Zotac GeForce RTX 5070 Solid Review 34

Zotac GeForce RTX 5070 Solid Review

(34 Comments) »

Value and Conclusion

  • The Zotac GeForce RTX 5070 Solid is currently listed for $600 in the States, but out of stock, the Solid OC is in-stock for $580. In Europe, the Solid is widely available at a price point of €568 including VAT, which converts to $570 before tax.
  • DLSS 4 Frame Generation and Transformer Upscaling
  • Quiet
  • Only small price increase over MSRP
  • Fantastic overclocking potential
  • Dual-slot design will fit all cases
  • Good temperatures
  • Extremely energy-efficient
  • Idle fan-stop
  • NVIDIA SFF-Ready form factor
  • Support for both HDMI 2.1b & DisplayPort 2.1b
  • PCI-Express 5.0
  • Good video encode/decode hardware acceleration support
  • Gen-over-gen improvement not that impressive (except for multi-frame generation)
  • 12 GB VRAM not enough for 4K + RT + Frame Gen in some titles
  • Non-gaming power levels increased
  • Memory overclocking artificially limited by the driver
Positioning & Architecture
With NVIDIA's GeForce RTX 50 product stack fully complete, we now have time to cover some cards that we weren't able to look at during the original coverage. The Zotac GeForce RTX 5070 Solid is a cost-efficient custom-design model that is designed to offer a slightly premium experience, without breaking the bank. The RTX 5070 is based on NVIDIA's GB205 graphics processor and features 6,144 cores enabled, vs 8,960 on the GB203 RTX 5070 Ti, a 45% difference. Other unit counts have been scaled accordingly, you get 80 ROPs, 192 TMUs and 48 RT cores. The memory subsystem uses GDDR7, too, like the other RTX 50-series cards, but there's only 12 GB VRAM on a 192-bit wide memory bus, clocked at 28 Gbps.

The Blackwell architecture introduces several architectural improvements under the hood, like giving all shaders the ability to run FP32 or INT32 instructions, on Ada only half the cores could do that. The Tensor Cores are now accessible from the shaders through a new Microsoft DirectX API, and they now support FP4 and INT4 instructions which run at lower precision, but much faster with less memory usage.

From a fabrication perspective nothing has changed though—Blackwell is built on the same 5 nanometer "NVIDIA 4N" TSMC node as last generation's Ada. NVIDIA claims this is a "4 nanometer process," but during Ada it was confirmed that NVIDIA 4N is actually not TSMC N4 (note the order of N and 4), but 5 nanometer. At the end of the day the actual number doesn't matter much, what's important is that NVIDIA is using the same process node.

Performance
At 1440p, with pure rasterization, without ray tracing or DLSS, we measured a 20% performance uplift over the RTX 4070, which is a bit lower than expected, but alright I'd say. At 4K, the increase is 25%, which is certainly better than the meager 15% that we got on RTX 5080, the RTX 5090 got +36%, and the RTX 5070 Ti was 27% faster. Just like on the RTX 5080, NVIDIA is unable to achieve their "twice the performance every second generation" rule with the RTX 5070, which is only 64% faster at 4K, and 54% at 1440p than the RTX 3070. Overall performance is roughly similar to the RTX 4070 Ti, 10% behind the RTX 4070 Ti Super. Compared to the RTX 4070 Super the performance uplift is 7%. When compared against AMD's offerings, the RTX 5070 ends up a little bit behind the RX 9070 non-XT, but is much faster than the RX 9060 XT.

The Zotac Solid, which is the company's slightly premium model, above the Twin Edge, ticks at a boost frequency matching the Founders Edition. There's also an OC model, with a small factory overclock, at a higher price. Some competing cards have a slightly higher boost rating, but generally, the differences from factory OC are small. Our manual overclocking worked much better and achieved over 12% performance increases—no idea why all the board partners are so conservative with their overclocks.

While the RTX 5070 is definitely a great card for 1440p, it's a little bit weak for 4K gaming without upscaling, that's why we recommend the card for a resolution of 1440p, especially when you consider that future games will have higher hardware requirements.

Ray Tracing & Neural Rendering
NVIDIA is betting on ray tracing and Blackwell comes with several improvements here. Interestingly, when comparing RTX 4070 with the RTX 5070, the performance gain with RT is smaller than with raster. We saw RTX 5070 trading blows with 4070 Ti in raster, with RT it's clearly behind and is only able to match RTX 4070 Super. Still, RT performance is good, but AMD's new Radeon RX 9070 clearly outperforms the 5070, by roughly the same 10% margin that we saw in non-RT testing, too. This is quite an achievement for AMD—they are no longer behind in ray tracing. With Blackwell, NVIDIA is introducing several new technologies. The most interesting one is Neural Rendering, which is exposed through a Microsoft DirectX API (Cooperative Vectors). This ensures that the feature is universally available for all GPU vendors to implement, so game developers should be highly motivated to pick it up.

VRAM
The RTX 5070 comes with 12 GB VRAM, just like the RTX 4070, RTX 4070 Super and RTX 4070 Ti—so no gen-over-gen increase here, except for the bandwidth, which is 33% higher thanks to the 28 Gbps GDDR7. The RTX 4070 Ti Super and RTX 5070 Ti come with 16 GB. AMD offers 16 GB on both the RX 9070 non-XT and RX 9070 XT. While our data clearly shows that 12 GB VRAM is sufficient for gaming at 1440p, even with RT enabled, it won't be enough for RT at 4K with Frame Generation in several titles. I'm sure that eventually there will be games, too, that hit 12 GB in 1440p, so the RTX 5070 is definitely not future-proof for 1440p, unless you are willing to dial down the settings slightly in those cases, which isn't an unreasonable ask I'd say. From an engineering perspective, 12 GB does make a lot of sense though. Increasing the memory size further would have required a wider memory bus and a GPU design with support for the extra bus width and more pins in the design, etc. All these changes would make the card more expensive, for relatively small gains. More memory does not automagically turn into additional performance. You would see scaling only in those rare cases that go beyond 12 GB. Still, AMD offers 16 GB on the RX 9070-series and that will be a selling point for them.

DLSS 4 Upscaling & Frame Generation
NVIDIA made a big marketing push to tell everyone how awesome DLSS 4 is, and they are not wrong. First of all, DLSS 4 Multi-Frame-Generation. While DLSS 3 doubled the framerates by generating a single new frame, DLSS 4 can now triple or quadruple the frame count. In our testing this worked very well and delivered the expected FPS rates. When using FG, gaming latency does NOT scale linearly with FPS, but given a base FPS of like 40 or 50, DLSS x4 works great to achieve the smoothness of over 150 FPS, with similar latency than you started out with. Image quality is good, if you know what to look for you can see some halos around the player, but that's nothing you'd notice in actual gameplay.

Just to clarify, even multi-frame-generation will not turn 15 base FPS into 60 "playable" FPS. Sure, the movements will be very smooth, just like 60, but the game would react to your inputs at a rate of 15 FPS, which is noticeably sluggish. That's why the "base" FPS value is so important, as mentioned before, 40+ works very well, but it also depends on the type of game. Some slower titles, not shooters, will run great, even with a base FPS of 30.

Want lower latency? Then activate DLSS 4 Upscaling, which lowers the render resolution and scales up the native frame. In the past there were a lot of debates whether DLSS upscaling image quality is good enough, some people even claimed "better than native"—I strongly disagree with that—I'm one of the people who are allergic to DLSS 3 upscaling, even at "quality." With Blackwell, NVIDIA is introducing a "Transformer" upscaling model for DLSS, which is a major improvement over the previous "CNN" model. I tested Transformer and I'm in love. The image quality is so good, "Quality" looks like native, sometimes better. There is no more flickering or low-res smeared out textures on the horizon. Thin wires are crystal clear, even at sub-4K resolution! You really have to see it for yourself to appreciate it, it's almost like magic. The best thing? DLSS Transformer is available not only on GeForce 50-series, but uses Tensor Cores on all GeForce RTX cards! NVIDIA has constantly been working on DLSS Transformer support, and it's now available in virtually all modern titles, also thanks to their brilliant override mechanism. While AMD's FSR 4 has definitely improved quality a lot, Transformer still looks a little bit better, and FSR 4 support isn't as common as NVIDIA DLSS 4.

Physical Design, Heat & Noise
Zotac's Solid comes with good build quality, includes a metal backplate, and the cooler shroud—while made of plastic—definitely does not feel plasticky. I also like that the card is only two slots thick, which means it will fit in virtually all cases, as long as they can take a card of 30.5 cm length. Temperatures are reaching only 65°C—excellent. With 29 dBA the card is very quiet, much quieter than the NVIDIA Founders Edition. Compared to other more premium models, the Zotac 5070 sits roughly in the middle of the pack, which is a good result, considering that all the quieter cards are (much) more expensive. If you are looking for a RTX 5070 model that's not too expensive, but still low-noise, then the Zotac Solid is definitely an excellent choice.

Our apples-to-apples cooler comparison test, which runs the cards in a normalized way reveals that the cooler on the Zotac Solid is much stronger than on the Founders Edition, by almost 14°C at the same heat load. Compared to other custom designs, the cooler is somewhat weaker, but not significantly so. The card makes up for this with good energy efficiency, which generates less heat, allowing the cooler to do its job with less effort.

PCI-Express 5.0
NVIDIA's GeForce Blackwell graphics cards are the first high-end consumer models to support PCI-Express 5.0. This increases the available PCIe bandwidth to the GPU, yielding a small performance benefit. Of course PCIe Gen 5 is backwards compatible with older versions, so you'll be able to run the RTX 5070 even in an older computer.

Just like we've done over the years, we took a detailed look at PCI-Express scaling in a separate article. Testing includes x8 Gen 5, for instances when an SSD is eating some lanes. The popular x16 4.0 was tested, which is common on many older CPUs and entry-level motherboards. Finally, some additional combinations were run, down to PCIe x16 1.1. The results confirm that unless you are on an ancient machine, PCIe bandwidth won't be a problem at all.

Power Consumption
While we saw quite high power consumption in idle, multi-monitor and media playback on the other Blackwell cards, this isn't a problem on the RTX 5070 in general. However, the Zotac Solid does use a few watts more in those states, which is a bit unexpected. The difference is not huge, but definitely noteworthy. Gaming power draw on the other hand is improved, even vs the Founders Edition, which is a nice achievement (207 W vs 229 W). This is a decent improvement over even RTX 4070 Ti (252 W), but gen-over-gen, the power draw has increased by 10%, which is pretty close to the performance gains. Not surprising, considering that NVIDIA is using the same process on these GPUs. Overall gaming efficiency is really good, near the top of our charts, considerably better than anything AMD offers, except for the RX 9070 non-XT.

Overclocking
Overclocking worked extremely well on the RTX 5070, reaching +13% in real-life performance. This is an exceptional OC result, and much higher than what we usually see on modern GPUs. No idea why NVIDIA didn't clock it higher—they are masters at eking the last bits of performance out of their GPUs, even at stock. Going from a real measured clock frequency of ~2800 MHz to ~3150 MHz is massive—why didn't they clock these cards higher? Maybe to keep the gap to 5070 Ti bigger, for the upsell? Perhaps to improve harvesting, so that the best GPUs can go to higher-margin products? Either way it's a lost opportunity, but excellent potential for people who are willing to do some manual overclocking. Memory overclocking is equally impressive and topped out at NVIDIA's artificial cap of +375 MHz / +6000 MT/s—the chips could certainly take more.

Pricing & Alternatives
NVIDIA's MSRP for the RTX 5070 is set at $550, which is very reasonable for what's offered. Actually that price is $50 lower than the MSRP of the RTX 4070, which launched at $600. While it was pretty impossible to find any of the new cards at promised pricing in the first months of this year, supply has improved a lot. Especially in Europe it's no problem finding GeForce 50-series cards in-stock at MSRP. While the situation is more complicated in the United States, due to the tariffs, I'm happy to report that several RTX 5070 models can be found online at $550. The Zotac RTX 5070 Solid non-OC from this review is listed at $600 out of stock, but the OC model is in-stock for $580. This of course doesn't make sense, the non-OC version should be a bit cheaper than the OC card. In Europe, the pricing aligns as anticipated. The Twin Edge is available at the manufacturer's suggested retail price (MSRP), the Solid non-OC is priced slightly higher, and the Solid OC is at an even higher price point. When converting for tax and USD, the European pricing puts the card at $570, just $20 above the MSRP. At that price difference, it's really a no-brainer, the better cooler its lower noise levels are definitely worth it over the cheapest baseline cards at MSRP. I don't think I'd pay $50 extra for it though. In terms of looks, the NVIDIA Founders Edition is a strong alternative of course, but it's hard to find in stock. Probably the most interesting alternative to consider is the AMD Radeon RX 9070 non-XT. It is priced at the same $550 MSRP yet offers higher performance in both raster and RT, and 16 GB VRAM for additional peace of mind. The only thing that's missing is support for NVIDIA's exclusive DLSS 3 and 4 technologies—especially the lack of a solution comparable to DLSS Transformer upscaling is problematic. Overclocking on the RTX 5070 does work very well though, so after manual OC I'd say both cards offer roughly similar raw performance. Besides that there aren't really a lot of good alternatives. The RTX 5070 Ti is quite a bit more expensive for not that much of a performance improvement. A used RX 6900 XT will sell for close to $450, but it has much lower RT performance and lacks the RDNA 4-exclusive FSR 4 upscaler support, and DLSS as well. RTX 4070 Ti used is still far too expensive to make it on the list, and last generation's RTX 4070/4070 Super haven't really gone down a lot in price, so I'd rather take the RTX 5070, also for its better resale value.

Looking at future options, there isn't much. Maybe Intel's Arc B770, but it's completely unknown when that launches, if ever. AMD RX 9070 GRE is a China-exclusive model that might make it to the West eventually, but it's unlikely that it will revolutionize this segment. We've been hearing that NVIDIA has been working on their RTX 50 Super refresh, which brings bigger VRAM sizes thanks to 3 GB-sized chips. A RTX 5070 Super with 18 GB VRAM would definitely be interesting, but I'm sure it will come with a price increase, too.
Recommended
Discuss(34 Comments)
View as single page
Nov 16th, 2026 03:40 CST change timezone

New Forum Posts

Popular Reviews

TPU on YouTube

Controversial News Posts