NVIDIA’s laptop GPUs are being throttled — modder blows past limits with 250W RTX 5090 and unlocks 40% more performance

by

When I got the ASUS ROG Strix SCAR 18 (G835L) gaming laptop for a review earlier this year, I quickly found out that the NVIDIA RTX 5090 Laptop GPU within wasn’t delivering its usual performance level.

Or, maybe it was expected, depending on whether or not you work for NVIDIA.

To put it simply, finding the RTX 5090 Laptop GPU underperforming significantly in a spacious 18-inch desktop computer equipped with an expansive cooling system was far from enjoyable.

How well did our tested device perform compared to a Lenovo Legion Pro 7i (Gen 10) with an RTX 5080 Laptop GPU? Remarkably, they delivered almost identical FPS in games like Cyberpunk 2077 and Forza Horizon 5, but the cost was significantly lower by around $1,000 or more.

In simpler terms, you had both of your mobile graphics processors (GPUs) operating at their maximum power consumption of 175 watts. Translated plainly, the RTX 5090 Laptop GPU isn’t a good investment for gamers.

Couldn’t it potentially be configured differently? A recent video by Youtuber GizmoSlipTech reveals that the mobile version of the RTX 5090 holds significantly more untapped power; it’s currently restricted due to power limitations.

What does a mobile RTX 5090 look like at 250W compared to 175W?

As a tech enthusiast, diving into the world of pushing boundaries, I’ve discovered that GizmoSlipTech had to take an unconventional approach to get more juice out of the mobile RTX 5090. They resorted to a technique known as “shunt modding” the card. Here’s the lowdown on what this entails and why it’s best left for those who are seasoned tinkerers with extensive experience in handling PC components: Shunt modding essentially bypasses the card’s 175W power ceiling, which can lead to improved performance but also increased risks of overheating, damage, or even system instability. So, unless you’re well-versed in the art of component tweaking and are ready to shoulder those potential consequences, it might be wise to steer clear of this mod for now.

Despite being designed for lower power consumption, the modified Eluktronics Hydroc 16 G2 laptop with an RTX 5090 graphics card could run steadily at approximately 75 watts beyond its intended limit, reaching a maximum of 250 watts.

As an analyst, I can express this finding as follows: When tested using the 3DMark Time Spy benchmark, the RTX 5090 Laptop GPU achieved a score of 33,056. In contrast, other unmodified RTX 5090 laptops operating at 175W managed to score approximately 24,868. This implies a notable difference in performance, with the modified GPU delivering around 27% more output.

Transitioning to 3DMark Steel Nomad, it’s worth noting that the performance gap between the overclocked mobile RTX 5090 (modded at the shunt) and its stock counterpart with a 175W power limit expands to approximately 41%. This significant disparity showcases the substantial potential output being held back by NVIDIA’s restrictions.

Additionally, GizmoSlipsTech conducts detailed game tests, consistently favoring the powerful RTX 5090 graphics card.

In Black Myph: Wukong, there is a 23.3% power difference when comparing the 175W RTX 5090 to a 250W RTX 5090, considering the adjustments made to the shunt’s clock speeds.

In the game Hogwarts Legacy, the customized RTX 5090 graphics card, running at boosted speeds, performs 14.2% quicker; however, it simultaneously becomes 16.4% warmer.

In Rainbow Six: Siege, there’s a significant boost of 20.4%; on the RTX 5090 that was temporarily removed from service. Shadow of the Tomb Raider jumps by just 8.2%, and The Witcher 3 experiences an impressive surge with a 29.7% improvement, accompanied by a rise in heat by 23.2%.

Eluktronics and XMG respond to requests for NVIDIA to raise power limits

In simpler terms, towards the end of their video, GizmoSlipTech asks NVIDIA to consider increasing the power limits on mobile GPUs. They question why these same chips can handle over 325W in desktop configurations but not run at a higher power draw when used in mobile devices.

1. Instead of missing out on modern advancements in laptop cooling, why not take advantage of them?

2. Rather than neglecting the latest innovations in laptop cooling, how about utilizing them?

3. Instead of overlooking the new developments in laptop cooling technology, wouldn’t it be beneficial to make use of them?

4. Why pass up on the recent breakthroughs in laptop cooling? Let’s explore their potential!

5. Why not leverage the progress made in laptop cooling technology instead of ignoring it?

6. Instead of disregarding the advancements in laptop cooling, why not capitalize on them?

7. Rather than turning a blind eye to the improvements in laptop cooling, how about making use of them for our benefit?

8. Why not seize the opportunity presented by the recent developments in laptop cooling technology?

9. Instead of dismissing the advancements in laptop cooling, why not embrace them and harness their power?

10. Why not make use of the latest advancements in laptop cooling for the betterment of individual companies’ specific supporting hardware configurations regarding NVIDIA GPU power draw?

11. Instead of ignoring the recent improvements in laptop cooling, how about allowing companies to safely configure NVIDIA GPU power draw based on their unique supporting hardware?

As a passionate follower, I must say that the unexpected reactions from Eluktronics, the creator of the tested laptop, and XMG, another laptop brand mentioned in the video, are truly astonishing in this shunt modding exposé I’ve come across.

According to the Eluktronics Youtube account:

Over the past few years, we’ve been working diligently to obtain Nvidia’s approval for manufacturers to set their own GPU wattage limits. Traditionally, mobile systems needed larger and bulkier designs to accommodate GPUs consuming over 200 watts. But today, that’s no longer a requirement. Progress in system engineering, such as thermal interface innovations and our liquid cooling technology, has greatly enhanced power and heat management in portable devices.

In every advancement of desktop graphics cards, Nvidia increases the power limit to achieve better performance. Similarly, mobile graphics cards need similar attention to reach their full potential. Fortunately, the constraints are becoming more noticeable, particularly in the 50xx series.

As a researcher eagerly anticipating approval, I am thrilled to express our readiness to showcase the remarkable capabilities of today’s mobile GPUs. This endeavor signifies a significant stride towards bridging the performance disparity between mobile and desktop graphics solutions. Can we not make it now that laptops deliver performance worthy of an upgrade?

As an analyst, I find this reaction from a significant PC manufacturer quite concerning, and it’s not hard to imagine that other prominent players might have similarly petitioned NVIDIA for greater influence over GPU capabilities.

In simpler terms, unlike the RTX 5090, neither AMD nor Intel currently offer a comparable product, meaning there’s no external pressure for NVIDIA to boost its power levels to match other hardware in the market.

In essence, it’s really aggravating for customers dealing with this predicament. The RTX 5090 is undeniably one of the priciest hardware options in contemporary gaming laptops. However, if reports are true that it may be losing up to 40% of its potential performance due to unnecessary constraints, then this situation calls for a change.

What is shunt modding and why should you NOT do it yourself?

According to the video, unlike its desktop counterpart, the RTX 5090 Laptop GPU doesn’t sport the GB202 chip. Instead, it employs the GB203 chip, which is shared by both the desktop and laptop versions of the RTX 5080.

The RTX 5080 graphics card on your desktop boasts significantly increased power headroom, enabling it to potentially overclock above 400 watts, although it typically operates at a base power of 325 watts.

As a researcher delving into this subject, I’ve found that these numerical data provide insights into the origins of performance discrepancies and shed light on how enhancing power significantly increases performance capabilities.

In simpler terms, shunt modding refers to a physical adjustment made to a device, enabling it to operate at increased power levels or faster speeds beyond its standard operation.

GizmoSlipTech has secretly boosted the power consumption of the mobile RTX 5090 by swapping out a resistor inside the GPU. This trick makes the GPU believe it’s still operating at 175W, but in reality, it runs at 250W without setting off any warning signals or safety mechanisms.

Apart from identifying the correct resistor for replacement, it’s equally important that you are adept at executing a proper replacement process. Similarly, on the software front, your skills should encompass adjusting the power curve to align with the physical modifications made.

Just a heads-up, self-attempting shunt modding isn’t advisable. Please be cautious about it as I wouldn’t want you to encounter any issues!

Read More

2025-07-22 00:10