Microsoft discloses a four-tier roadmap for Windows AI PCs, revealing in advance the codenames of AMD and Intel's next-generation chips

More Than Just an NPU Upgrade: Microsoft Teases the Next-Generation AI PC Chip Lineup Source: Semiconductor Horizons Published: 2026-11-04 18:29 Microsoft chip design generated poster Microsoft recently disclosed a Windows AI PC hardware roadmap involving multiple next-generation processor codenames from vendors that have not yet been officially announced, including Ryzen AI 500 and Nova Lake.

Recently, Microsoft disclosed a Windows AI PC hardware roadmap. Using local AI computing capability as the yardstick, it divides Windows PCs into four capability tiers, covering everything from entry-level models with 8GB of memory priced at about $500 to workstation-class supercomputing systems equipped with 512GB of single-card VRAM.

Notably, this roadmap also "accidentally" leaked the codenames of next-generation processors from vendors such as AMD, Intel, Qualcomm, and Nvidia that have not yet been officially announced.

Four-Tier Classification: Memory Capacity Becomes the New Yardstick for AI PCs

According to this roadmap, Microsoft for the first time has put the concept that "Windows is the home of hybrid intelligence" into concrete hardware classification standards. And the four tiers of the entire classification are almost entirely bounded by a single dimension: memory capacity.

Tier 0 is positioned as the entry-level AI PC, with 8GB of memory as standard, corresponding to affordable thin-and-light laptops in the roughly $500 range on the market, capable of handling basic context awareness and simple local AI functions; the compatible hardware lineup covers Intel Core Series 3/4 (Wildcat Lake), AMD Ryzen 100/200 series, and the full Qualcomm Snapdragon C/X/X Plus platform lineup.

Tier 1 is the mainstream Copilot+ PC in the current market, with memory ranging from 16GB to 64GB. Microsoft directly spells out the actual experience corresponding to different memory capacities in very specific terms: 16GB is just enough to run basic AI functions such as "quick Q&A," 32GB can enable "task delegation," and 64GB is the threshold for supporting long-running continuous AI workflows and heavy, complex tasks. Hardware in this tier already on the market includes Intel Core Ultra Series 3 (Panther Lake), AMD Ryzen AI 400 (Gorgon Point), and Snapdragon X2 Plus/Elite, and will later add the next-generation platforms Ryzen AI 500 and Core Ultra Series 4 that were revealed ahead of schedule this time.

Tier 2 targets the developer community and requires no less than 128GB of memory, with the core capability requirement being "completing AI model training, fine-tuning, and deployment services locally."

Tier 3 targets high-end workstation and desktop supercomputing scenarios, with the core goal of achieving "smooth fully local operation of frontier large models."

Two years ago, the hard threshold Microsoft set for Copilot+ PCs was that a device's NPU computing power had to reach 40 TOPS. In this new roadmap, however, the core yardstick for classification has quietly shifted to memory capacity and the scale of models a device can run. The TOPS computing power metric has not been abolished, but it is no longer the core basis for dividing AI PC capability tiers.

New Product Lineup: The Next-Generation Chips Microsoft Revealed in Advance

The most eye-catching content of this roadmap is the advance disclosure of next-generation product codenames from multiple chip vendors that have not yet been officially announced:

AMD Ryzen AI 500 (Medusa Point): a new-generation mobile platform based on the Zen 6 architecture, positioned in the mainstream AI device segment. Engineering sample information has already surfaced, showing a 10-core, 20-thread design, boost frequencies up to 5.37GHz, a single-core score of 3329 and a multi-core score of 16555, representing a very considerable performance improvement over the previous generation Strix Point.

Intel Core Ultra Series 4 (Nova Lake): expected to feature the sixth-generation NPU 6 AI engine, with INT8 computing power of up to 74 TOPS, and native support for DDR5-8000 memory and the PCIe 5.0 interface standard.

Snapdragon Next: the next-generation Snapdragon PC chip positioned ahead of the current X2 series, about which very little information has been disclosed so far.

In addition, the roadmap also mentions Intel Core Ultra Next, as well as AMD's next-generation hardware for which specific details have not yet been announced; these products will together form a high-end hardware matrix for running local large models.

As of press time, neither AMD nor Intel has officially confirmed the formal commercial names of their next-generation client processors. This amounts to Microsoft directly placing the product names that the two vendors had kept tightly under wraps onto the table in advance in an internal presentation material. Considering Windows' core dominant position in the x86 and Arm ecosystems, this seemingly accidental "information jump-start" is essentially not an ordinary leak, but rather the platform party's public endorsement of the entire industry's silicon roadmap—Microsoft, using its own AI PC classification system, has directly brought its partners' product rollout pace for next year into its official ecosystem planning framework ahead of time.

Looking Ahead to 2027: The Golden Window for the AI PC Chip Market

Combining the roadmap details with existing industry leaks, the period from late 2026 to 2027 will be a window of concentrated, intensive iteration for AI PC hardware. Intel's Nova Lake is expected to officially enter mass production in the fourth quarter of 2026; engineering samples of AMD's Zen 6 mobile version, namely Ryzen AI 500, have entered the actual testing stage, with an expected release time from late 2026 to 2027; Qualcomm's Snapdragon X2 series is also steadily accelerating its R&D pace, and its NPU computing power may reach a new height of 80 TOPS.

By disclosing this tiered roadmap at this time, Microsoft has two core objectives: on the one hand, to set a unified hardware admission baseline for AI functions across the entire Windows system; on the other hand, to convey clear product upgrade expectations to OEM manufacturers, downstream channels, and ordinary consumers.

The next stage of competition for Windows AI PCs will no longer remain at the level of simply stacking NPU computing power TOPS figures, but will instead reorganize the product lineup of the entire chip ecosystem around a new yardstick of "memory capacity × runnable model scale." AMD's Zen 6, Intel's Nova Lake, Qualcomm's Snapdragon Next, and Nvidia's RTX Spark have all been incorporated by Microsoft into the same ecosystem rollout timetable.

This article is reproduced from media reports or online platforms and represents the author's personal stance or views. Our reproduction is solely for sharing and does not mean that we approve of or agree with it. If the source attribution is incorrect or your lawful rights and interests are infringed, please contact customer service promptly, and we, as a neutral platform service provider, will promptly correct, delete, or handle it in accordance with the law. Comments No user comments yet Latest News More Related Recommendations

DRAM prices hit a new high again in October; institutions: the price increase trend will continue in the fourth quarter 2026-11-04 18:12

DeepSeek bets on Ascend, a year of breakthrough for domestic AI chips 2026-11-04 17:59

The world's second-largest photoresist company is about to be sold 2026-11-04 17:56