Whirlwind Accelerator Completes Infrastructure Expansion
On September 22, 2026, Whirlwind Accelerator disclosed it had completed a new round of infrastructure expansion, adding six core data centers and raising backbone bandwidth by 60%. This is another step up in underlying resources after the earlier node expansion.
Active users grew markedly over the past two quarters, pushing peak-hour concurrent connections up. Some hot nodes showed latency fluctuation under high load, concentrating user complaints. This expansion aims to shore up peak-hour stability.
1. Expansion scope
The round adds six core data centers in Singapore, Frankfurt, Los Angeles, Tokyo, Hong Kong and São Paulo, and raises backbone bandwidth from 120Gbps to 192Gbps (+60%). The team also upgraded 18 edge nodes with new hardware.
| Item | Before | After | Gain |
|---|---|---|---|
| Core data centers | 22 | 28 | +27% |
| Backbone bandwidth | 120Gbps | 192Gbps | +60% |
| Edge node upgrades | — | 18 | New |
2. User pressure behind it
The team disclosed that Q3 peak-hour concurrent connections grew 45% year-over-year, with hot-node peak load once nearing the 85% safety threshold. Without expansion, peak-hour latency fluctuation would widen. A real case: the Tokyo node once queued connections during an esports event, and user feedback directly drove this project.
3. Direct user impact
After expansion, peak-hour latency fluctuation on hot nodes narrowed from ±18ms to ±6ms, and connection queues essentially disappeared. For ordinary users, the most tangible change is "no more node grabbing" at peak hours.
| Metric | Before | After |
|---|---|---|
| Peak-hour fluctuation | ±18ms | ±6ms |
| Hot-node peak load | 85% | 62% |
| Connection queues | Occasional | Gone |
Infrastructure expansion is an "invisible" job — users only feel its value at peak hours. For an accelerator company, investing in the invisible is exactly how a stability promise is honored. Download from the official site to test the post-expansion peak-hour experience.
User Comments
No more node switching at peak hours, much more stable.
Tokyo node no longer queues during events, timely expansion.
Peak-hour latency is steadier, hope for more nodes.
