Intel Ships First U.S.-Made 18A Laptop Chips as ‘AI PC’ Push Hits Stores

A familiar laptop—new silicon inside

LAS VEGAS — The laptops going on sale Jan. 27 with Intel’s latest chip inside look like the familiar high-end PCs that line electronics store shelves every winter: thin aluminum shells, bright OLED screens, a splashy “AI PC” sticker near the keyboard.

What is new is the silicon at their center — and where it is made.

Intel’s Core Ultra Series 3 processors, code-named Panther Lake and formally introduced Jan. 5 at CES 2026, are the company’s first mass-market client chips built on its 2-nanometer-class Intel 18A manufacturing process. Intel says that process is developed and produced in the United States, primarily at its Fab 52 complex in Chandler, Arizona, which has received billions of dollars in federal support under the CHIPS and Science Act.

The launch turns years of policy speeches and factory groundbreakings into a tangible product: a consumer laptop that, for the first time in decades, relies on a leading-edge logic chip fabricated at scale on U.S. soil rather than in Taiwan or South Korea.

“We are entering an exciting new era of computing, made possible by great leaps forward in semiconductor technology,” Intel Chief Executive Officer Lip-Bu Tan said at an investor event previewing the platform. Intel describes 18A as “the most advanced semiconductor node developed and manufactured in the United States.”

The stakes extend well beyond a single generation of CPUs. Panther Lake is a linchpin in three overlapping campaigns: Intel’s attempt to reclaim technical leadership in mobile computing, Washington’s effort to rebuild domestic chipmaking with public money, and the PC industry’s push to turn “AI PC” from a marketing slogan into a real category with on-device artificial intelligence.

New chip, new category

Core Ultra Series 3 is a family of mobile processors aimed at laptops, mini PCs and edge systems. The flagship consumer part, the Core Ultra X9 388H, combines 16 CPU cores — four high-performance “P-cores,” eight efficiency “E-cores” and four low-power E-cores — with integrated Intel Arc Xe3 graphics and a fifth-generation neural processing unit, or NPU.

Intel says the shift to 18A, along with architectural changes, delivers up to 60% higher multithreaded CPU performance and as much as 77% faster gaming compared with last year’s Core Ultra “Lunar Lake” chips at similar power levels. The company also advertises up to 27 hours of video playback on a reference thin-and-light design.

Early independent tests suggest the gains are substantial, though not always as dramatic as Intel’s best-case figures. Tom’s Guide reported that an Asus Zenbook Duo 2026 equipped with the X9 388H scored about 3,000 in single-core and 17,000 in multi-core Geekbench 6 tests, outpacing Intel’s prior generation and matching or beating the latest x86 rival chips from Advanced Micro Devices. Wired called Panther Lake “Intel’s biggest win in years,” noting roughly one-third higher multi-core performance than Apple’s M5 in its benchmarks and battery life that stretched toward 22 hours on a high-end configuration.

On the graphics side, PC Gamer found that the Arc B390 integrated GPU, with up to 12 Xe3 cores, could rival or beat Nvidia’s GeForce RTX 4050 Laptop GPU in several 1080p games when using Intel’s XeSS upscaling, a notable milestone for a chip without a separate graphics card.

Beyond raw performance, the defining feature is the NPU. Intel’s NPU 5 can deliver around 50 trillion operations per second (TOPS) on AI workloads in higher-end Panther Lake models. Across CPU, GPU and NPU, Intel advertises up to 180 TOPS of combined AI throughput.

That matters because Microsoft’s Copilot+ PC program, which underpins new AI features in Windows, requires at least 40 TOPS of NPU performance. Until now, only Qualcomm’s Arm-based Snapdragon X series met that bar. With Panther Lake, Intel systems can join the same tier of devices designed for local AI assistants, real-time transcription and small language models running on the laptop itself rather than in the cloud.

“With Series 3, we are laser focused on improving power efficiency, adding more CPU performance, a bigger GPU in a class of its own, more AI compute,” Jim Johnson, senior vice president and general manager of Intel’s Client Computing Group, said at CES. He said the company expects “over 200” PC designs based on Panther Lake, calling it “the most broadly adopted and globally available AI PC platform we have ever delivered.”

CHIPS Act in silicon

The chips at the heart of those PCs are also the most visible test yet of the Biden administration’s semiconductor strategy.

The U.S. Department of Commerce has awarded Intel up to roughly $7.9 billion to $8.5 billion in direct funding, along with access to as much as $11 billion in federal loans, to help finance new and expanded factories in Arizona, Ohio, Oregon and New Mexico. Intel also expects to claim a 25% investment tax credit on more than $100 billion in planned U.S. manufacturing and packaging investments.

Commerce Secretary Gina Raimondo has set a goal of increasing the U.S. share of the world’s most advanced chip production from effectively zero to about 20% by 2030. She has cited Intel’s 18A technology and facilities in Arizona as central to that target.

For years, the highest-performance logic chips used in smartphones, servers and many PCs have been manufactured primarily by Taiwan Semiconductor Manufacturing Co. in Taiwan and, to a lesser extent, by Samsung in South Korea. That concentration has been viewed by U.S. officials as a strategic vulnerability, especially as tensions between China and Taiwan have risen.

By moving a 2-nanometer-class node such as 18A into high-volume production in Arizona, Intel and the administration are attempting to show that public subsidies can attract or sustain leading-edge manufacturing at home. The same law has also steered billions to TSMC for its Arizona projects, underscoring that Washington wants both domestic and allied capacity onshore.

Competitors raise the bar

Intel is not alone in chasing the AI PC label.

AMD used CES 2026 to introduce its Ryzen AI 400 series laptop chips, built on its Zen 5 architecture with an XDNA 2 NPU rated up to 60 TOPS and RDNA 3.5 integrated graphics. Qualcomm has begun sampling its Snapdragon X2 Elite Extreme processors for Windows on Arm systems, which early benchmarks suggest can exceed Apple’s M4 and some x86 rivals in multi-core tests while offering around 80 TOPS of NPU performance.

Apple’s M5 processors, introduced in 2025 for MacBook Pro and iPad Pro models, increased CPU performance and neural engine throughput over the M4, while maintaining battery life the company markets as reaching up to 24 hours in some laptops.

In that context, Panther Lake’s technical achievements are competitive rather than unchallenged. Reviewers generally find that Intel’s new chips erase much of the performance gap that opened when Apple moved its MacBooks to in-house Arm-based silicon in 2020, particularly in multi-core workloads. Battery life on top-tier Windows laptops is edging closer to, and sometimes matching, Apple’s figures in specific tests. Arm-based Windows machines, however, still hold advantages in some efficiency metrics and cellular connectivity.

New power, higher prices

For consumers, the first Panther Lake systems arrive mainly at the premium end of the market.

An Asus Zenbook Duo configuration with a Core Ultra X9 388H, dual 14-inch 3K OLED displays, 32 gigabytes of memory and a 1-terabyte solid-state drive is listed around $2,300. Samsung’s Galaxy Book6 Pro and Book6 Ultra laptops with Core Ultra Series 3 processors start in a similar price range in South Korea, with some configurations approaching or exceeding the equivalent of $3,000.

Samsung has indicated that prices for some Galaxy Book6 models are more than 20% higher than those of the previous generation, attributing the increase to factors including component costs and the move to the new AI PC platform. Analysts expect more midrange and entry-level systems based on lower-tier Panther Lake parts later in the year, but the initial wave underscores that advanced AI capabilities currently carry a premium.

That raises questions for schools, small businesses and consumers in lower-income markets if AI-integrated PCs become the expected baseline for work and education software.

Beyond the laptop

Intel is also positioning Panther Lake for devices that will never sit on a desk.

The same Series 3 silicon is offered in embedded and industrial variants certified for operation across wide temperature ranges and continuous 24/7 use. Intel and its partners are targeting applications such as warehouse robots, factory automation controllers, hospital equipment, smart cameras and retail analytics systems.

The company claims those edge-oriented parts can outperform popular embedded platforms such as Nvidia’s Jetson Orin AGX in some large language model and video analytics workloads, while using less power and lowering total system cost by integrating CPU, GPU and NPU on a single chip.

That push means the AI capabilities now being touted for laptops — local language models, real-time vision, autonomous decision-making — are also likely to appear in kiosks, surveillance systems and mobile robots over the next several years, with implications for productivity and employment in logistics, security, health care and retail.

A turning point, with open questions

For Intel, Panther Lake and 18A mark the culmination of a turnaround plan that promised “five nodes in four years” after repeated process delays. For the federal government, the chips shipping Jan. 27 offer a high-profile example of CHIPS Act dollars at work. For PC makers, they provide the hardware foundation for a new wave of features billed as making AI a standard part of personal computing.

Whether those bets pay off will depend on factors that go beyond transistor density. Software developers must deliver compelling AI applications that run best — or only — on machines with modern NPUs. Consumers and enterprises will weigh the benefits of fast, local AI against higher purchase prices. Policymakers will watch to see if factories in Arizona and Ohio reach the scale and sophistication that Washington envisioned when it committed public money.

In the meantime, the first judgment will come from buyers flipping open new laptops this week. Their experience, more than any press release or ribbon-cutting, will determine if the AI PC with an American-made heart feels like a new era of computing or simply the next step in a long, incremental march.