NVIDIA's GeForce RTX 2060 Founders Edition is designed to hit the "sweetspot" market segment around $350 hard—this is where AMD's Vega cards hang out. The RTX 2060 is built using the same GPU as the RTX 2070, just with fewer shaders (1920 instead of 2304), texture units (120 instead of 144), ROPs (48 instead of 64), memory bus width (192 bit instead of 256 bit), and memory capacity (6 GB instead of 8 GB). The board design, VRM circuitry, and power delivery is identical, at least when comparing the Founders Editions.

With those specs, the RTX 2060 sits roughly between the GTX 1070 Ti and GTX 1080 in performance, which means it beats the AMD Radeon RX Vega 64, which is the fastest card AMD has to offer. Pretty impressive for a card ending in "..60", a designation that used to be reserved for the mid-range. At 1440p, NVIDIA's RTX 2070 is 18% faster; the RTX 2080 is 41% ahead, and the RTX 2080 Ti has a 71% advantage. This means that we can easily recommend the RTX 2060 for all 1440p gaming at maximum details or 1080p with a high-refresh-rate monitor;144 Hz, for example.

As mentioned before, the GeForce RTX 2060 uses the same form factor as the RTX 2070 with the same cooling configuration and sexy looks, but terribly complicated disassembly process. Since the heat output of the RTX 2060 is lower, the cooler has an easier time handling the heat. NVIDIA was wise not to downgrade cooling capacity since that helps achieve a quieter operation in gaming. With gaming noise levels of 32 dBA, the RTX 2060 is quieter than anything that has come out of AMD for several years with the exception of the watercooled Fury X at the same noise level as the air-cooled RTX 2060. Unfortunately, what's still missing is the idle-fan-off feature that completely turns off the fans when the card is lightly loaded or sitting idle, but with 28 dBA, idle noise levels have definitely been improved and are whisper quiet now.

Power is delivered through a single 8-pin power connector, which is in a somewhat odd location near the short back edge of the card. However, the card is short enough for it not to be an issue in any case we know of outside of some tiny ITX cases. A single 8-pin is good for 225 W of power draw, which the card isn't even close to using. We measured power draw of up to 170 W, which ensures pretty much every power supply will be able to handle the RTX 2060.

Just like all other Turing cards, the RTX 2060 is no longer limited by heat, but by power instead. NVIDIA is using the power limiter capability of the cards to dial in a very specific power draw profile that exactly defines thermal limits and finds the efficiency sweet spot. As a result, the card will nearly always sit in its power-limited state when gaming, which isn't an issue unless you start overclocking. Basically, you can no longer reach the maximum boost clock state reliably, which leads to stability testing being limited to whatever frequency your test load is running at, and there is a chance that lighter games might run at higher boost clocks and crash the card.

Our manual overclocking tests worked well, though, and we gained an additional 10% real-life performance, which is quite good, especially considering competing AMD cards have little additional overclocking headroom. Our review sample from NVIDIA uses Samsung GDDR6 memory, which overclocks about 100 MHz higher than the GDDR6 chips from Micron. We have reached out to NVIDIA to confirm whether all retail Founders Edition cards will use Samsung memory and will update this space as soon as we hear back from them.

NVIDIA RTX technology is finally available in Battlefield V, and several other RTX titles are coming soon, too. While the adoption rate could be faster, the results in Battlefield are impressive, especially when you think of the paradigm shift that comes with ray tracing. We tested this card across all DXR Reflections quality settings and resolutions and summarized our findings. The RTX 2060 is capable of playing Battlefield V at both 1080p and 1440p with quality set to "high." You get comfortable frame rates in excess of 60 FPS at 1080p and 42 FPS at 1440p, at the "high" setting. Since there is little visual difference between the RTX quality levels, especially during busy gameplay, we recommend setting RTX to "low" to enjoy the benefits of the technology with a performance hit that is as small as possible.

The second big novelty of Turing is acceleration for artificial intelligence. At first thought, many assumed that it won't do anything for gamers, but NVIDIA devised a clever new anti-aliasing algorithm called DLSS (Deep Learning Super-Sampling) which utilizes Turing's artificial intelligence engine. DLSS is designed to achieve quality similar to temporal anti-aliasing and solve some of its shortcomings with a much smaller performance hit. We tested the technology in Final Fantasy XV, which is the first game to support DLSS. Find our detailed review here.

The US$349 price for the RTX 2060 may look daunting if you consider that predecessor GTX 1060 6 GB launched at $249 ($299 for Founders Edition), but you must take into account the massive performance increase over the GTX 1060, and we're not even counting the additional capabilities that tensor cores and RT cores bring to the table. By all intents and purposes, the RTX 2060 belongs to a higher market segment than the GTX 1060, and this is reflected in the card's performance.

At $350, the RTX 2060 renders a whole spectrum of previous-generation graphics cards obsolete. Given it performs on par with the GTX 1080, it no longer makes sense to pick up a "Pascal" GTX 1070 Ti or even its AMD rivals, the RX Vega 56 and RX Vega 64. It now makes sense to pick the RTX 2060 over any similarly priced Pascal or Vega graphics card for the simple reason that you get GTX 1080/Vega 64-like performance with the added advantage of RTX and DXR readiness. NVIDIA is serious about getting as many game developers to implement RTX as possible. As if that were not all, DLSS is a very tangible feature-set addition that offers better visuals and performance than temporal anti-aliasing.