressed as qualitative tiers: transaction counts are not disclosed, so they can't be used to estimate the platform's business scale.
- Not stratified by card product type: this period does not distinguish standalone-limit cards from shared-limit cards.
- Overseas payment compatibility report governs the measure: a few pages on the site carry historical data with longer time windows that may differ from this period.
Machine-readable data (JSON / CSV)
Period 2026-10 · Observation window 2026-08-01 to 2026-10-27 · Approval rates from 2026-08-01 to 2026-08-31 · Dataset revised 2026-10-29 · Next update 2026-11-30; the page and data files are generated from the same source.
Overseas payment compatibility report's matrix has a same-source machine-readable version that updates in sync with the page and likewise contains no transaction counts. In the CSV, the first row is a metadata comment line starting with #, and the fields match the JSON.
- JSON dataset: ai-payment-compatibility.json (includes period / updated / methodology / rows)
- CSV dataset: ai-payment-compatibility.csv
Two further datasets: measured daily operations and an estimated platform allocation
The two differ in data nature; cite each one on its own terms.
- Measured daily operations: 88 days from 2026-08-01 to 2026-10-27, covering daily cards issued, activation rate, transaction counts and success rate, USDT top-up amounts, platforms on sale, support tickets and first-response time. Single-platform daily totals, not industry averages. JSON · CSV
- Estimated platform allocation: each day's measured totals split across 138 platform entries and Visa / Mastercard by platform weight (starred entries 3, all others 1) — 24,288 daily records, with daily totals matching the measured totals (card-issued difference 0, top-up difference 0). These are allocated values, not measured per-platform detail, and contain no per-platform transaction counts, approval rates or evidence of acceptance. JSON · CSV · platform totals CSV
Neither dataset feeds the approval matrix on Overseas payment compatibility report. The 12 cells come from filtered authorization records; the allocation data contains no authorization outcomes and cannot be used to derive any platform's approval rate.