Future TechnologyNvidia shipped zero new gaming GPUs this year, the first time since 1999
- 2026 is the first year since 1999 that Nvidia has not released a new consumer graphics card
- Samsung, SK Hynix, and Micron have sold out their entire 2026 HBM output to AI data centre customers
- The RTX 60 series based on the Rubin architecture is now expected in 2028 at the earliest
Twenty-seven years of new GPUs, then nothing
Nvidia has shipped at least one new consumer graphics card every year since 1999. The GeForce 256, the 8800 GT, the GTX 1080, the RTX 3080. Every generation brought something for gamers to argue about. In 2026, there is nothing to argue about, because nothing shipped.
No RTX 50 Super refresh. No mid-range 5060 Ti. The company that built its name on gaming GPUs went an entire calendar year without releasing one.
HBM ate the supply chain
The reason is three letters: HBM. High Bandwidth Memory is the stacked DRAM that powers Nvidia's AI accelerators, and Samsung, SK Hynix, and Micron have sold out their entire 2026 production run to data centre customers. There is no spare memory capacity for consumer cards.
Nvidia's AI business generates significantly higher margins than gaming. When memory prices are climbing and supply is constrained, the maths is simple: every gigabyte of HBM allocated to an H200 or B200 earns more than it would inside an RTX 5090 sitting on a shelf.
The RTX 60 series is now a 2028 story
The RTX 50 Super series, which would normally have filled this year's product calendar, has been shelved. The next-generation RTX 60 series, codenamed Rubin, was originally scheduled to begin mass production in late 2027. Current supply projections push that into 2028 at the earliest.
For anyone still running an RTX 30 or 40 series card, the RTX 5090 and 5080 from earlier this year are the only new options until 2028 at the soonest.
What this tells you
Nvidia's market cap crossed 4 trillion dollars in 2026. Its gaming division, once the company's entire identity, now accounts for a shrinking slice of revenue. Skipping a full year of consumer GPUs is not a supply chain accident. It is a resource allocation decision, and the decision was that AI infrastructure comes first.
The company that made its name selling graphics cards to gamers now makes most of its money selling AI compute to hyperscalers. That gap is only getting wider.
Some links in this article are affiliate links. We may earn a small commission at no extra cost to you.
