Region codes
| Code | Where | Typical latency from major hubs |
|---|---|---|
US-East | Virginia, NJ, Toronto, Miami | NYC ~10ms, LON ~80ms |
US-Central | Chicago, Dallas, Iowa | NYC ~30ms, SFO ~45ms |
US-West | Oregon, California | SFO ~10ms, NYC ~60ms |
CA | Quebec, Vancouver | NYC ~15ms |
EU-West | Ireland, Netherlands, UK | LON ~5ms, NYC ~80ms |
EU-Central | Germany, Switzerland, Poland | FRA ~3ms |
EU-North | Sweden, Finland | STO ~5ms |
JP | Tokyo, Osaka | TYO ~5ms |
KR | Seoul | SEL ~5ms |
SG | Singapore | SIN ~3ms |
AU | Sydney | SYD ~5ms |
BR | São Paulo | GRU ~5ms |
How regions affect price
Same GPU often differs by ±20% across regions. US-East and EU-West are usually the cheapest because supply is densest. APAC and South America typically price 15–30% higher.How regions affect availability
H100s skew US/EU. Blackwell B200/B300 are scarce everywhere — if you need B200, US-East is your best bet. RTX 4090s are nearly everywhere.Specifying a region
Via dashboard: there’s a region filter chip in the Create page. Pick one or leave on “Any”. Via API:503 provider_pool_empty — no fallback to other regions. We don’t
silently relocate your workload.
Data residency
If you need GDPR-compliant EU-only processing, setregion to one of the
EU-* codes and we’ll only allocate from EU providers. Note that we
ourselves (the marketplace) operate from the US, so the session cookie and
ledger live in US-East. The compute itself stays where you specified.
For strict residency requirements (e.g. healthcare, public sector), email
business@gpuoutlet.ai — we have a few
specifically-certified providers we can pin you to.