Gemini API suddenly returns "User location is not supported" after enabling Billing, and AI Studio cannot list projects

Hello,

I encountered an issue with the Gemini API today.

Timeline:

  1. I created a new Google Cloud Project.
  2. Enabled Billing successfully.
  3. Created Gemini API keys.
  4. Initially, everything worked correctly:
    • Google AI Studio could access my project.
    • API keys were created successfully.
    • Gemini API requests worked.

About one hour later, the following issues suddenly appeared:

  1. Google AI Studio now shows:

Failed to list imported projects: permission denied.
Please try again.

  1. Both Gemini models return the following error:

Error 400
User location is not supported for the API use.
FAILED_PRECONDITION

Environment:

Could you please help determine whether this is a temporary backend issue, an account region restriction, or a project permission problem?

Thank you.

Are you using a VPN or a VPS? Some VPSs and VPNs are banned by Google Cloud services.

Hi, thanks for your reply.

Yes, I am using a VPN with a US IP address (Los Angeles).

I checked my IP using ipinfo .io, and it shows:

  • Country: United States
  • City: Los Angeles
  • ISP: HostPapa

The strange part is that everything worked normally for about one hour after I enabled billing and created the API key. Then suddenly:

  • Google AI Studio started showing “Failed to list imported projects: permission denied.”
  • Both Gemini 2.5 Flash and Pro APIs began returning:
    User location is not supported for the API use.

I haven’t changed the VPN or my IP during this time.

Do you think this IP might have been blocked, or could this be an issue with my Google account or project?

Since it worked initially and then stopped without any configuration changes, I’m wondering if this could be a temporary backend issue or a false regional detection.

I can’t comment much on the VPN usage, but it is possible that your connection leaked your source IP, which is not supported by Gemini during the authentication. My advice is to tunnel all your connections to the VPN or try changing the VPN server. “Technically,” users are not allowed to access Gemini via a VPN.

Thank you for your reply.

I am using a US VPN and my public IP is located in Los Angeles (verified via ipinfo..io).

Everything worked correctly for about one hour after enabling billing and creating the API key, then it suddenly started returning “User location is not supported” and AI Studio could no longer list my projects.

I will test another VPN server and also check for possible DNS/IP leaks. Thank you for the suggestion.

Hello @yixin ,
This could be specific to your IP address. Please report it on our Report IP problems page. Suggested workarounds which may help for now:

  • Try temporarily switching to a different network (e.g. try with the hotspot from your phone if available).

  • Try asking your host for a new IP.

  • Trying switching between IPv6 / IPv4.

All API keys and models in project gemini-abc-50590.. return 400 FAILED_PRECONDITION "User location is not supported", while the outbound IP is in a supported region (Tokyo, JP)

Body

Copy

Environment
-----------
Project name:    gemini ABC
Project ID:      gemini-abc-50590.
Project number:  5491978302..
Billing:         prepay, active, balance normal, no payment issues
Client:          Node.js 24.18.1 / Electron 43.3.0, HTTPS requests via a local proxy
Affected models: all (tested gemini-3.6-flash); also /v1beta/models
Affected routes: /v1beta/models, :generateContent, :streamGenerateContent

Symptoms
--------
1. Any request with a valid key returns:

   HTTP 400
   {"error":{"code":400,
             "message":"User location is not supported for the API use.",
             "status":"FAILED_PRECONDITION"}}

2. Google AI Studio shows:
   "Failed to list projects: permission denied. Please try again."
   The API key list is therefore empty as well.

3. This affects both an older key (created 2026-08-18) and a key created
   today in the same project — identical error for both.

Troubleshooting already performed
---------------------------------
Network / egress (all verified, not assumed)
- Outbound public IP confirmed as Tokyo, Japan:
  14.137.235.33 (AS2497 IIJ) and later 216.23.122.13 (AS205548).
  Japan is a supported region per the available-regions documentation.
- Local proxy logs confirm generativelanguage.googleapis.com is matched by
  the Google routing rule and exits through the Japan node — it is not
  falling through to DIRECT.
- No public IPv6 on the host, so there is no IPv6 leak revealing a different
  region. Only one proxy is present.
- For comparison, on the same node and at the same moment:
  www.google.com, ai.google.dev and api.github.com all return HTTP 200.
  Only the Gemini endpoint fails.

Account and billing region (every location signal resolves to a supported region)
- The Google account's home address and work address are both in Japan.
- The payments profile is denominated in JPY (Japan).
- Together with the Tokyo request IP, every location signal Google can
  observe for this account and project is inside a supported region.
  A genuine geographic restriction therefore cannot be what this
  precondition is reporting.

The credential is a valid API key, not an OAuth token
- Request with no key            -> 403 "Method doesn't allow unregistered callers"
- Request with an invalid AIza.. -> 400 INVALID_ARGUMENT "API key not valid"
- Our key                        -> 400 FAILED_PRECONDITION location error
  i.e. the key passes identity validation and is then rejected on location.
- `Authorization: Bearer <key>`  -> 401 UNAUTHENTICATED
- `?access_token=<key>`          -> 403 unregistered callers
  Both confirm it is an API key (a service-account-bound auth key), not an
  OAuth 2 access token.

Key restrictions are not the cause
- The key is restricted to one API: "Gemini API".
- Application restrictions are set to "None".
- The console does not flag either restriction as misconfigured.

Other standard remedies already tried
- Created a brand new API key in the same project — same error.
- Switched to a different egress node in the same country — same error.
- Billing balance is normal and the billing account is active.

Why I believe this is not a location issue
------------------------------------------
The identical combination — AI Studio "Failed to list projects: permission
denied" plus API 400 "User location is not supported", with a working
configuration and a supported-region IP — has been reported by several other
users of this forum, from locations that are unambiguously supported:
Los Angeles (US), Frankfurt (DE) and Poland (EU). In those threads the same
two remedies (new key, new outbound IP) also failed.

That pattern points at a backend/billing-state problem rather than an actual
regional restriction. The error message appears to be a misleading
catch-all for more than one precondition failure.

Reference threads
- /t/gemini-api-suddenly-returns-user-location-is-not-supported-after-enabling-billing-and-ai-studio-cannot-list-projects/172717
  (tagged "bug"; identical symptom pair, reporter egress in Los Angeles)
- /t/gemini-api-suddenly-returns-user-location-is-not-supported-worked-fine-for-a-week-broke-suddenly-persists-even-with-a-brand-new-api-key/176357
  (prepay billing, worked ~1 week, new key and new outbound IP both ineffective)

Request
-------
Could someone from the team please check the account/project state for
project number 5491978302..?

Specifically:
- whether the billing account behind this project is in a state that makes
  the Gemini API reject every key, and
- whether the misleading "User location is not supported" message can be
  replaced with an accurate diagnostic, since it currently sends developers
  into long, unproductive network troubleshooting.

Thank you.

Hello @lanxin ,

This could be specific to your IP address. Please report it on our Report IP problems page. Suggested workarounds which may help for now:

  • Try temporarily switching to a different network (e.g. try with the hotspot from your phone if available).

  • Try asking your host for a new IP.

  • Trying switching between IPv6 / IPv4.