Best Buy is running a back-to-school sale on AI-powered laptops with discounts reaching $900, but the timing is critical: memory prices are set to surge through 2027, making current inventory the last chance to buy at 2026 prices. Every device in the promotion was manufactured under older, cheaper memory contracts, and analysts project that replacement stock will cost significantly more once current supplies sell out.

Why Are AI PC Prices About to Jump?

The semiconductor industry is facing a DRAM (dynamic random-access memory) shortage that has driven costs to record highs. Gartner projects a combined 130% surge in DRAM and solid-state drive (SSD) prices by the end of 2026, which would raise average PC prices by 17% overall. TrendForce, another major market research firm, confirms that upward pressure on memory pricing will continue through the third and fourth quarters of 2026 and into 2027.

The devices on sale now were priced under memory contracts signed before the shortage hit. Once current inventory turns over, new laptops will reflect the higher component costs, making today's deals a finite window for buyers who need a new machine.

What Exactly Is a Copilot+ PC?

Copilot+ PCs are laptops and desktops certified by Microsoft to run a specific suite of on-device artificial intelligence features without requiring a cloud connection. The certification requires a minimum of 40 TOPS (tera operations per second) from a dedicated neural processing unit (NPU), at least 16 gigabytes of RAM, a 256-gigabyte SSD, and Windows 11.

The NPU is a specialized processor designed to handle AI inference, which is the computational work of running trained AI models. The Copilot+ features that unlock with this hardware include Windows Recall (which creates a searchable index of your screen activity), Click to Do (contextual actions on screen content), live captions with real-time translation across more than 40 languages, and enhanced Windows Studio Effects for video calls. One common misconception worth correcting: the Copilot chatbot that answers general questions remains cloud-dependent. The hardware certification unlocks the on-device Windows features, not an offline AI assistant.

Which Chips Power Today's AI PC Deals?

Three processor families dominate the current Copilot+ PC market, and all three clear Microsoft's 40 TOPS NPU floor comfortably. Qualcomm's Snapdragon X, AMD's Ryzen AI 400 series, and Intel's Panther Lake each take a different architectural approach, with trade-offs in battery life, software compatibility, and GPU-accelerated performance.

• Qualcomm Snapdragon X: Runs on ARM64 architecture, the same instruction-set family as every smartphone processor since 2009. Delivers exceptional battery life, often 18 to 20-plus hours of real-world use, because the RISC-based architecture uses significantly less power per clock cycle. The trade-off is software compatibility; Windows on ARM runs most consumer applications natively or through an emulation layer called Prism, but may fail with software that installs kernel-level drivers, certain VPN clients, professional audio and video tools, anti-cheat systems for games, and specialized engineering software.
• AMD Ryzen AI 400 Series: Delivers up to 12 Zen 5 CPU cores with clock speeds reaching 5.2 gigahertz, alongside RDNA 3.5 integrated graphics and LPDDR5X memory support. The x86 architecture means full compatibility with the entire Windows software ecosystem, including kernel-level drivers. AMD's integrated GPU provides competitive performance for light creative work and better real-world GPU-accelerated AI task throughput than the NPU alone.
• Intel Panther Lake: Offers a balanced approach between the ARM-based Snapdragon and x86-based AMD, with full Windows compatibility and integrated graphics. The specific architectural advantages and trade-offs remain less detailed in current coverage, but the platform clears the Copilot+ certification floor.

What Are the Actual Deals Available Right Now?

Best Buy's promotion includes entry-level machines starting at $299.99 and premium options reaching $1,700. The headline deal is HP's OmniBook 3 14-inch with a Qualcomm Snapdragon X chip, 16 gigabytes of RAM, a 512-gigabyte SSD, and a 2K touchscreen, priced at $549.99, a reduction of approximately $400 from its standard $950 retail price. For buyers who want a larger-screen 2-in-1 alternative, the HP OmniBook X Flip 16-inch with an OLED display is listed at $800, down from $1,350.

On the AMD side, the LG gram Book 16-inch equipped with a Ryzen AI 7 445 processor, 16 gigabytes of RAM, and 1 terabyte of storage carries a sale price of $900 at Best Buy, a $650 discount from its standard $1,550 retail tag. The LG gram Book has also hit $849.99 in some listings, marking its lowest price to date. The Lenovo Yoga 7a, a 14-inch 2-in-1 with an OLED panel and the Ryzen AI 5 430 chip, is available at approximately $800. The Acer Swift Go 16 with a Ryzen AI 7 445 and 1 terabyte of storage completes the AMD lineup at $850.

At the premium end of the sale, Samsung's Galaxy Book6 Pro 16-inch at Best Buy, featuring an AMOLED display and Intel Core Ultra 7 silicon, is listed at $1,700, compared to its usual $2,200. A Dell 27-inch all-in-one desktop with a Core Ultra 7 processor and touchscreen rounds out the lineup at $1,450, down from $1,660.

How to Choose the Right AI PC for Your Needs

• Battery Life Priority: If you need a laptop that lasts all day without charging, choose a Snapdragon X-based machine. These ARM-based systems typically deliver 18 to 20-plus hours of real-world battery life, significantly longer than Intel and AMD counterparts, because the RISC architecture uses less power per clock cycle.
• Software Compatibility Concerns: If your workflow depends on specialized software with kernel-level drivers, VPN clients, professional audio and video tools, anti-cheat gaming systems, or engineering software, verify ARM64 compatibility before purchasing a Snapdragon machine. AMD and Intel x86 systems offer full Windows software ecosystem compatibility without these restrictions.
• Creative Work and GPU Tasks: If you do light image editing, video work, or plan to run local AI image generation models, AMD's Ryzen AI 400 series with integrated RDNA 3.5 graphics offers better real-world GPU-accelerated performance than NPU-only solutions. Independent testing found that image generation tasks completed significantly faster on the integrated GPU than on the XDNA 2 NPU.
• Budget Constraints: Entry-level Snapdragon X machines start at $299.99 and offer excellent value for everyday productivity and web browsing. Mid-range AMD and Intel options at $800 to $900 provide more software flexibility and stronger creative performance.

Is Samsung Planning to Enter the AI PC Market?

Samsung's System LSI division is developing a dedicated AI accelerator for PCs called Gaia, according to reporting from the Korea Economic Daily. The company has reportedly supplied early samples to Lenovo and HP for performance testing, with mass production potentially beginning as early as 2027. Samsung will manufacture Gaia using a 4-nanometer process, and its architecture centers on an optimized NPU capable of accelerating generative AI and other on-device workloads without constantly relying on cloud servers.

The company is also exploring integration with processing-in-memory technology, which enables certain calculations to happen directly inside memory, reducing the amount of data shuffled between the processor and RAM. Samsung could potentially pair Gaia with its own memory products, giving it greater control over several important parts of the AI computing pipeline.

A dedicated discrete accelerator could allow PC manufacturers to add stronger local AI capabilities without rebuilding an entire laptop around a new processor architecture. Samsung's history in PC silicon stretches back more than a decade; the company had its Exynos 5 chips power some Chromebooks in the early 2010s. Today, Samsung's Galaxy Book laptops largely rely on processors from Intel and Qualcomm, but Gaia could eventually give Samsung a proprietary piece of that hardware stack once again.

What Should Buyers Know About TOPS Ratings?

A common source of confusion in AI PC marketing is the TOPS (tera operations per second) metric, which vendors use to advertise their chips' AI performance. AMD's NPU is rated at 60 TOPS, Intel Panther Lake claims 172 "total AI TOPS," and Qualcomm's Snapdragon X delivers 45 to 85 TOPS depending on the model. These numbers are not directly comparable because vendors measure TOPS at different numerical precisions without disclosing it.

A chip claiming 60 TOPS at INT4 precision is doing 4-bit integer operations, which introduce quantization errors that affect model accuracy. A chip claiming 30 TOPS at INT8 may produce more accurate inference for the same model. Intel's "172 total AI TOPS" figure aggregates compute across the CPU, GPU, and NPU, a combined platform figure that is not the same measurement as Qualcomm's 45 TOPS NPU-specific rating. AMD similarly bundles CPU, GPU, and XDNA 2 NPU contributions into its aggregate "platform AI" claims.

For the Copilot+ features covered in this article, Recall, live captions, Windows Studio Effects, and Click to Do, this distinction makes no practical difference. Microsoft's 40 TOPS NPU floor is what governs feature access, and all three platforms clear it. The aggregate TOPS competition only becomes relevant for GPU-accelerated workloads like image generation or running local large language models, where memory bandwidth limits AI inference more than raw TOPS at model sizes above 3 billion parameters.

For everyday Copilot+ feature use, an Intel buyer who sees "172 total TOPS" and a Qualcomm buyer who sees "45 TOPS" are both buying machines that will run Recall and Live Captions at identical quality.