THE CRUNCH
Nvidia has reportedly stopped allocating its GB202 chip to the GeForce RTX 5090, redirecting the largest Blackwell graphics processor to workstation and data centre cards instead. The claim comes from leaker MEGAsizeGPU on X, who posted "No more GB202 for 5090, confirmed" and said all future GB202 supplies are RTX Pro exclusive. A second leaker, hongxing2020, shared an apparent factory notice saying Nvidia will no longer sell any RTX 5090 model, including China-market D variants. Nvidia has not confirmed any of this, and Tom's Hardware says the company has yet to respond to a request for comment.
The same GB202 chip powers the RTX Pro 6000 with 24,064 CUDA cores and 96 GB of memory, which can sell for around $15,000, as well as the newly unveiled RTX Pro 5500. That card matches the RTX 5090's 21,760 CUDA cores but carries 84 GB of memory instead of 32 GB, and targets professional visualisation and AI work at far higher prices. Shifting scarce chip allocation to those products is an obvious commercial move, though it would leave the GeForce line without a GB202-class flagship.
According to the leaked notice, the top consumer card would become an RTX 5080 24GB, while the 16 GB RTX 5080 would be discontinued. That sounds like the long-rumoured RTX 5080 Super, which was expected to pair 24 GB of GDDR7 with the smaller GB203 chip, though it is not clear whether the two are the same card. Super cards reportedly reached board partners in July before being put on hold over GDDR7 pricing.
Even if the reports are accurate, RTX 5090 boards will not vanish overnight, as partners and retailers hold inventories that could sustain availability for some time. But with no fresh supply, availability would decline and prices would likely rise. Anyone wanting the current top GeForce card may be better off buying sooner rather than later, and with next-generation Rubin-based GeForce cards rumoured for late 2027 or later, the gap could last a while.