# Microsoft brings AI inference closer to enterprise data with Windows Server 2028

> Source: <https://www.networkworld.com/article/4250500/microsoft-brings-ai-inference-closer-to-enterprise-data-with-windows-server-2028.html>
> Published: 2026-11-19 02:25:03+00:00

**Key takeaway:** Windows Server 2028's public preview lets enterprises run inference on-premises via AI accelerators such as AMD's Instinct MI350P (preview in the first half of 2027) and support for open-weight and custom models across Foundry Local, Windows ML, vLLM and Llama.cpp.

Microsoft is betting on a hybrid future where AI inferencing takes place where it makes the most sense. Which, logically, is where data already lives: On enterprise-controlled infrastructure.

To that end, this week at [Microsoft Ignite](https://ignite.microsoft.com/en-US/home), the tech giant announced Windows Server 2028, its largest platform update since Windows Server 2025.

Now in public preview, the [next generation server](https://www.networkworld.com/article/4206611/agentic-ai-could-force-a-rethink-of-enterprise-ai-server-design-researchers-say.html?utm_source=chatgpt.com) is positioned as a foundation for AI inference that supports “enterprise hybrid intelligence.” It introduces new AI accelerators so enterprises can run inference on their own infrastructure, storage tools for accelerated data transfer, security controls to establish a “secure baseline from day one,” and simplified admin and management dashboards.

But while it promises new capabilities across automation, management, identity, security, recovery, storage, and virtualization, it’s still early days for Windows Server 2028, with various features at different levels of availability and maturity.

“It’s important to remember that this is the public preview phase, so a lot still remains a promise on the roadmap,” noted Infotech associate VP [John Annand](https://www.infotech.com/profiles/john-annand).

Windows Server 2028 gives customers the ability to run inferencing workloads close to physical operations and in their own controlled ecosystems. This means they don’t have to build and operate a “separate AI stack for every location,” Windows Server product team leader [Neil Hinnant](https://www.microsoft.com/en-us/windows-server/blog/author/neil-hinnant/) explained in a [blog post](https://techcommunity.microsoft.com/blog/windowsservernewsandbestpractices/windows-server-2028-a-trusted-foundation-for-an-evolving-enterprise/4574112).

Microsoft is introducing AI accelerators in its servers; notably, AMD’s [Instinct MI350P](https://www.amd.com/en/blogs/2026/more-ai-more-control-amd-instinct-mi350p-comes-to-windows.html) card is be in preview in the first half of 2027 for supported Windows Server systems, including Server 2028. The company is also developing tools to manage that hardware, deploy and configure [inference environments](https://www.networkworld.com/article/4122439/microsoft-launches-its-second-generation-ai-inference-chip-maia-200.html?utm_source=chatgpt.com), and direct model requests to appropriate environments.

Microsoft said the offering also begins its long-term investments to extend [Microsoft Foundry](https://www.infoworld.com/article/4245368/exploring-microsoft-ai-foundation-models.html?utm_source=chatgpt.com) to “fleets of local token factories,” and to support open-weight and custom models across Foundry Local, Windows ML, vLLM, and Llama.cpp.

“As AI-accelerated workloads diffuse across the enterprise, different workloads have different needs,” Hinnant wrote. “Cost, latency, IP protection, data residency, and regulatory requirements all shape that decision.”

From a security standpoint, Windows Server 2028 has a hardened Server Core configuration “by default,” with hundreds of security settings enabled on first boot, aligned with industry benchmarks.

“The secure configuration is the starting state, not a project you run after setup,” [Microsoft explained](https://learn.microsoft.com/en-us/windows-server/get-started/whats-new-windows-server-2028). Enterprises need only relax the settings their workloads need. “That closes the window where a newly deployed server sits exposed while it waits to be hardened.”

The new server incorporates a containment layer for isolation, Agent 365 for agent identity and governance, and virtual machine (VM) protection to build a chain of trust. Microsoft says it is also investing in automatic certificate management to verify both human and machine identities, prevent compromise, and encrypt traffic, alongside post-quantum cryptography. This can balance day-to-day security tasks with longer term needs to modernize “against the threats of tomorrow.”

Also by default, Windows Server 2028 supports native non-volatile memory express (NVMe) storage to “give demanding workloads more room to grow.”

Enterprises need to be able to apply policy, manage updates, and get a view of what’s running across their estates. Intune management for Windows Servers, now in preview, brings endpoint management workflows including configuration, compliance, and updates to eligible and always-on servers. Windows Admin Center Virtualization Mode (WAC vMode), now in general availability on Windows Servers, gives admins further ability to work with specific servers and infrastructure “as one fabric,” Hinnant explained.

“When a workload needs attention or a virtualization host needs troubleshooting, administrators need to get close to the machine and understand what is happening,” he said.

The hybrid intelligence capabilities are available to a limited number of customers through an early-access program, which includes hardware requirements, Microsoft said. Customers interested in early access to Windows Server 2028 can apply through the [Windows Server Technology Adoption Program](https://aka.ms/WSTAP) (TAP). NVMe and Intune management for servers are available for testing now.

If customers want to use Server 2028 for local AI on existing hardware, they would have to add Nvidia or AMD accelerators (there’s a lot of interest in AMD’s ROCm open-source GPU), Infotech’s Annand noted, not to mention adding “lots and lots of ever more expensive RAM,”. Further, the 10GbE network cards that were “blazingly fast” 3 years ago will likely be a bottleneck until they upgrade to 50 or 100GbE+.

“To the surprise of no one, if you want to use Server 2028 for AI, you’ll need to invest in a new kit,” Annand said.

Many enterprises are more interested in the quality-of-life improvements than the AI hype around platform scaffolding, he said. For these customers, Server 2028 is a “viable midpoint upgrade for existing hardware.”

Microsoft’s value proposition is ecosystem, or at least product bundling, and enterprise control, Annand said. The power of AI on Server 2028 will come from integrability with Microsoft Foundry and other Azure cloud services. Local inference, from both a security and cost control standpoint, is a hot button topic in the target Microsoft customer base.

Running AI in the public cloud has proven to be expensive, and implied in the Windows 2028 feature set released is a “tacit acknowledgement that some Microsoft customers want to return to on-prem from the blue skies of their Azure environments,” Annand said.

Microsoft emphasized that while it is moving toward general availability, Windows Server 2028 is a [prerelease product](https://learn.microsoft.com/en-us/windows-server/get-started/whats-new-windows-server-2028). Some elements may be “substantially modified” before its official release.

It’s also important to note that, because security and memory defaults will be locked down, some things that worked on Windows Server 2025 may not work on Windows Server 2028 without configuration changes. Problematic features could include older protocols, legacy authentication, and some management tools and scripts, Microsoft said. The company advised enterprises to test apps and operational tooling against these configurations before deployment, and to relax settings only when required.

Admittedly, not all enterprises will see a need to immediately upgrade to the new servers once they’re officially deployed, experts noted. After all, Windows Server 2025 will remain under extended support until 2034, and previous editions are also covered through Windows [Long-Term Servicing Channel](https://www.computerworld.com/article/1716419/faq-windows-10-ltsb-explained.html) (LTSC): Windows Server 2022 until Oct 14, 2031; 2019 until Jan. 9, 2029; and Windows Server 2016 sunsets on January 12, 2027.

Formerly known as the Long-Term Servicing Branch (LTSB), Windows’ LTSC allows companies to maintain a stable, feature-locked OS. The offering is particularly geared to mission-critical, highly regulated, or restricted environments.

Many AI pioneers don’t wait for Microsoft to make something enterprise safe or ready; they’ve been spinning up their own Linux clusters; working with Neocloud providers; or working in AWS Bedrock or Azure Foundry, Annand pointed out.

“One strategy is to let everyone else pay the highest prices and experience the most technical difficulties when something first comes out,” he said. Later adopters may have the short path from investment to profitability that AI pioneers are still searching for.

“Enterprise AI on Server 2028 will eventually be an option for the fourth wave of AI adopters,” who are “the extremely cautious ones,” he said.
