It's been more than a decade since Intel launched the first NUC (Next Unit of Computing) mini PC, and over the years I've come to think of NUC family as reference designs for other small form factor computers. Sure, there have been a few lemons over the years, but for the most part each new model featured the latest hardware, and typically set a benchmark for performance, port offerings, cooling solutions, and noise emissions.

So when Intel announced last year that it was shutting down its NUC business, I was disappointed. And when ASUS announced it had signed a non-exclusive deal to carry on design, manufacturing, marketing, and support for the NUC family, I was somewhat apprehensive. I wondered what may happen to the NUC concept. So I was pleased to have an opportunity to test one of the first ASUS-branded NUC systems.

This year ASUS introduced the first NUC branded computers that didn't have Intel's name on them. The ASUS ROG NUC is a compact PC designed for gaming, with a 2.5 liter chassis and NVIDIA discrete graphics, while the NUC Pro and Pro+ are general-purpose computers with smaller bodies thanks to integrated graphics and 4'' x 4'' motherboards.

Taiwanese mini PC maker GEEKOM provided me with an ASUS NUC 14 Pro for testing. GEEKOM is best known for making its own computers, but the company is also now a Value Added Reseller for Asus NUC systems that it sells to business customers, which come configured with memory, storage, and a Windows license.

While GEEKOM doesn't sell the NUC 14 Pro directly to consumers, you can pick one up from multiple retailers including Amazon, the Asus Store, B&H, and Newegg with prices starting at around $500 for a model with an Intel Meteor Lake processor.

The computer featured in this review is based on the NUC14RVH-B kit from ASUS. It has an Intel Core Ultra 7 155H processor, and GEEKOM has equipped the system with 32GB of RAM and 1TB of storage. The only difference between a NUC14RVH-B and the NUC14RVH model is that the latter includes the ASUS logo on the case.

GEEKOM provided Liliputing with this NUC 14 Pro demo unit for free, with no requirement that the computer be returned upon completion of the review. This review is not sponsored by GEEKOM, and the company did not modify or approve the content of this article in any way.

The ASUS NUC 14 Pro (NUC14RVH) consists of a 117 x 117 x 54 mm (4.61 x 4.61 x 2.13 inches) square plastic case which covers an internal metal frame. It also includes a detachable base to provide access to internal components. It is an actively cooled mini PC with an internal fan.

The review model (NUC14RVHU7) came with an Intel Core Ultra 7 155H mobile processor consisting of 6 Redwood Cove "performance" P-cores, 8 Crestmont "efficiency" E-cores and 2 Crestmont "low-power efficiency" E-cores. The performance cores maximum turbo frequency is 4.8 GHz, the efficiency cores maximum turbo frequency is 3.8 GHz and the low power efficiency cores have a maximum turbo frequency of 2.5 GHz. Together they provide a total of 22 threads.

This processor also includes an integrated GPU from the Intel Arc graphics family and has 8 Xe-LPG graphics cores based on the Alchemist architecture with a maximum dynamic frequency of 2.25 GHz.

The Neural Processing Unit (NPU) provides integrated AI capabilities and can perform up to 11 TOPS making a total compute performance of 34 TOPS after including 5 TOPS from the CPU and 18 TOPS from the iGPU.

Looking at the device front-on, from left to right there are:

On the left side there is a Kensington Lock together with plenty of holes for ventilation.

There aren't any ports on the right side, as it just has ventilation holes.

The bulk of the ports are on the rear. These include, from left to right:

Immediately obvious is the lack of a 3.5 mm headphone jack. Evidently the "Pro" in the name targets business users who have been deemed unlikely to need one. I did consider the reason may be that as the mini PC is likely to be mounted on the rear of a monitor, a headphone port would be quite inconvenient to access, making Bluetooth the preferred option for audio connectivity. But then I thought the same applies to the USB and Thunderbolt ports.

This isn't something that started with ASUS though - Intel omitted headphone jacks on recent NUC Pro systems made for business customers, and it looks like ASUS has decided to continue with that approach.

All of the ports are annotated with labels describing their function and, in the case of the USB ports, their theoretical top speeds. Testing showed all of the ports both functioned correctly and are as per the specification, although I'll talk more about the Thunderbolt performance later.

The base has four rubber feet and two screw-holes for VESA mounting.

To access the memory, storage, and other internal components you just need to remove the base.

When ASUS introduced the NUC 14 Pro earlier this year, the company said it featured a tooless design, allowing you to remove the bottom panel without any tools. But that wasn't exactly true in my experience. There may not be any screws to remove, but a screwdriver or similar tool still comes in handy.

After sliding the catch left on the base, you then prise the bottom of the mini PC towards you. The first hurdle you could encounter is that the catch mechanism may be locked. There's a rotating dial that needs to be set vertically, pointing to the "red" dot, in order to release the catch. And the only way I could get this to turn was using a flathead or slotted screwdriver.

The screwdriver will also come in handy again as you can thread it through the hook below the dial to give you some purchase when prising the bottom off. Just don't blindly yank the base off as there is a cable attached to it which connects to the motherboard inside.

The underside of the base contains a mount for a 2.5" SSD or HDD SATA drive. It attaches to the motherboard through a ZIF cable.

On the right of the motherboard is a M.2 2280 PCIe Gen 4 x4 slot which for the review unit was populated with a 1 TB Kingston drive (OM8PGP41024N-A0) and appears to be from Kingston's Industrial SSD range.

In the centre, there is space for an additional M.2 PCIe Gen 4 x4 drive but the drive is restricted to 2242 size.

On the left side there are dual SO-DIMM slots with support up to DDR5-5600 memory. In the review unit, these were populated with two Crucial 16 GB DDR5 5600 MHz (CT16G56C46S5) sticks of memory.

Underneath the main M.2 2280 drive is a soldered-on Intel WiFi 6E (AX211AX211D2W) which also provides support for Bluetooth 5.3 (LMP 12.15715). This frankly is disappointing. I would have expected to see WiFi 7 as a minimum on an NUC system released in 2024. Failing that, it should also be a physical network card allowing the user to upgrade the wireless hardware.

At the top and bottom of the above picture you can also see further evidence of the toolless design. Neither of the M.2 drives are secured using screws. Instead the drives are secured in place using plastic plugs or stoppers which are conveniently attached to the mount so you don't lose them.

I get that if you don't lock the base, then once inside you don't need a screwdriver. However, you still have to fight to get the base off the very first time you open the computer, although it does loosen up over time. But you still have that inconvenient and fiddly ZIF cable which arguably is more of a pain than simply using a small screwdriver.

The NUC 14 Pro review unit came with Windows 11 Pro Version 23H2 OS build 22631.3958 pre-installed, which I upgraded to OS build 22631.4541 for testing.

I also shrank the Windows partition down by 100 GB so I could install Ubuntu 24.10 (Oracular Oriole).

One key group of settings that can have a significant impact on the performance, are those controlling Power Limits. I was slightly surprised to see that by default, the setting for PL2, which controls the turbo boost power limit, was set the same as PL1 and both were set to 64 watts.

Even though Tau was set to 28 seconds, there would not be any additional power to boost the processor to temporarily run at a higher frequency. Having said that, the recommendation from Intel is to set PL1 equal to the Processor Base Power, which in this case would be 28 watts.

So ASUS must be very confident in the NUC 14 Pro's cooling solution by allowing the mini PC to effectively run at PL2 indefinitely. It also means that cooling will have to rely on frequency throttling should the temperature reach the maximum operating temperature.

All the testing I performed was using Windows' power mode set to "Best performance" through Settings.

I started performance testing by using Crystal Dew World's CrystalDiskMark to measure the included NVMe drive's performance. The results were not particularly outstanding for a PCIe Gen 4 drive running with four allocated lanes. The sequential read speed was around 4100 MB/s and the write speed around 3200 MB/s.

I then tested the USB ports. It is good to see a USB Type-C 3.2 Gen 2×2 port on the device as this provides one of the USB specification's optional speeds of 20 Gbit/s. The actual port speed was observed running with sequential read and write speeds of just over 2000 MB/s.

The rest of the more commonly found USB Type-A 3.2 Gen 2×1 mini PC ports ran at 10 Gbit/s and when tested the sequential read and write speeds was just over 1000 MB/s.

But the ports I was most interested in testing were the two Thunderbolt 4 ports on the rear of the NUC 14 Pro. These are both genuine, fully Intel licensed Thunderbolt 4 ports, which are marketed as being capable of providing throughput of up to 40 Gbits/s.

When I plugged in my external NVMe drive enclosure, which uses the ASM2464PD chip, it connected to a "USB Type-C Subsystem PCIe Root Port" with the hardware Vendor ID of 8086 (Intel) and Device ID of 7EC4 (Meteor Lake-P Thunderbolt 4 PCI Express Root Port). The sequential reads were just over 3800 MB/s and the sequential write speeds were lower at just over 3100 MB/s. This is interesting, and I'll cover why later.

To see how the NUC 14 Pro compares with other mini PCs, I've run a number of performance benchmarks that I've also run on several other mini PCs I've tested recently. These include models with Intel processors, such as the Beelink GTi12 Ultra (Intel Core i9-12900H), the Simply NUC Onyx NUC13OXv9 (Intel Core i9-13900H) and the GEEKOM GT1 Mega (Intel Core Ultra 9 185H) as well as an AMD mini PC, the MINISFORUM UM890 Pro (AMD Ryzen 9 8945HS).

On Windows I ran:

Obviously the Intel Core Ultra 9 185H used by GEEKOM GT1 Mega mini PC has a slightly higher specification than the Intel Core Ultra 7 155H found in the NUC 14 Pro.

Firstly it has a higher maximum Turbo Frequency of 5.1 GHz compared to 4.8 GHz for the performance cores. Also, for each of the performance, efficiency and low-power efficiency cores, the base frequency is substantially higher in the Core Ultra 9 with the frequency increased by 900 MHz for both the performance and efficiency cores, and by 300 MHz for the low-power efficiency cores. The GT1 Mega also implements power control differently by making use of Intel's Turbo Boost Technology 2.0 by configuring PL1, Tau and PL2.

This results in making the GT1 Mega around 10% better than the NUC 14 Pro across CPU benchmarks and 5% across GPU benchmarks..

I also ran 3DMark Time Spy, which achieved a CPU score of 11092 and a GPU score of 3379 for an overall score of 3772.

The Intel ARC graphics performance was slightly lower than for the AMD and the GT1 Mega mini PCs. Realistically you are going to need to add an eGPU if you want to play the more recent "AAA" games, so I decided to connect one.

I first ran AIDA 64 to measure the throughput when writing to the eGPU's memory. And I was not surprised when I saw a write speed of only 2270 MB/s. This was very similar to the write speed I saw with the GEEKOM GT1 Mega and GPD G1 where it was 2277 MB/s.

What is interesting is that the NUC 14 Pro still managed to get an average FPS of 93 whilst running the in-built benchmark from Shadow of the Tomb Raider at 1080p resolution using the high quality graphics preset. This is exactly the same as I got on the GT1 Mega/GPD G1 despite the GT1 Mega performing better in the CPU benchmarks.

Based on these results we now have the Thunderbolt 4 ports from the NUC 14 Pro performing at the same speeds as the USB4 ports on the GT1 Mega.

But when I reviewed the Beelink GTi12 Ultra, which has a Thunderbolt 4 port, I found that Thunderbolt 4 was slower than the USB4 ports on the MINISFORUM UM890 Pro.

So which is it? Is USB4 faster than Thunderbolt 4 or are they both the same? I couldn't find a definitive answer from using search engines and AI chatbots on the internet, so instead I'll offer my experiences from using both types of port on the various mini PCs I've reviewed.

First it is worth noting some certification implications. Thunderbolt 4 devices are certified by Intel, while USB4 devices are not. So if you have an uncertified "Thunderbolt 4" port on a mini PC it ends up being called a USB4 port. To be certified as Thunderbolt 4, the port must be capable of data transfer speeds up to 40 Gbps (5000 MB/s). However the USB4 specification only requires ports to be capable of data transfer speeds up to 20 Gbps (2500 MB/s). As there's no guarantee of the speed of a generic USB4 port, on a mini PC each port needs to be checked against the manufacturer's specification for the port to confirm its data transfer speed.

And secondly, two of the most significant factors that I have seen from my review testing is the performance from a USB4 port on a mini PC depends on the processor and how the port is connected to the PCI BUS. More specifically, what chip is used for the USB4 port connection in the mini PC and what is the certification for the cable used to connect to the port.

This can be demonstrated by looking at the CrystalDiskMark results for some Intel and AMD mini PCs I've reviewed with either USB4 or Thunderbolt 4 ports. The drive being tested was a 500 GB Samsung 980 PRO PCle 4.0 NVMe M.2 (MZ-V8P500BW) drive capable of sequential read speeds of up to 6,900 MB/s and sequential write speeds of up to 5,000 MB/s. It was housed in an external NVMe drive enclosure which uses the ASM2464PD chip, and was connected using a Plugable Thunderbolt 4 Cable (Certified Thunderbolt 4 / USB4).

As seen in this review, the ASUS NUC 14 Pro Thunderbolt 4 ports use a "USB Type-C Subsystem PCIe Root Port" with the hardware Vendor ID of 8086 (Intel) and Device ID of 7EC4 (Meteor Lake-P Thunderbolt 4 PCI Express Root Port). So it has "Intel Thunderbolt 4" ports.

The GEEKOM GT1 Mega I recently reviewed has USB4 ports that also use the same "USB Type-C Subsystem PCIe Root Port" with the hardware Vendor ID of 8086 (Intel) and Device ID of 7EC4 (Meteor Lake-P Thunderbolt 4 PCI Express Root Port).

So whilst it has ports that are described as USB4, they're effectively uncertified Intel Thunderbolt 4 ports.

I also reviewed the Beelink GTi12 Ultra with its rear Thunderbolt 4 port that uses an "Integrated Thunderbolt PCIe" PCI Bus connection with a hardware Vendor ID of 8086 (Intel) and Device ID of 7EC4 (Alder Lake-P Thunderbolt 4 PCI Express Root Port). So it also has an "Intel Thunderbolt 4" port.

This shows that the new models of Intel mini PCs that offer either Thunderbolt 4 or USB4 ports are physically connected by Thunderbolt 4 PCI Express Root Ports and are effectively Thunderbolt 4 ports. It is just that they may not necessarily be certified and therefore cannot use the term Thunderbolt 4 in their description.

And we know from Intel that Thunderbolt 4 data transfer speeds can only reach a maximum of 3,000 MB/s which is why the "Thunderbolt 4" AIDA 64 write speed for the NUC 14 Pro of 2270 MB/s was similar to the GT1 Mega "USB4" write speed 2277 MB/s.

By comparison, all of the AMD-powered mini PCs with USB4 ports that I've tested recently have been connected by a PCIe USB4 bridge.

Using the MINISFORUM UM890 Pro mini PC as an example, the USB4 bridge shows a hardware Vendor ID of 1022 (AMD) and Device ID of 14EF (Family 19h USB4/Thunderbolt PCIe tunnel). So it has "AMD USB4" ports.

We also know from USB-IF that USB4 data transfer speeds can reach approximately 3,800 MB/s which is why the "USB4" AIDA 64 write speed for the MINISFORUM UM890 Pro was 3789 MB/s.

Returning to the earlier question of "Is USB4 faster than Thunderbolt 4 or are they both the same?". My experience shows that answer to be equivocal. This is because there are several factors that can affect the performance, including:

The best AMD mini PCs with USB4 ports I have tested had data speeds of up to 3800 MB/s whereas the best Intel mini PCs with either Thunderbolt 4 or USB4 ports were slower with data speeds of only up to 3000 MB/s.

And just for the record, currently some AMD motherboards have Thunderbolt 4 ports (e.g. ASUS ProArt B550-CREATOR but I've not seen any Thunderbolt 4 ports on AMD mini PCs yet.

Having installed Ubuntu 24.10 as dual boot, I tested the basic functionality and confirmed everything worked as expected.

There were no problems with wireless performance, audio, or video output. I also checked that the Thunderbolt 4 ports operate at 40 Gb/s (i.e. with two lanes of 20 Gb/s).

The 2.5 Gb Ethernet port performs as expected being similar to other mini PCs at around 2.37 Gbps for both upload and download.

WiFi was good for upload but somewhat disappointing for download but that is probably because the network card is only WiFi 6E. Download on the 5 GHz band was 1.20 Gbps and upload was 1.66 Gbps. On the 6 GHz band, the throughput was similar with download reaching 1.16 Gbps and upload reaching 1.72 GHz.

The UEFI (BIOS) interface is the same as on earlier generations of Intel NUC mini PCs. The key difference to other mini PCs is related to the settings of PL1 and PL2 values. Intel configures these based on the "External Ambient Temperature Tolerance". The default setting of 35°C or the optional setting of 40°C both set PL1 and PL2 to 64 W and Tau to 28 seconds on the NUC 14 Pro.

Setting the ambient temperature to "User Defined" allows independent control of PL1 and PL2. First you are offered a "Power" submenu where you can change from a default of "Balance Enabled" to "Low Power Enabled" or "Max Performance Enabled" together with a couple of other settings. Secondly you can manually set the values of PL1, PL2 and Tau. And thirdly under a "Performance" submenu and then a "Processor" sub-submenu, you can toggle Hyper-Threading, Intel Turbo Boost Technology or set the number of active Performance or Efficiency Cores.

There is also a Fan Mode that defaults to "Custom" that offers a "Cooling" submenu or the option of "Fixed Cooling". Under the "Cooling" submenu you can set the controls for the fan curve and you can change the Max, Min and Hysteresis Temperature, Duty Cycles and other settings typically associated with fan curves.

The NUC 14 Pro is audible when under load with the fan assisting in keeping the thermals under control. At idle, the fan noise measured 31 dBA and was effectively silent.

On Windows, when looping the Cinebench R23 benchmark the temperature climbed to an average of around 100°C. The fan consequently ramped up and the noise measured around 47.7 dBA where it remained whilst running the benchmark.

On Ubuntu when idle the CPU's average frequency was around 4300 MHz. Running a "stress" test on all cores for 20 minutes results in thermal throttling being used to control the CPU temperature.

When the test starts, the CPU temperature climbs rapidly to 100°C during which the CPU frequency is throttled down to 4100 MHz. As the CPU temperature still rises to 101°C, the CPU is forced to drop its average frequency down to 3800 MHz resulting in an immediate drop in CPU temperature to 95°C. The CPU frequency then bounces between 3800 MHz and 3900 MHz while the CPU temperature gradually climbs back up to a maximum of 101°C.

The CPU temperature then remains hovering around 100°C which after about 1 1/2 minutes causes further throttling of the CPU frequency down to 3600 MHz before climbing slightly to average around 3700 MHz for the rest of the test.

Power consumption was measured by connecting the NUC 14 Pro to a power meter that was in turn connected to a UPS power outlet to provide clean power.

It is hard to be positive about the NUC 14 Pro. Yes, the performance was okay and it wasn't too noisy to run. But as an NUC (Next Unit of Computing), it's supposed to set a standard by featuring the latest hardware and best-in-class performance. And it fails miserably on those fronts.

The NUC14RVH-B was announced the same week as the WiFi 7 standard became official, but this little computer comes with WiFi 6E instead of the latest wireless networking hardware. Worse still, it's soldered down, which means that the only way to add WiFi 7 support is through a USB adapter.

The computer does have 2.5 Gbps Ethernet, but it only has a single port at a time when many other mini PCs have two. There's no SD card reader or 3.5mm audio jack. The second M.2 expansion slot only supports M.2 2242 cards. And sure, the system has Thunderbolt 4 ports, but they have slower throughput than USB4 ports on AMD devices. .

I should point out that it was rumoured that the NUC 14 Pro was designed by Intel before the company exited the NUC business, which means that one of the first ASUS NUC devices is really an Intel system that ASUS brought to market. But that doesn't excuse its lacklustre offerings.

However, if you weren't looking to be dazzled by the branding, then the pricing may give you the shock you desire. ASUS sells the barebones NUC 14 Pro model with this Core Ultra 7 155H processor for $660 on Amazon. For the same configuration as the review unit (32 GB RAM and 1 TB storage) it is $996. Even though it will be cheaper through GEEKOM, there are better options available for lower prices.

At the end of the day, businesses who are buying NUCs by the pallet load and have previously placed their trust in Intel for a small, reliable mini PC, are not going to be let down by the ASUS NUC 14 Pro.

But for consumers, and especially for those on a tight value-for-money budget, then this may not be the mini PC for you. I've always held NUCs in high regard and recommended them. But this looks like cashing in on the NUC brand and signals bad omens for its future.

I'd like to thank GEEKOM for providing the review unit.