Nvidia Pauses GeForce Refresh in 2026 Amid Memory Shortage

Nvidia will not release a new GeForce GPU family in 2026, breaking an annual release streak that began in the 1990s. The company plans to cut production of gaming graphics cards by as much as 40% as memory suppliers allocate components to Blackwell AI processors.

Nvidia’s computer and networking division, which supplies AI chips, posted an average profit margin of roughly 69% over the past three years. The consumer gaming graphics segment averaged about a 40% margin. Blackwell AI accelerators can carry price tags up to $40,000, while consumer GeForce cards typically retail between about $299 and $1,999.

High Bandwidth Memory (HBM), used in many AI accelerators, requires about four times the number of silicon wafers to produce compared with standard memory used in consumer devices. That production profile reduces the number of memory chips available for gaming cards and other consumer products.

Memory manufacturers have warned the shortage could be extended. One major supplier has flagged a long-term shortfall in the types of memory used for AI compute. Industry forecasts indicate the shortage could push PC prices up roughly 17% this year and cause PC shipments to fall about 10.4%. Some analysts project that entry-level consumer PCs could largely disappear by 2028 if supply constraints continue.

Rivals AMD and Intel have been affected. AMD raised prices across its Radeon RX 9000 series by roughly 10% to 17% depending on the model: the Radeon RX 9070 XT rose about 17%, the RX 9060 XT 8GB about 10%, and the 16GB RX 9060 XT about 14%. David McAfee, who oversees AMD’s Radeon business, acknowledged the company is working with memory suppliers to limit price increases but said maintaining those efforts is becoming increasingly difficult given current supply limits.

Intel abandoned plans for a consumer Arc B770 gaming card and will instead release an Arc Pro B70 workstation card with 32GB of memory aimed at AI workloads, citing limited memory availability and low financial viability for a gaming version.

Stacy Rasgon of Bernstein Research observed that gaming is no longer the primary profit driver it once was, adding, “Every bit of memory that’s out there, I think is really getting prioritized to AI compute.” Tim Gettys, co-host of a gaming podcast, warned that higher-margin AI businesses are drawing corporate focus and said, “If they’re making three times the money and the stockholders are three times happier, then yeah, I do think that they will abandon gaming despite it being what got them there.”

Nvidia introduced its first GPU, the GeForce 256, in 1999. Early consumer demand played a key role in building the company. The current shift to AI hardware follows changes in profit margins and the availability of key components.