Nvidia GeForce RTX 40 Series: Everything We Know So Far

Following the success of Nvidia’s current-gen RTX 30 Series GPUs which offered significant gains in raw performance over their predecessors, several rumors surrounding the launch date and specifications for the upcoming RTX 40 Series lineup have been making rounds on the Internet every now and then.

According to these rumors, the next-gen RTX 40 Series GPUs will provide a similar jump in performance compared to the RTX 30 Series lineup and is still in active development behind closed doors at Nvidia HQ.

Furthermore, recent leaks suggest that Nvidia’s Ada Lovelace GPU architecture will be based on TSMC’s shiny new 4nm process node (N4) as opposed to Samsung’s 8nm wafers which power the current-gen Ampere lineup.

For PC gamers who prefer staying on the bleeding edge of technological innovations related to GPU architecture, we’ve gathered a comprehensive list of all the relevant details for Nvidia’s upcoming RTX 40 Series lineup of GPUs.

From their expected release date and pricing structure to their leaked specifications and performance metrics, here’s everything you need to know about the Nvidia GeForce RTX 40 Series GPUs.

Nvidia GeForce RTX 40 Series: Rumored Specifications

Based on the latest information provided by anonymous tipsters like @greymon55 and @kopite7kimi, Nvidia’s initial reveal might feature three new GPUs from their RTX 40 Series lineup:

• Nvidia GeForce RTX 4090 (Codename: AD102-300)

• Nvidia GeForce RTX 4080 (Codename: AD103-300)

• Nvidia GeForce RTX 4070 (Codename: AD104-275/400)

RTX 4090 Specifications:

According to hardware leaker @kopite7kimi, Nvidia‘s next-gen flagship, the RTX 4090 is supposed to feature 28 Streaming Multiprocessors, providing 16384 CUDA Cores in its entirety.

Furthermore, the AD102 GPU comes equipped with 24GB of GDDR6X memory, clocked at around 21Gbps on a wider 384-bit Memory Bus.

Judging by the specifications alone, the RTX 4090 is allegedly twice as fast as Nvidia’s current-gen ‘BFGPU’, the RTX 3090 which boasts 10496 CUDA Cores to its name.

Regarding GPU Clock Speeds, @kopite7kimi’s estimations suggest that the RTX 4090 apparently has a Base Clock of 2235Mhz and a Boost Clock of 2520Mhz with a 2750Mhz actual clock in-game, an impressive feat when compared to the existing RTX 3090 SKU wherein the Base and Boost Clocks max out at 1395 and 1695Mhz respectively.

Despite being built on TSMC’s N4 architecture, the RTX 4090 is expected to reach a TBP of 450W, indicating a major uptick in power consumption over the RTX 3090’s 350W while maintaining a similar power draw with Nvidia’s top-of-the-line RTX 3090Ti.

Alongside these leaks, one of @kopite7kimi’s recent tweets hinted at the performance improvements for the RTX 4090 over its previous-gen counterparts.

In 3D Mark’s Time Spy Extreme GPU Benchmark, the RTX 4090 managed an astonishing score of over 19000 points, suggesting that the new AD102 GPU from Nvidia is around 66% faster than the RTX 3090Ti and 82% faster than the RTX 3090.

With that said, a single benchmark score like the one from Time Spy Extreme isn’t indicative of the actual performance of this GPU and will likely vary depending on a bunch of synthetic workloads.

Nevertheless, it is impressive to see such notable improvements from a generational leap, considering that the RTX 3090Ti was already a top performer across every synthetic benchmark we’ve had so far.

RTX 4080 Specifications:

At the time of writing this article, several speculations suggest that the RTX 4080 will feature 76 Streaming Multiprocessors or 9728 CUDA Cores with 16GB of GDDR6X memory, clocked at around 21Gbps on a 256-bit Memory Bus.

In terms of Power Consumption, earlier leaks pointed to the fact that the RTX 4080 is going to require a TBP of 420W, a staggering 100W higher power draw than its previous-gen counterpart, the RTX 3080.

However, recent claims from @kopite7kimi emphasized upon a significant decrease in power draw for the RTX 4080 – from 420W down to 320W, thereby keeping it in line with RTX 3080’s TBP.

Regarding performance improvements for the RTX 4080 over its predecessors, @kopite7kimi’s latest revelations were solely based on the specifications of the AD103 GPU.

In 3D Mark’s Time Spy Extreme GPU Benchmark, the RTX 4080 manages to score over 15000 points, a whopping 40-50% increase over the RTX 3080 which typically sits at around 10000 points.

From the benchmark score alone, it is evident that the AD103 GPU is significantly faster than the RTX 3080 and aligns itself with the likes of Nvidia’s current-gen RTX 3090 which is often available at a premium price tag.

RTX 4070 Specifications:

After a handful of unconvincing claims about the RTX 4070 being equipped with 10GB of GDDR6 memory on a 160-bit Memory Bus, it seems like Nvidia might have made up their mind on upgrading the SKU with 12GB of GDDR6X memory on a 192-bit Memory Interface.

Based on @kopite7kimi’s recent speculations, the AD104 GPU features 60 Streaming Multiprocessors or 7680 CUDA Cores with Clock Speeds hovering around 2310Mhz for the Base Clock and 2610Mhz for the Boost Clock.

Furthermore, the RTX 4070 is rumored to have a TBP of 285W, a modest increase in power consumption of 45W over its previous-gen counterpart, the RTX 3070.

In 3D Mark’s Time Spy Extreme GPU Benchmark, @kopite7kimi affirms that the RTX 4070 is likely to receive a total score of 11000 points, outperforming even the RTX 3080 by a significant margin.

If the rumored specifications are anything to go by, then the AD104 GPU might be a perfect choice for PC gamers who are on a tight budget and are willing to grab something that offers exceptional performance at a reasonable price tag.

Nvidia GeForce RTX 40 Series: Expected Release Date

Although several claims made by reliable hardware leakers like @greymon55 and @kopite7kimi adhered to the fact that the RTX 40 Series lineup will arrive within Q3 of 2022, there hasn’t been any official announcement from Nvidia till date.

Based on the latest information from Moore’s Law Is Dead, the RTX 4090 is supposedly the first GPU to be unveiled by Nvidia, somewhere in October 2022.

While earlier leaks suggested that the RTX 4080 and 4070 might launch in Q1, 2023, recent reports proclaim that we are likely to see a paper launch for both of these GPUs by November 2022.

One of the major reasons for such inconsistencies regarding the launch of Nvidia’s RTX 40 Series lineup is partly due to the overproduction and saturation of GPUs in the used market amidst the recent cryptocurrency crash.

By delaying the release date of the RTX 40 Series GPUs even further, Nvidia will have gained an additional few months to sell the existing RTX 30 Series stock without incurring a huge financial loss.

Nvidia GeForce RTX 40 Series: Estimated Pricing Structure

Before indulging ourselves in the estimated pricing structure for Nvidia’s RTX 40 Series lineup of GPUs, let us take a look at the current pricing for the RTX 30 Series GPUs:

Based on the aforementioned pricing structure for Nvidia’s RTX 30-Series GPUs, we can expect a similar price tag for the more budget-oriented RTX 4050 and 4060 or even the mid-range RTX 4060Ti, 4070, and 4070Ti from the entire RTX 40 Series lineup.

However, if we take into consideration specific parameters like inflation and component costs over the past few years, we might see a price bump of about $50-$100 for the rest of the stack.

As of now, we strongly believe that the RTX 4090/4090Ti might arrive with a hefty price tag of $1999/2499 whereas the RTX 4080/4080Ti should benefit from competitive pricing of $799/$1499.

Since the estimated pricing structure for these GPUs is heavily reliant on major speculations, we recommend that you take such claims with a grain of salt.