A standalone drill-down on Vietnam's new-user surge: where the growth comes from, whether the new users hold assets, and whether they convert to PayFi and keep paying. Facts and definitions only; open questions for the Vietnam team at the end. Domain: wallet (Bitget Wallet). Snapshot 2026-07-07. Sources: device attribution table (#872), Spend / scan-to-pay orders table (#4845).
Vietnam new devices jumped +142% Q1 to Q2 (30,338 to 73,492). The growth is almost entirely Android, accelerating in May and June (June Android 32,325). iOS, Chrome, and Telegram are flat or down. It is not referral-driven (only 6.8% of Q2 devices were invited).
The new users activate and return well (Q2 wallet activation 93.8%, next-day retention 22.4%, both higher than Q1), but they hold almost no assets: 71% have zero balance, 98.6% of those who do hold have under USD 10, median holding is USD 0, and only 1.9% ever did a Swap.
Where the growth goes is PayFi scan-to-pay. Vietnam scan-to-pay payers grew from 7,872 (Apr) to 40,219 (Jun), on the same scale as new devices. So the growth is scan-to-pay-driven, not trading or holding.
Are they paying continuously? Within a month, yes: 4 to 7 payments per active user, but tiny (median ticket about USD 1). Across months, no: only 22% of April's new payers paid again in May and 11% in June; the May cohort retained just 9% into June. Most pay for one month then churn. A small core (2% of payers doing 30+ payments) drives about 30% of all transactions.
## Where the growth comes from
### New devices by month and platform:
The growth is Android-only, ramping in May (18,276) and exploding in June (32,325). iOS holds flat at 2,000 to 3,000, Chrome is declining, Telegram is negligible. This is a mobile Android acquisition wave, not platform-wide growth.
### Q2 new devices by acquisition source:
The largest block (44,747) is untagged installs with no attribution source, and it is the lowest quality (only 0.4% hold >= USD 10). The second block is organic (28,104). Referral is negligible. So the surge is an untagged plus organic Android install wave, not a referral campaign.
## Do the new users hold assets?
### Q1 vs Q2 new-device funnel (current snapshot):
