Rumour has it that the Nvidia RTX 5090 will surpass the potent RTX 4090 by a significant margin, making it a game-changer. The next flagship GPU from Nvidia is expected to be a truly high-performance successor, as shown by spec leaks that suggest a 40% rise in energy consumption and increased processing power.
Although a formal release date has not yet been set, speculation is that it will happen in late 2024 or early 2025, which is in line with Nvidia’s typical 18–24 month cycle. According to certain rumours, the RTX 5080 might arrive before the 5090, which could cause the introduction to be postponed until 2025. It is still possible that both models may be unveiled at CES 2025.
It is anticipated that the RTX 5090 would have a sturdy 16-pin 12V-2×6 power connector, Nvidia’s Blackwell architecture, and performance-oriented improvements such as a 448-bit memory bus, 28Gbps GDDR7 memory (33% faster than the RTX 4090’s 21Gbps), and 28GB of VRAM. It’s uncertain if this holds true for the 5090, but leaks indicate that Nvidia may use a multi-chiplet module (MCM) design for some Blackwell GPUs.
Notably, the 5090 might have a faster base clock speed, which could result in performance that is 1.5–1.7 times that of the RTX 4090, a significant generational leap. Despite these improvements, some suggest it may still be smaller than the 4090, which would be an impressive achievement.
The launch volume of the RTX 5090 may be constrained by Nvidia’s emphasis on AI GPUs. There may be a shortage of the 5090 when it is released if the RTX 5080 sees significant stock. A strong, state-of-the-art GPU that is anticipated to boost performance is depicted in these leaks for the time being. The tech community is eagerly awaiting more information, but it won’t be known until Nvidia confirms these specifics whether the 5090 lives up to the expectations.