Review: Intel Lunar Lake CPUs combine good battery life and x86 compatibility

Two things can be true about Intel’s new Core Ultra 200-series processors, codenamed Lunar Lake: They can be impressive and embarrassing at the same time.

It’s impressive because it performs reasonably well, despite some setbacks and inconsistencies, and because it gives Intel’s battery life a much-needed boost as the company competes with Qualcomm’s new Snapdragon X Elite processors and AMD’s Ryzen AI chips. It will also be the first Intel chip to meet Microsoft’s performance requirements for Windows 11’s Copilot+ features.

Embarrassing because to reach this goal, Intel had to use another company’s manufacturing facilities to produce a competitive chip.

Intel claims this is a temporary arrangement, just a bump in the road while the company prepares to scale up its upcoming 18A manufacturing process so it can bring its own chips back in-house. That may be true! But years of manufacturing failures (and early reports of 18A problems) have made me reflexively skeptical of any timelines the company gives for its own manufacturing. Intel has been outsourcing some of its own manufacturing at the same time that it is desperately trying to convince other chip designers to make their products in Intel’s factories.

This is a review of Intel’s latest mobile silicon chip, the Asus Zenbook UX5406S, provided by Intel, and not a chronicle of Intel’s manufacturing decline and ongoing financial woes. I’ll focus primarily on telling you whether the chip performs well and whether you should buy it. But this is a rare situation where a chip is not a sure thing for Intel, which could affect our overall analysis.

About Moon Lake

Let’s talk about the formation of Moon Lake, briefly.

Like the Core Ultra 100 chips launched last year that are based on Meteor Lake, Lunar Lake is a collection of tiny chips that are stitched together using Intel’s Foveros technology. In Meteor Lake, Intel used this technology to integrate multiple silicon chips made by different companies — Intel made the compute tiles that house the main CPU cores, while TSMC made the tiles for graphics, I/O, and other functions.

On Lunar Lake, Intel is still using Foveros — using a “core tile” of silicon as a spacer that allows communication between different chips — to put the chips together. But the CPU, GPU, and NPU are combined into a single compute tile, and input, output, and other functions are handled by the platform controller tile (sometimes called the platform controller core, or PCH, in previous Intel CPUs). There’s also a “filler tile,” which is just there so the final product is rectangular. Both the compute tile and the platform controller tile are manufactured by TSMC this time around.

Intel still splits its CPU cores between power-efficient E cores and highly efficient P cores, but the core count overall is down slightly compared to previous-generation Core Ultra chips and older 12th- and 13th-gen Core chips.

Lunar Lake features four E cores and four P cores, a configuration common in Apple’s M-series chips but not yet in Intel chips. For example, the Meteor Lake Core Ultra 7 155H features six P cores and a total of 10 E cores. The Core i7-1255U features two P cores and eight E cores. Intel has also removed Hyperthreading from the CPU architecture it uses for the P cores, claiming that the silicon space is better spent improving single-core performance. You might expect this to boost single-core performance in Lunar Lake and hurt multi-core performance compared to previous generations, and to spoil our performance section a bit, that’s basically what’s happening, but not to the extent you might expect.

Intel is also shipping a new GPU architecture with Lunar Lake, codenamed Battlemage — which will also power the next wave of Arc desktop GPUs, when we get them (Intel hasn’t said anything about this, but it’s been canceling or passing on a number of its side projects lately). It said the Arc 140V integrated GPU is 31 percent faster on average than the older Meteor Lake Arc GPU in games, and 16 percent faster than AMD’s latest Radeon 980M, though performance will vary widely based on the game. The Arc 130V GPU has fewer Xe cores than Intel (7, instead of 8) and lower clock speeds.

The final piece of the computing puzzle is the Neural Processing Unit (NPU), which can process some AI and machine learning workloads locally rather than sending them to the cloud. Windows and most apps still don’t do much that way, but Intel estimates Lunar Lake’s NPUs to be between 40 and 48 trillion operations per second (TOPS) depending on which chip you buy, meeting or exceeding Microsoft’s 40 TOPS requirement and generally about four times faster than the NPU in Meteor Lake (11.5 TOPS).

And there’s one last big change: For these Core Ultra chips, Intel is integrating the RAM into the CPU package, rather than letting PC makers solder it to the motherboard separately or provide DIMM slots—again, something we see in Apple Silicon chips in the Mac. The Lunar Lake chips ship with either 16GB or 32GB of RAM, and most variants can be had with either amount (in the chips Intel has announced so far, model numbers ending in 8 have 32GB, and model numbers ending in 6 have 16GB). Packing the memory this way saves motherboard space and, according to Intel, reduces power usage, because it shortens the physical distance the data has to travel.

I’m reasonably confident we’ll see more Lunar Lake variants with more CPU cores and external memory — I don’t see Intel abandoning high-performance, high-margin laptop processors, and these chips will need to compete with AMD’s high-end performance and offer additional RAM. But if these chips are coming, Intel hasn’t announced them yet.