Skip to content

Est. 2011ยทMicrosoft Partner ID 7033487ยทDelivery under 3 minยทSupport 7 days a week

Your vault is empty.

Free Fix 0x80090325

Remote Desktop cannot verify the identity: certificate error 0x80090325

11 min read Updated October 5, 2026 Networking, Sharing & Printing

Fix it now

0x80090325 is SEC_E_UNTRUSTED_ROOT: the Remote Desktop client built a chain for the host’s certificate and it ended in a root this machine does not trust. By default a Windows host signs its own Remote Desktop certificate, so the warning is correct until you either issue the host a certificate from an authority the client already trusts, or trust that specific certificate deliberately.

Run these on the Remote Desktop host, elevated; the first in Command Prompt, the second in PowerShell

certutil -store "Remote Desktop"
Get-CimInstance -Namespace Root\CIMv2\TerminalServices -ClassName Win32_TSGeneralSetting | Format-List SSLCertificateSHA1Hash,SecurityLayer,UserAuthenticationRequired
  1. Read the subject and the validity dates the first command prints. Then connect using exactly the name on that certificate – an IP address will never match a certificate issued for a host name.
  2. On a domain, issue the host a certificate from your own certification authority using a template that allows server authentication. Import it into the computer account’s Personal store first – Microsoft documents that binding a certificate that is not there returns an Invalid Parameter error – then bind it by writing its SHA1 thumbprint into SSLCertificateSHA1Hash on Win32_TSGeneralSetting.
  3. Confirm the client trusts the issuer: open certlm.msc on the client and look for your authority under Trusted Root Certification Authorities.
  4. If the certificate has simply expired, delete it from the host’s Remote Desktop store and restart the Remote Desktop Services service so the host generates a fresh self-signed one.

Turning the client’s server-authentication warning off does not fix anything. The session stays encrypted either way; what you lose is the only signal that would tell you something else is answering on that address.

If the warning is gone rather than dismissed, you are done. If not, the next section separates the five different things this dialog is actually reporting.

Why it happens

Remote Desktop server authentication is an ordinary certificate check. The client builds a chain from the certificate the host presents up to a root, and it wants three things to be true: the chain ends in a root this machine trusts, nothing in the chain is outside its validity period, and the name that was dialled appears on the certificate. Each of those failures has its own code, and the dialog wording is nearly identical for all of them, which is why people fix the wrong one.

A Windows host with no certificate of its own makes one. That certificate is issued for the machine’s own name and signed by nobody, so no client anywhere has a reason to accept it. For a machine you can physically identify, accepting it once is a defensible risk. For routine access it is not, because an untrusted certificate is exactly what an impersonating host would also present.

The companion codes come from two different layers. 0x80090327 and 0x80090328 are SSPI reporting on the certificate itself – an unspecified processing error, and an expired certificate. 0x800B0108 and 0x800B0111 come from the chain engine: a certificate missing a required field, and a certificate that somebody has explicitly marked as untrusted on this machine. That last one is worth knowing, because it is the only code here that is not an accident.

The host is presenting its own self-signed certificate

You have this one if The subject and the issuer on the certificate are both the host’s own name, and every client warns, including ones that have never connected before.

  1. On a domain, request a certificate from your internal certification authority using a template that permits server authentication.
  2. Import it into the Personal store of the computer account with certlm.msc. Microsoft documents this as a prerequisite: bind a certificate that is not in that store and the operation fails with Invalid Parameter.
  3. Bind it by thumbprint: set SSLCertificateSHA1Hash on the Win32_TSGeneralSetting instance in the Root\CIMv2\TerminalServices namespace.
  4. Make sure the authority’s root certificate reaches every client’s Trusted Root Certification Authorities store. Domain members get this through Group Policy already.
  5. Reconnect and confirm the certificate now shows your authority as the issuer.

Importing a host’s self-signed certificate into a client’s trusted roots makes that client trust anything the certificate signs. Do it for named machines you control, not as a habit.

You are connecting by a name the certificate does not carry

You have this one if The fully qualified name connects cleanly and the short name or the IP address does not.

  1. Read the subject and the subject alternative names on the host’s certificate.
  2. Connect using one of those names exactly, and update saved .rdp files and shortcuts so nobody keeps using the old address.
  3. Where people legitimately connect by several names, reissue with every one of them listed as a subject alternative name.

The certificate has expired, or the clock says it has

You have this one if 0x80090328, and the connection worked without complaint until a specific date – or one of the two machines has a wrong clock and nothing expired at all.

  1. Check the validity dates on the host’s certificate first.
  2. For a self-signed certificate, delete it and restart the Remote Desktop Services service so a new one is generated.
  3. For an issued certificate, renew it and rebind the new thumbprint before removing the old certificate.
  4. If the dates look fine, compare both clocks in UTC and run w32tm /resync on the machine that is wrong.

The chain is incomplete, or the root never reached this client

You have this one if The certificate came from your own authority, domain members accept it, and machines outside the domain do not.

  1. On the host, confirm the intermediate certificates are present in Intermediate Certification Authorities.
  2. On a failing client, open the certificate from the warning and read the Certification Path tab to see where the chain stops.
  3. Distribute the root, and any intermediates, to non-domain clients through your management tool or by import.

The certificate has been explicitly distrusted on this client

You have this one if 0x800B0111, and one client refuses a certificate that its neighbours accept without complaint.

  1. On the failing client, open certlm.msc and look in the Untrusted Certificates store.
  2. If the host’s certificate or its issuer is sitting there, someone or something put it there deliberately. Find out why before removing it.
  3. Remove the entry only when you know the distrust was a mistake, then reconnect.

Explicit distrust is not a chain failure. Microsoft’s wording is that the certificate was marked as untrusted by the user, so treat it as a decision somebody made rather than as damage.

Full reference

Where each thing lives on the host

What you need Where it is
The certificate the host presents The Remote Desktop store in the local computer certificate store; certutil -store "Remote Desktop" prints it
The certificate currently bound to RDP SSLCertificateSHA1Hash on Win32_TSGeneralSetting, in the Root\CIMv2\TerminalServices namespace
Whether network level authentication is demanded UserAuthenticationRequired on the same class; 1 means the client must authenticate before a session is created
Which security layer is negotiated SecurityLayer on the same class
Trusted roots on the client certlm.msc, Trusted Root Certification Authorities
Certificates somebody has distrusted certlm.msc, Untrusted Certificates – the store behind 0x800B0111

Binding by thumbprint is the part worth getting right. Microsoft documents SSLCertificateSHA1Hash as the SHA1 hash, in hexadecimal, of the certificate the target server should use, and notes that you can read a thumbprint from the Details tab of the certificate properties page in the Certificates snap-in. Putting a certificate in the store is not enough on its own; something has to point Remote Desktop at it.

Reading the dialog properly

Code Constant What it actually says
0x80090325 SEC_E_UNTRUSTED_ROOT The chain was issued by an authority that is not trusted
0x80090327 SEC_E_CERT_UNKNOWN An unknown error occurred while processing the certificate
0x80090328 SEC_E_CERT_EXPIRED The received certificate has expired
0x800B0108 CERT_E_MALFORMED A required field such as the subject or issuer name is missing or empty
0x800B0111 TRUST_E_EXPLICIT_DISTRUST The certificate was explicitly marked as untrusted by the user

Network level authentication and this error

Where the host requires network level authentication, the client has to authenticate before a session exists, and a certificate problem is met earlier and more bluntly than it would be otherwise. That is a good thing, but it means the same underlying certificate fault can surface either as an identity warning you can click past or as a refusal you cannot, depending on how the host is configured. Check UserAuthenticationRequired before concluding that two machines are behaving differently for some deeper reason.

When none of the five causes fits

  • Confirm you are looking at the certificate the host is actually using. A host can hold several certificates in that store while RDP is bound to a different one; the thumbprint in SSLCertificateSHA1Hash is the authoritative answer.
  • Check whether the machine was renamed. Renaming does not reissue a certificate, and the old name stays on it until you delete the certificate and let the service build a new one.
  • Check whether a connection broker, gateway or load balancer is answering instead of the host you named. The name on the certificate you are shown tells you which machine really replied.
  • On a non-domain client, remember that trust decisions are per machine. A certificate imported into a user’s own store does nothing for a service or a scheduled task on the same machine.

Public certificates, internal names

A public certification authority will not issue for an internal-only name, and it does not need to: your own clients already trust your own authority, and you control the renewal. Public certificates belong on names that people outside your organisation resolve and connect to, which in practice means a Remote Desktop Gateway rather than an individual session host. Buying one for an internal host is a cost with no benefit, and it usually still leaves you binding it by thumbprint.

Every code this article covers

Code What it points at Source
0x80090325 SEC_E_UNTRUSTED_ROOT: the certificate chain was issued by an authority that is not trusted Microsoft Learn
0x80090327 SEC_E_CERT_UNKNOWN: an unknown error occurred while processing the certificate Microsoft Learn
0x80090328 SEC_E_CERT_EXPIRED: the received certificate has expired Microsoft Learn
0x800B0111 TRUST_E_EXPLICIT_DISTRUST: the certificate was explicitly marked as untrusted by the user, not rejected by chain building Microsoft Learn
0x800B0108 CERT_E_MALFORMED: a certificate is missing or has an empty value for an important field, such as a subject or issuer name Microsoft Learn

Confirm the fix worked

  1. Connect using the name on the certificate and confirm no identity warning appears at all, rather than being dismissed.
  2. Open the certificate from the connection and confirm the issuer is your certification authority, not the host itself.
  3. On the host, confirm SSLCertificateSHA1Hash matches the thumbprint of the certificate you intended to bind.
  4. Note the expiry date and put a renewal in the calendar; certificate expiry is entirely predictable and should not surprise you twice.
  5. Test from a second client, ideally one outside the domain, so you know the trust chain reaches everywhere it needs to.

Questions people ask about this

Does fixing this cost anything?

No. A domain with Active Directory Certificate Services issues Remote Desktop certificates at no extra cost, and a host with no certificate makes its own for free. Money only enters the picture if you choose a public certification authority for a gateway that people outside your organisation connect to, which is a different scenario from an internal host.

Can I just tick the box that stops it asking?

You can, and the session stays encrypted. What you give up is the ability to notice that something else has started answering on that address. On a network you physically control that may be an acceptable trade; across the internet it is not.

Why did this start after we renamed the server?

Because the certificate still carries the old name and renaming a machine does not reissue it. Delete the self-signed certificate and restart the Remote Desktop Services service, or reissue from your authority with the new name and rebind the thumbprint.

One laptop refuses a certificate every other machine accepts. Why?

Look at 0x800B0111 first. That code means the certificate has been explicitly marked as untrusted on that machine, which is a per-machine decision recorded in the Untrusted Certificates store rather than anything wrong with the certificate itself.

Is a certificate warning the same as an insecure connection?

No, but it is the absence of a guarantee. The transport is encrypted either way. What the check gives you is confidence about which machine is on the other end, and an untrusted certificate withholds exactly that.

Related error codes

Was this article helpful?

Your feedback helps us improve our documentation.

Related articles

Free Fix VPN error 798 and 13806: no usable certificate for IKEv2 or EAP logon Free Fix Error 0xC000006D: logon failure when connecting to a server or share License Error Error 0x80070047: no more connections can be made to this remote computer License Error Scan to folder fails with 0xC00000CC: the copier cannot reach your share
โ† Back to Knowledge Base