Gigabyte’s so-called “Server-Grade Thermal Conductive Gel” and a degrading thermal paste on the Radeon RX 9070 Gaming OC

1 - Overview and an important foreword 2 - Thermal putty: Consistency, microscopy and issues 3 - Thermal Putty: Measurements and material analysis 4 - Thermal paste: problem case and microscopy 5 - Thermal paste: measurements and material analysis 6 - Summary and conlusion

The mysterious “gel” on Gigabyte’s new graphics cards – analysis instead of alarmism

Today’s article is dedicated to a topic that has attracted a lot of attention in recent days: the so-called “Thermal Conductive Gel” that Gigabyte uses on its current graphics card models and the accusation that this material is leaking or is generally problematic. It is understandable that images of leaking thermal conductive gels spread quickly and cause justified irritation. However, it is all the more important to take a differentiated view of this topic and not to fall for bold clickbait mechanisms straight away. This much in advance: it is very difficult to make a general assessment, as many negative factors have to interact, but it is also clear from my research that something like this can (but does not necessarily have to) happen. But the “can” alone is a complete absurdity with such an expensive product.

I have therefore made a conscious decision to base this article on a sound, scientifically comprehensible foundation. Because the actual topic is far more complex than simply observing a material that shows up on a card edge or has slipped. It is about the question of why a manufacturer like Gigabyte chooses a thermal solution that appears unnecessarily complicated and expensive at first glance, but at the same time has a deep impact on the production processes and only works if numerous factors, from the choice of material to the application method to the control of the application quantity, are optimally coordinated.

To assess this, it is not enough to document visible symptoms. You have to understand what these materials actually are, how they are structured, how they can be processed and what pitfalls arise from the seemingly simple application by dispenser. This is exactly what I will do in this article – objectively, comprehensibly and on the basis of my own measurements, material analyses and microscopic examinations. Because if you want to understand the causes, you can’t stop at the surface.

Marketing vs. reality

In an increasingly saturated market, in which technical differences between graphics card models are often only marginal, marketing seems to be looking ever more desperately for ways to stand out from the competition. All too often, technical substance gives way to the facade of euphonious slogans. Those who used to use “Ultra Durable” or “Military Class” to set themselves apart from the competition are now resorting to terms such as “Server-Grade Thermal Conductive Gel”, always in the hope that it sounds like precision, high availability and professional standards, even if there is often little more behind it than a standard material that can be sprayed.

The fact that the mere mention of high-quality PTM pads on the GPU die is no longer a unique selling point is a development that comes as no surprise. I myself have been preaching the technical advantages of these materials for years, and it is gratifying that many manufacturers have now adopted this approach. But it is precisely because of this that the use of such solutions in the GPU sector is no longer an exception, but rather a sensible industry standard. For marketing purposes, this step therefore loses its appeal because it can no longer be sold spectacularly as a special feature.

Instead, the focus is now shifting to thermal putty, another topic that I have been working on for years and which I have helped to shape in public through detailed tests and practical analyses. After all, I don’t make my investments for nothing. It was foreseeable that this trend would now be taken up by marketing. However, there is a deep gulf between controlled application by private users or in test laboratories and large-scale use in industrial mass production. What works excellently on a small scale comes up against physical, manufacturing and logistical limits in series production when the surfaces are so large.

In the case of Gigabyte, the question therefore remains as to whether the company was actually aware of all the consequences when it decided to use a sprayable thermal putty (because the “gel” is nothing else). The challenges begin with the selection of a material-saving but structurally stable formulated product, and range from coordination with the dispenser and precise dosing to ensuring process stability over tens of thousands of units. And even if all this succeeds, there is still a residual risk of material behavior that only becomes apparent after months in the field, a risk that could be avoided from the outset with appropriately selected pads.

In retrospect, it almost seems as if we have jumped on the next best buzzword that has not yet been completely used up without questioning its implications. After all, the real unique selling point in the hardware sector is ultimately not the choice of terms on the packaging, but the technical consistency in the details, and this is something that is not revealed in the brochure, but under the microscope. I have respect for the step that Gigabyte has dared to take, but I would also like to express my doubts that this will work in the long term.

Pros and cons

The switch to thermal-putty-based solutions in series production is undoubtedly a technological step with a good approach, which has clear advantages. The aim is obviously to reduce the complexity of the cooler design by no longer having to compensate for the technical differences in height between the GPU, VRAM, voltage converters and other components with a large number of thermal pads of different thicknesses. Instead, the putty, with its significantly better malleability and flowability, takes on the function of a compensating material. This simplifies the heatsink design, reduces the need for steps, undercuts or precisely milled recesses in the heatsink and thus saves mechanical production costs.

Another positive effect is the easing of tight tolerance limits. Where pad thicknesses previously had to be adjusted with tenths of a millimeter precision in order to achieve optimum thermal contact pressure everywhere, a well-designed thermal putty allows significantly more leeway. The structure is more robust against production fluctuations and there is no need for fragmented assembly plans with different pad sizes. This simplifies the process chain considerably, especially in large-scale production.

However, this method also brings with it serious challenges that cannot be ignored. Switching to dispenser-based application systems is not trivial, as it requires considerable investment in machines, precisely controllable application hardware, sensor technology and maintenance capacities. Simply changing the material is not enough: good putties must be precisely matched in terms of viscosity, stability and shear behavior so that they can be processed without errors. Each material requires its own pressure parameters, nozzle profiles, delivery times and adjustment for thermal and mechanical stability under production conditions. And the initial post on Quasar Zone and the pictures on Techpowerup show that mistakes can happen.

In addition, the use of Putty significantly shifts the effort required for RMA cases and repairs. What was a simple removal and reapplication with pads quickly becomes a challenge with viscous materials. Separated assemblies must be completely cleaned, residues removed and then rebuilt with precisely dosed replacement material. The time required for reworking increases considerably, as does the susceptibility to errors during servicing.

Last but not least, the long-term stability of putty compared to high-quality pads is also a potential risk. Migration, shrinkage, bleeding or uncontrolled flow under load can occur if the choice of material, dosage and dispenser parameters are not precisely coordinated. In this case, the simplification of the cooler design simply does not pay off in case of doubt with subsequent thermal problems. Problems that would have been much easier to calculate with a conventional design with mechanically defined pad thicknesses. The installation direction also plays an important role, as the industry rarely tests orthogonal (suspended) operation (e.g. Thermaltakes Tower 300) in advance.

The idea of using thermal putty to create more tolerance leeway and simpler cooling solutions is understandable and technologically sound. However, the path to achieving this is rocky, cost-intensive and full of potential pitfalls in its operational implementation, which can place a considerable burden on both production and after-sales. It is a good idea, but it is only viable if it is implemented with the same consistency with which it is advertised as progress in promotional materials.

Today’s review of a Gigabyte Radeon RX 9070 Gaming OC

Today’s review is about a Gigabyte Radeon RX 9070XT Gaming OC, which was kindly provided to me by a reader. This AMD card has a different memory positioning than the NVIDIA cards, so that there are no areas with putty above the PCIe slot. I have to mention this for the sake of completeness, but it also provides a good basis for comparison. At this point, of course, I would like to express my thanks to this supporter, because without such a sample, an in-depth technical analysis of this card would simply not have been possible at present. As some readers will know, I’m obviously no longer one of Gigabyte’s preferred media representatives, as they obviously don’t like to see critical reporting.

As a result, I am no longer provided with official samples and requests for information are also completely ignored. This makes the contribution of the community all the more important, as this is the only way to continue to provide an independent, fact-based and reliable assessment of the manufacturer’s products, such as the RX 9070XT Gaming OC. Despite this well-known history, I will keep the review factual and objective as usual, as I am not in the least interested in the personal sensitivities of individual employees. After all, the technical quality of a product is not measured by the whims of marketing, but by the facts.

In fact, I would like to make a constructive contribution with this article. Because if I can use analytical work to help critically scrutinize certain decisions in product design and make the causes of specific problems comprehensible, then this will ultimately also help Gigabyte itself. Maybe not immediately, but in the medium term on the way to better products. And that, at least from a technical point of view, is actually the goal of everyone involved.

And I am referring explicitly to part 2 of today’s article, which is about the thermal paste of horror.