Microsoft has patched a maximum-severity vulnerability in Azure AI Foundry that could allow unauthorized attackers to escalate their privileges over a network. Tracked as CVE-2026-85889, the flaw carries a CVSS score of 10.0, making it one of the most serious security issues affecting Microsoft’s cloud-based artificial intelligence services. The company said the vulnerability resulted from missing authentication for a critical function within Azure AI Foundry. There is currently no evidence that the flaw has been exploited in real-world attacks.
Azure AI Foundry, also known as Microsoft Foundry, is an enterprise platform used to build, deploy and manage generative AI applications and autonomous agents. A successful exploit could have enabled an unauthorized attacker to gain elevated privileges, potentially increasing access to sensitive resources or administrative functions. Security researcher Rémy Marot, who uses the handle @R_Marot, discovered and reported the vulnerability to Microsoft. The company has already addressed the issue across its cloud infrastructure.
Microsoft said customers do not need to take any action because the affected cloud services have been fully mitigated. Alongside CVE-2026-85889, the company also fixed several other critical cloud vulnerabilities. These include CVE-2026-85885, a command injection flaw in Microsoft 365 Copilot rated 9.9; CVE-2026-85878, an improper authorization issue in Azure Database for PostgreSQL rated 9.9; and CVE-2026-87701, an improper neutralization vulnerability in Azure Cosmos DB rated 9.6.
The tech giant separately released updates for two Windows vulnerabilities. CVE-2026-62721 affects the Windows User-Mode Power Service and could allow an authorized local attacker to gain SYSTEM privileges. CVE-2026-85921 is a double-free vulnerability in Windows Secure Kernel Mode that could enable privilege escalation to Virtual Trust Level 1. Both vulnerabilities were fixed through an out-of-band update for Windows 11 version 26H1, distributed as cumulative update KB5129194 for arm64 and x64 systems.
The latest fixes arrive shortly after Microsoft patched 974 vulnerabilities across its software portfolio. Two of those flaws, affecting Windows Advanced Local Procedure Call and the Windows Update Stack, were reportedly exploited in active attacks. Proofpoint and Volexity said the ALPC flaw was chained with two Google Chrome vulnerabilities to create the BlueMoon exploit kit, which multiple espionage-linked threat actors allegedly used to deliver malicious payloads. The developments highlight the growing security risks surrounding cloud platforms, enterprise AI tools and widely deployed operating systems.
Read the original article: