Nvidia reportedly halts RTX 5090 production to prioritize AI workstation silicon
Nvidia has reportedly ceased manufacturing the GeForce RTX 5090, including Chinese market variants, to reserve GB202 silicon for high-margin professional AI cards. A leaked factory notice suggests the RTX 5080 24GB will become the new top-tier consumer GPU, while the 16GB version is discontinued. This move exacerbates existing memory supply constraints, delaying rumored RTX 50 Super refreshes and pushing next-gen Rubin-based cards to late 2027 or 2028.
Key points
- Leaked factory notices from leakers @hongxing2020 and @Zed__Wang claim Nvidia has stopped all RTX 5090 production, including D variants for China.
- The GB202 chip previously used in the RTX 5090 will now be exclusively allocated to data center and professional workstation cards.
- The RTX 5080 24GB is expected to replace the 5090 as the highest-end consumer GeForce card, while the 16GB version is reportedly discontinued.
- Nvidia has not officially confirmed these changes, and Tom's Hardware reported that the company did not respond to comment requests.
- An upcoming RTX 50 Super refresh, potentially featuring a 24GB RTX 5080 Super, has been delayed due to memory supply constraints.
- Next-generation Rubin-based RTX 60-series cards are rumored to arrive no earlier than late 2027, possibly extending to 2028.
Background
This development follows a significant price surge for the RTX 5090 in September 2026, where retail prices reached approximately $7,500 due to AI server demand and supply issues. Earlier reports in October 2026 also indicated that Nvidia was halting production to prioritize high-margin AI workstation cards, aligning with the current leaked information. The broader industry is facing memory shortages, which have also impacted console timelines and other hardware availability.
Why it matters
The discontinuation of the RTX 5090 signals a strategic shift by Nvidia to prioritize high-margin AI and professional markets over consumer gaming hardware. This decision may lead to further scarcity and price increases for existing stock, while delaying the availability of next-generation consumer GPUs until 2027 or later. It highlights the ongoing tension between consumer demand and corporate profit optimization in the face of global memory supply constraints.
What to watch
Consumers may see the RTX 5080 24GB become the new flagship consumer card, though its release is uncertain due to memory costs. The RTX 50 Super refresh is likely to be delayed further. Next-generation Rubin-based RTX 60-series cards are expected to launch in late 2027 or 2028, potentially leaving a long gap in high-end consumer GPU updates.
Want the full story? Read the original reporting
Read on GamesIndustry.biz