Windows 11 26H1: Built for New Silicon. Nvidia PC Chips Are Next

This Week in IT

This Week in IT

LISTEN ON:

Nvidia is about to crash the Windows‑on‑Arm party. Its first PC chips could arrive with a special Windows 11 build just for new silicon. At the same time, Mandiant just made it trivial to crack a decades‑old Microsoft protocol that some companies still run in production. And a new class of AI bugs shows how Copilots can be turned into cloud‑credential vending machines.

Thanks to Cayosoft for sponsoring this episode!

Links and resources

Episode overview

This Week in IT – Key Updates and Insights:

1. Nvidia entering Windows on Arm PC market

Nvidia is preparing to launch its first ARM‑based PC chips, called N1X, potentially shipping by the end of Q1. These chips are expected to coincide with a special build of Windows 11 (26H1) designed exclusively for new AI‑capable hardware, without new user‑facing features.

The architecture is based on Nvidia’s DGX Spark platform, combining ARM CPUs, Blackwell GPUs, and 128GB unified memory—similar to Qualcomm’s Snapdragon design. This creates new competition in the Windows‑on‑ARM ecosystem, offering potential benefits in performance, battery life, pricing, and native ARM software availability.

The host notes long‑standing rumors that AMD may also enter the ARM PC chip space, which could accelerate industry adoption and challenge Intel’s dominance—especially if Intel cannot match ARM battery efficiency.

2. Mandiant releases rainbow tables for NTLMv1

Mandiant unexpectedly released pre‑computed rainbow tables that make it extremely easy to crack NTLMv1, an outdated Windows authentication protocol. With only ~$600 of consumer hardware, an attacker could crack NTLMv1 in under 12 hours.

This was positioned as “tough love”: organizations should disable NTLMv1, audit systems still using it, and upgrade or isolate any applications relying on it. Microsoft already disables NTLMv1 by default on newer versions of Windows 10 and 11.

3. New vulnerabilities in MCP servers

Researchers identified several security vulnerabilities in Anthropic’s Git MCP server, including remote code execution and server‑side request forgery (SSRF). They also found an SSRF issue in Microsoft’s Markdown MCP server running on AWS when using Instance Metadata Service v1.

Because MCP servers act as translators between natural language and service APIs, their growing use makes them increasingly attractive to attackers seeking credentials or internal data.

Organizations are advised to:

  • Audit MCP server use
  • Sandbox execution
  • Restrict metadata access (e.g., upgrade AWS IMDS to v2)
  • Apply vendor patches promptly