Intel's Arrow Lake (Core Ultra 200S) processors launched last year to underwhelming reviews. Between divided opinions on the built-in NPU, the integrated graphics tile, and the same Lion Cove P-cores and Skymont E-cores found in the ultra-thin Lunar Lake mobile chipset, Arrow Lake failed to capture the gaming enthusiast market.

While Intel increased the performance of the chips with subsequent tweaks to the Arrow Lake platform, such as the pre-configured 200S Boost overclock profile and various BIOS updates, Arrow Lake never managed to overtake AMD's Ryzen 9 9950X3D processor as the top choice for gamers this generation. With the Arrow Lake Refresh chips due in early 2026, here are a few improvements I'd like to see Intel bring to the table with the revamped chips.

Better performance

At least compete with Raptor Lake Refresh

Arrow Lake's initial disappointments stemmed from an underwhelming performance. At best, the Intel Core Ultra 200 generation of desktop CPUs performed as well as the previous 14th generation chips. At worst, the Intel Core Ultra 200 variants performed worse than their predecessors. There were several reasons for this, including an internal focus on power efficiency compared to raw power, and a significant architectural change that included a modest NPU alongside the CPU and integrated GPU.

While that's all well and good for people who don't care much about their CPU performance, Intel marketed the Arrow Lake 200S generation chips to enthusiasts, which ultimately backfired. To Team Blue's credit, Intel has admitted the base Arrow Lake CPUs weren't its best work.

This is ultimately something I'd like to see corrected with the new generation, and a modest 5–10% increase in performance could keep Arrow Lake CPUs performing at Raptor Lake levels if not higher. Of course, I'd always love to see a 20% increase in performance generation over generation, but Arrow Lake Refresh is based on the same basic architecture as Arrow Lake, so I'm keeping my expectations reasonable.

Higher clock speeds

This one is basically confirmed already

Thanks to recent leaks, it's already pretty much confirmed that the Intel Core Ultra 300S generation Arrow Lake Refresh CPUs will come with higher clock speeds.

Initial reports suggest slight clock speed increases, enabling only incremental improvements to CPU performance through overclocking. But even a slight clock speed adjustment can go a long way, particularly in demanding tasks. While you won't notice a difference when surfing the web or checking your email, having that additional CPU speed overhead when overclocking, gaming, or doing high-intensity photo or video editing can make a noticeable difference.

This could be how Intel achieves a 5–10% performance increase without sacrificing overall power efficiency (even if bursting to higher clock speeds would likely consume more power in isolation), but we'll have to wait for more information on the CPUs to confirm.

Better efficiency

If Arrow Lake is the "efficiency" chip, the power savings should be significant

Power efficiency in a desktop CPU isn't usually very high on the list of expectations. Desktop CPUs designed with computing enthusiasts and gamers in mind tend to be power hogs, especially when you get into the 16+ core segment. Although some chips have a tendency to draw more power than others, Intel's 13th and 14th-generation chips were more power-hungry than most. The Intel Core Ultra 9 285K doesn't demand the most power, it is far from "efficient" when compared to AMD's Ryzen 9000 generation, which isn't ideal for Intel's push into more efficient CPUs.

With rising energy prices, increasing environmental concerns, and the need for powerful cooling when dealing with an energy-hungry chip, power efficiency is now a more significant consideration for desktop owners than ever. And if Arrow Lake is going to be an "efficient" CPU, then I'd like to see that power draw drop with the refresh.

Of course, this would present a difficult balancing act for Intel between increasing performance for the refresh while simultaneously increasing efficiency. However, if Intel's committed to the power efficiency aspect, I'd like to see that come with a significant reduction in CPU power draw. And I could easily be persuaded to let go of the performance increase and clock speed adjustments in favor of a far more efficient CPU.

Stronger NPU

If we need to have an AI processor, give us the best.

While many gamers have no desire to deal with an NPU on a desktop, Intel's defense has always been akin to future-proofing, comparing the integrated NPU on a chip to integrated graphics, which can now power your display, thereby saving your GPU resources for more demanding tasks.

While the need for an NPU is still somewhat divisive, as a GPU is always going to be a more powerful accelerator, if Intel is going to put an NPU on a desktop CPU, it may as well be a proper, Copilot+ NPU offering at least 40 TOPS of processing power, rather than the smaller NPU 3 taken from Meteor Lake that's currently used in Arrow Lake. While an NPU is mostly helpful for conserving battery life on a laptop, it's designed for AI workloads and can be significantly more efficient at some AI tasks than the CPU or GPU.

Adding a beefier NPU could also be a way for Intel to decrease power draw on the CPU by offloading more Windows 11 AI features to the NPU, rather than relying on the CPU and GPU. Rumors have been back and forth on whether Intel will upgrade to NPU 4 for its Arrow Lake Refresh CPUs, though it's all ultimately just speculation for now. However, it's the kind of speculation I can get behind.

This is just my personal wishlist

Arrow Lake Refresh is already in testing at Intel, with a launch expected at the end of this year or early 2026. Whether Intel can meet this personal wishlist of mine will have to wait until we receive more information on the chips or run the new CPUs through testing once they've launched.

While I'd love to say that I want to see Arrow Lake Refresh take over as the top gaming CPU against AMD's upcoming Zen 6 processors, expecting that kind of performance boost in a mid-cycle refresh chip is asking to be disappointed. Hopefully, Intel's Nova Lake can help bridge that gap later next year.