Nvidia Corp. senior vice president Jeff Fisher stated that the GeForce RTX 5090 graphics processing unit has not been discontinued, addressing market rumors that production of the flagship hardware had ceased to reallocate silicon capacity to AI accelerators. Fisher’s remarks, reported by Taiwan-based outlets 4Gamers and SJTV, arrive as the high-end gaming card remains absent from major retail channels and commands secondary-market prices reaching thousands of dollars above retail expectations.
Production Status and Retail Pricing Pressures
Rumors regarding the termination of the RTX 5090 circulated following reports that Nvidia had delayed the rollout of its RTX 50 Super series, stoking concerns among consumers that the top-tier consumer GPU would vanish from shelves. Speaking after an industry event, Fisher clarified that production remains active, though he acknowledged severe constraints within the manufacturing supply chain. “Supply chain capacity is somewhat tight, but Nvidia is doing everything possible to bring this product to market,” Fisher told 4Gamers. “Even though there are other variants of the RTX 5090 chip, production has not stopped, and Nvidia will continue shipping.” Tom’s Hardware reported that it reached out to Nvidia for additional official comment regarding the retail shortages but received no immediate response.
Despite Nvidia’s assurances, retail availability remains fractured. The RTX 5090 is entirely absent from official Nvidia marketplaces, Best Buy storefronts, and Micro Center inventories. Where the card does appear on third-party channels, pricing reflects extreme scarcity. Amazon listings priced third-party units around $8,500, while Newegg listed configurations near $7,300 with shipping estimates pointing to overseas fulfillment, leaving buyers exposed to lingering United States import tariffs.
Power Connector Modification and Thermal Mitigations
Beyond supply constraints, technical scrutiny has focused on the thermal demands of the RTX 5090 and its reliance on the 16-pin power connector (12VHPWR and the revised 12V-2×6 standard), which has experienced documented melting incidents under high loads. Addressing these engineering challenges, a hardware modder known as DallasGrave developed a custom quad 8-pin power conversion kit designed to distribute electrical load away from the single high-density connector.

Unlike standard aftermarket adapter cables, the direct-mount kit solders straight onto the graphics card’s printed circuit board power rails. The system channels a 600-watt load across four traditional 8-pin PCIe connectors rather than forcing current through a single 16-pin interface. According to PCI-SIG specifications, a standard 16-pin connector houses six power pins rated for a theoretical maximum current capacity of 684 watts, forcing individual pins to operate near strict thermal limits during peak operation. By contrast, a four-connector 8-pin array provides 12 distinct power pins, effectively halving the amperage per pin to approximately 4.17 amperes under a 50-ampere load, which significantly reduces thermal accumulation and terminal degradation risks.
The developer designed the conversion kit to trick the 16-pin sense pins into recognizing a full 600-watt delivery while testing successfully up to 800-watt configurations via shunt modifications. The hardware functions either independently or alongside the stock 16-pin interface to balance electrical distribution. Although DallasGrave intends to offer models compatible with a wide range of market variants, including Nvidia’s Founders Edition cards, official release dates, pricing tiers, and installation logistics remain unannounced.
Frequently Asked Questions
Why is the RTX 5090 missing from major retailers like Best Buy and Micro Center?
Retail inventories dried up amid supply chain constraints and rumors of production shifts toward AI accelerators. While Nvidia executive Jeff Fisher confirmed that manufacturing continues, actual stock allocations remain extremely limited across official storefronts.
How much does the RTX 5090 cost on secondary markets?
Third-party sellers on Amazon have listed the flagship GPU for approximately $8,500, while Newegg marketplace listings sourced from international vendors hover near $7,300 before accounting for import tariffs and shipping fees.
What does the quad 8-pin conversion kit fix on the hardware?
Created by modder DallasGrave, the direct-mount PCB kit splits a 600-watt power load across 12 separate pins from four traditional 8-pin PCIe cables, reducing current density per pin and mitigating the melting risks associated with compact 16-pin 12VHPWR connectors.
Did Nvidia cancel the RTX 5090 to prioritize artificial intelligence chips?
No. Senior vice president Jeff Fisher explicitly denied rumors of cancellation in statements to Taiwan-based media outlets 4Gamers and SJTV, asserting that Nvidia continues to manufacture and ship the graphics cards despite constrained component capacity.
Nvidia maintains that its consumer graphics ecosystem remains foundational to the company’s broader technological footprint, even as high-performance computing and artificial intelligence segments dominate its quarterly financial growth.