ACCORDING to The Information, the meteoric surge in demand for AI data centre computational power, compounded by a severe deficit in critical memory components, has allegedly compelled NVIDIA to defer the launch of its RTX 50-series revisions—widely anticipated as the RTX 50 Super series—originally slated for 2026.
The report suggests NVIDIA may refrain from introducing any novel consumer‑grade graphics hardware throughout the current calendar year, marking a historic hiatus in the company's gaming portfolio for the first time in three decades. It attributes the decision to pragmatic fiscal viability, noting that generative AI is rapidly becoming ubiquitous and the appetite for high‑tier AI acceleration silicon remains insatiable.
These AI accelerators reportedly command a profit margin of around 65%, far exceeding the 40% yielded by gaming GPUs, reinforcing the prioritisation of high‑margin AI ventures amidst finite production capacity.
The piece also highlights that gaming GPUs constituted 35% of revenue from January to September 2022, versus 8% for January to September 2025, underscoring the shift in focus amid a broader memory bottleneck that is driving a global move away from standard DRAM toward High Bandwidth Memory (HBM) for AI silicon and away from GDDR memory for consumer graphics.
It notes NVIDIA is seemingly reluctant to “squander” wafer and memory quotas on the lower‑margin gaming sector, with potential knock‑on effects including the cancellation of the RTX 50 Super and a possible postponement of the RTX 60-series, which was projected for late 2027, leaving the existing RTX 50-series as the consumer market’s mainstay for an unusually extended period.