The Crypto Card Report — August 2026
In August 2026 I logged 21 data changes across 13 tracked crypto cards. The headline number for the month: across the 9 cards that rank at $1,000/month in the US, the average marketed top rate is 3.44% while the average rate cardholders actually keep is 2.08% — a gap of 1.36 percentage points.
Data last verified: 2026-08-20 · changes logged 2026-08-02 to 2026-08-17
Key numbers, August 2026
- Average effective cashback
- 2.08%
- Across 9 ranked cards at $1,000/month, US, after all fees and caps.
- Marketed-to-real gap
- 1.36pp
- Average distance between a card's headline rate and what it actually pays.
- Median effective cashback
- 1.50%
- Half the ranked cards pay less than this.
- Data changes logged
- 21
- Corrections, resolutions and additions, each dated below.
Widest gap this month: Crypto.com Card markets 5.00% and computes to 1.50% at $1,000/month — a 3.50 percentage-point difference once fees and caps apply.
Top three by spend level (US)
| Monthly spend | 1st | 2nd | 3rd |
|---|---|---|---|
| $500 | Coinbase One Card3.17% · $16/mo | ether.fi Cash3.00% · $15/mo | Lava Card3.00% · $15/mo |
| $1,000 | Tria Card3.92% · $39/mo | Coinbase One Card3.58% · $36/mo | ether.fi Cash3.00% · $30/mo |
| $5,000 | Coinbase One Card3.92% · $196/mo | ether.fi Cash3.00% · $150/mo | Lava Card3.00% · $150/mo |
Marketed rate vs what you keep
| Card | Marketed | Effective | Gap |
|---|---|---|---|
| Crypto.com Card | 5.00% | 1.50% | 3.50pp |
| Plasma One | 4.00% | 1.34% | 2.66pp |
| Tria Card | 6.00% | 3.92% | 2.08pp |
| MetaMask Card | 3.00% | 1.01% | 1.99pp |
| KAST Card | 3.00% | 1.50% | 1.50pp |
| Coinbase One Card | 4.00% | 3.58% | 0.42pp |
| Avici Card | 0.00% | -0.12% | 0.12pp |
| ether.fi Cash | 3.00% | 3.00% | 0.00pp |
| Lava Card | 3.00% | 3.00% | 0.00pp |
"Marketed" is the highest rate any tier of that card advertises. "Effective" is what the engine computes at $1,000/month in the US after tier bands, monthly caps, fees, FX and staking haircut. Method at /methodology.
What changed in August 2026
2026-08-17August re-verification, 11 of 12 cards. ether.fi's referral landing page now advertises up to 3% rather than the boosted grocery and restaurant rate recorded in July, so that figure came off every page carrying it. Gnosis Pay turned out to cap eligible spend by GNO balance as well as setting the rate, and that cap is now modelled. Gnosis Pay's GNO holding requirement is also priced for the first time, which drops it off the top of the EU ranking. Nine other cards re-checked and confirmed unchanged, and ether.fi's neobank relaunch is logged. Coinbase keeps its July date — its rewards page blocks automated access.
| Card | Type | Detail |
|---|---|---|
| ether.fi Cash | corrected | The ether.fi referee benefit is now up to 3% cashback on everyday purchases, replacing the higher category-specific figure carried here since July. Re-checked by direct browser render of ether.fi/@start on 2026-08-17, which resolves to www.ether.fi/app/cash/referral?ref_code=start: the page states 'Use the referral code when you sign up to earn up to 3% cashback on everyday purchases, plus more rewards.' A full-text scan of the rendered page found the only rates stated are that 3% and a separate 1% paid to the referrer. Whether ether.fi changed the offer or the July reading was wrong can't be determined after the fact, so this records only what the page says today. Worth knowing: 3% is the same headline rate the standard Core tier already pays on the first $2,000/month, so on current terms the code decides who gets credited for a signup rather than what you earn. The code itself is still live and the invite page still renders. |
| ether.fi Cash | corrected | Cashback bands re-verified unchanged against help.ether.fi: Core 3% to $2,000 then 1% to $3,000 then 0.5%; Luxe 3% to $10,000; Pinnacle 3% to $50,000; Business 1%. Legacy referral program closure re-confirmed — open only to accounts created before 2026-05-01, ending entirely 2026-10-31. |
| Tria Card | corrected | Re-verified unchanged against Tria's own Season 3 membership post and docs.tria.so/card-terms-us: Virtual 1.5% on the first $100/month then 0.5%, Signature 4.5% on the first $1,000 then 1%, Premium 6% on the first $2,000 then 1%; membership fees $25 virtual, $109 plastic, $250 metal; FX 'up to 3.0%'; ATM 'up to $2 + 3%'. |
| Avici Card | corrected | Re-verified unchanged against docs.avici.money: Platinum virtual $15 first card, Signature virtual $30 then $20/year, Platinum physical $50, Signature physical $75 then $20/year; 0% FX with a noted 0.4–1% Visa cross-border fee; ATM $1 + 0.65% on Platinum and $0 on Signature. No cashback is offered at any tier, which the documentation still confirms by omission. Referral terms unchanged: 10% off any card for the person entering a code, 20% documented to the referrer. |
| KAST Card | corrected | Tier pricing re-verified unchanged on kast.xyz/crypto-cards: Standard free at 1.5%, Premium $1,000/year at 2%, Private $10,000/year at 3%. The $2,000/month cap on Standard cashback could not be re-checked — its source, the concierge.kast.xyz fees article, returns HTTP 403 to automated access, as already recorded. KAST's page also now lists an invite-only 'Solana Solid Gold' tier at +2% on all spend, which is not tracked here because invite-only tiers can't be obtained on request. |
| MetaMask Card | corrected | Re-verified unchanged against metamask.io/card: Virtual 1% mUSD at no annual fee, Metal 3% mUSD on the first $10,000 then 1% at $199/year, no MetaMask foreign transaction fee. Signup status re-confirmed as closed in the US and UK specifically, with the card still available in 40+ other countries. |
| Crypto.com Card | corrected | Re-verified unchanged against Crypto.com's own help article, which is public again: Basic carries no spending rewards, Plus 'up to 2%' capped at $1,250/month of spend, Pro 'up to 3%' capped at $2,500/month, Private 'up to 5%' uncapped. Those spend caps reconcile exactly with the $25 and $75 monthly reward caps modelled here. |
| Nexo Card | corrected | Cashback re-verified unchanged on nexo.com/crypto-card: Base 0.5%, Silver 0.7%, Gold 1%, Platinum 2% in NEXO tokens, set by the ratio of NEXO held against the rest of the portfolio. Known gap now recorded: the same page gives FX as 0.2%-2.5% varying by weekday, weekend and location, while this dataset models a flat 0.2%. That is the favourable end of the range, so Nexo's computed effective rate here is optimistic for weekend or foreign spend. |
| Bybit Card | corrected | Base rate re-verified unchanged at 2% cashback in USDT. VIP levels are still advertised only as 'up to 10% cashback on all purchases' with no per-level breakdown published, so they stay unranked. Regions were not re-checked this pass — bybitcard.com states none and the help-centre fees article timed out. |
| Lava Card | corrected | Re-verified unchanged on lava.xyz/card and the card FAQ: no annual fee, no FX fee, and the card is described as 'a secured credit card, which means you spend directly from your dollar balance'. The headline 'up to 5% back in bitcoin' still carries the caveat that 'specific merchants or categories may not be eligible'. The 3% general and 1% international split this card is actually ranked on was not re-checked, because its source is a captured copy of the full FAQ text rather than a stable public URL. |
| Brighty Card | corrected | Re-confirmed that no fixed rate is published: the cashback terms state rates 'vary by merchant, category, country, and promotional period' and are shown only in the app. Brighty therefore stays excluded from the rankings, which is the same reason it was excluded originally. |
| Gnosis Pay Card | corrected | Gnosis Pay's GNO balance sets a maximum weekly spend eligible for cashback as well as the rate, and that cap is now modelled: $250/week at 0.1 GNO, $375 at 1, $500 at 10, $1,250 at 100, with no rollover of unused cap. The rate ladder itself is confirmed unchanged at 1/2/3/4% base. Caps are carried as cashback_bands converted to monthly eligible spend at 52/12 weeks per month ($1,083/$1,625/$2,167/$5,417) with a 0% band above, because the engine reasons in monthly spend. Effect: below roughly $5,400/month nothing moves, and at $10,000/month the top tier drops from 3.95% to 2.14% effective. Confirmed against the primary help-centre article, read in a logged-in browser by the site owner because automated access returns HTTP 403. |
| Gnosis Pay Card | corrected | The GNO you must hold to reach each cashback tier is now priced and charged. At $105/GNO that is $10.50, $105, $1,050 and $10,500 of locked capital by tier, against which the engine applies its standard 15% annual haircut ($0.13 to $131.25 a month). Until now staking_required_usd was null, so the engine treated a 100-GNO holding as free. Effect: Gnosis Pay no longer ranks first in the EU at any spend level tested — it falls to 3rd through 5th — and its computed rate at $1,000/month drops from 3.75% to 1.62%, because the engine now selects the 1-GNO tier rather than a 100-GNO tier whose haircut swamps the extra two points of rate. Price supplied by the site owner and corroborated the same day at $105.06 on metamask.io/price/gnosis, with other trackers in the $104-111 range. This is a snapshot and the field most likely to go stale: GNO has held roughly $105-120 for three months on a slow 2026 decline from $140+ in January, and a 20% move changes which tier the engine picks. The OG NFT bonus that would lift each tier by a point stays unmodelled, deliberately — it is earned by spending a set amount within 30 days of activation, so the base rate is the honest default. |
| ether.fi Cash | added | ether.fi relaunched its app as a full neobank on 2026-08-14: tokenized stock and metals trading, an integrated Aave market on Optimism for borrowing against a whole portfolio at around 4%, fiat on and off ramps in 30+ currencies, and new deposit methods including Cash App and Apple Pay. Tokenized stocks and metals are unavailable to US users. None of it changes this card's cashback or fees, so no figure here moved — logged because the product around the card changed shape, and because the coverage independently restates the card's rate as 3%. Worth noting how that 3% is reported: press coverage gives it flat, while ether.fi's own help centre gives 3% only on the first $2,000/month, then 1%, then 0.5%. The banded figure is the one modelled here. Deposit methods and currency support are not captured in this dataset's schema and were not added on the strength of an announcement — the funding field still reads USDC pending a primary-source check. |
2026-08-11KAST now has a real referral code, so its code page carries one for the first time. Tria's code page moved to /tria-access-code, and both cards' referral terms were re-verified against the issuers' own pages.
| Card | Type | Detail |
|---|---|---|
| KAST Card | added | Referral code OWS9Q04D and link app.kast.xyz/referral/OWS9Q04D added, generated from my own KAST account and confirmed by an in-app screenshot. This is the in-app 'Refer a Friend' program, which needs no application — not the separate KAST Affiliate Program, which I still haven't applied to and which would give you 20% off a first subscription instead of 10%. I'm telling you which one you're getting because the difference is yours, not mine. |
| KAST Card | corrected | Referee benefit restated in dollars from KAST's own referral page: $100 off a Premium membership, $1,000 off Private. Previously carried as '20% off first subscription (vs. 10% under the plain in-app referral)' — not wrong, but the affiliate figure was leading on a page where the plain referral is what's actually on offer. The dollar figures are exactly 10% of Premium's $1,000/yr and Private's $10,000/yr, so the two records reconcile. |
| KAST Card | corrected | Referrer benefit set to the in-app figure: up to $1,250 per referral ($20/$30/$50/$150 across the referee's $100/+$500/+$1,000/+$5,000 spend milestones, plus $100 or $1,000 if they buy a membership), replacing the affiliate-program tiers I hadn't earned. Disclosed on the page in full. |
| Tria Card | corrected | No number changed. The referral page moved from /tria-referral-code to /tria-access-code, matching what Tria's own signup calls it, with a permanent redirect from the old URL. Cashback bands and membership fees were re-verified against Tria's Season 3 post the same day and were already correct: Virtual 1.5% to $100/mo then 0.5% ($25/yr), Signature 4.5% to $1,000/mo then 1% ($109/yr), Premium 6% to $2,000/mo then 1% ($250/yr). |
2026-08-02Correction: Crypto.com's tier unlock costs were modelled wrong in both directions, which overstated the card in the rankings. Found during a pre-launch audit, fixed same day.
| Card | Type | Detail |
|---|---|---|
| Crypto.com Card | corrected | Pro tier: the $29.99/mo subscription needed to unlock it was documented in the tier's own notes but never entered as data — monthly_fee_usd and staking_required_usd were both null, so the engine charged nothing to unlock a paid tier and ranked Pro at a full 3.00% / $30/mo at $1,000/mo spend. monthly_fee_usd set to 29.99. Pro now computes to $0.01/mo at that spend — essentially break-even, which is the honest answer for a $29.99/mo subscription buying 3% capped at $75/mo. It only pulls ahead above roughly $1,000/mo in spend. |
| Crypto.com Card | corrected | Plus tier: the opposite error. Unlocking Plus costs $4.99/mo OR a $500 12-month CRO stake — either/or — but the data carried both, so the engine charged the subscription AND a $6.25/mo staking haircut on top. staking_required_usd set to 0; the cheaper of the two paths ($4.99/mo) is what's modelled. The stake alternative is still real, just not the route a rational user picks. |
| Crypto.com Card | corrected | Net ranking effect: Crypto.com's best tier at $1,000/mo in the US is now Plus at 1.50% / $15.01, not Pro at 3.00% / $30. The card drops from 5th to joint 6th. No cashback percentage changed — only the cost of getting the tier, which was always published and simply hadn't been modelled. |
Citing this edition
Everything here is CC-BY-4.0 — use it, quote it, just link back. Suggested citation: "Crypto Card Report, August 2026 edition (https://www.cryptocardreport.com/report/2026-08)". The underlying dataset is machine-readable at /api/cards.json, no key required, and every figure above traces to an issuer's own documentation with an access date recorded per field.