As the debut of the next generation RTX 40 series graphics cards based on the Ada Lovelace architecture approaches, the rumor mill has been spinning. According to new information, the next-generation cards will be released even sooner than thought, possibly in August if not July.

This information comes from well-known leaker Kopite7kimi, who suggested a probable launch date in early Q3, which would be July or around that time. Gamescom is also scheduled for August, so an August release could be a possibility. If the launch happens in early or mid-Q3, as this new information suggests, it will happen far sooner than expected. We had anticipated a Q4 release, possibly in September or October, based on historical launch dates. There are advantages to NVIDIA's early debut, which gives it a leg up on competitors AMD and even Intel, who are about to release their Arc series of graphics cards. This year, Computex will return, but with COVID restrictions, so anticipate a smaller event. NVIDIA is scheduled to give a keynote talk on May 23, although it's unclear whether any information about the RTX 40 series will be presented.

The high-end cards are expected to be released first. They generate the halo headlines, yet production numbers aren't as large as they would be for a mainstream release. This is supported by the fact that we've only heard a few code names and possible shader counts for the other GPUs in the range.

In fact, there is another full-fat AD102 SKU with 900W TGP, 48G 24Gbps GDDR6X, 2*16pin and higher frequency.
But no one knows whether it will become an actual product. Because the test board of AD102 has more than two 16pin connectors, so everything is possible.

— kopite7kimi (@kopite7kimi) April 27, 2022

Everything points to the high-end AD102 being a beast of a GPU. Ada Lovelace was once thought to be more of an evolution of Ampere than something entirely distinct, although that seems to have altered in recent months. Nvidia appears to be trying all out to keep its title of fastest GPU. Surely, a 900W GPU is overkill, but now that the RTX 3090 Ti has surpassed the 450W mark, those 600W reports are becoming increasingly plausible.

With a 600W TDP, third-party cards might feature dual 16-pin power connectors, allowing for far over 1KW of power delivery. Allow that to sink in. The RTX 4090 may require all 600 watts to power itself. By alone, 18432 cores and 96MB of L2 cache would give significant performance advantages over the 450W RTX 3090 Ti's 10752 cores and 6MB of L2 cache. However, when you factor in other architectural upgrades like updated RT and Tensor cores, the RTX 4090 is shaping up to be a pixel-pushing monster. To succeed, AMD will have its work cut out for it.