Nvidia shifts priorities
- Jensen Huang described Nvidia as executing an 'AI foundation portfolio', expanding investments beyond chips into AI firms. - Reports say Nvidia may skip its 2026 GeForce consumer GPU launch as AI accelerators and memory shortages take precedence. - The change signals an industry reallocation of scarce silicon and memory toward AI infrastructure rather than traditional gaming cadence. (finance.yahoo.com) (digitaltoday.co.kr)
Nvidia is steering more of its money, chips and memory toward artificial intelligence infrastructure, even as reports say its next GeForce gaming cycle could slip. (finance.yahoo.com) (digitaltoday.co.kr) Chief executive Jensen Huang said Nvidia is building an “AI foundation portfolio” that spreads bets across many artificial intelligence companies instead of trying to pick one winner. Yahoo Finance reported April 19 that Nvidia has taken multibillion-dollar stakes in companies including OpenAI and Anthropic and has also invested in suppliers such as Marvell, Lumentum and Coherent. (finance.yahoo.com) That investment push is landing alongside a manufacturing squeeze. Digital Today reported April 18 that Nvidia may skip a 2026 GeForce consumer graphics launch for the first time in roughly 30 years as artificial intelligence accelerators and memory constraints take priority. (digitaltoday.co.kr) (nvidia.com) Nvidia’s own numbers show where the business already sits. In its latest annual results, the company said fiscal 2026 revenue reached $215.9 billion, including $62.3 billion in fourth-quarter data center revenue, while gaming remained a much smaller business line. (investor.nvidia.com) The bottleneck is not just chip wafers. Artificial intelligence servers also need advanced memory, especially high-bandwidth memory, which stacks DRAM chips vertically to move data faster for training and inference workloads. (news.skhynix.com) (investors.micron.com) Memory makers are openly orienting production around that demand. SK hynix said in January that 2026 demand for HBM3E and HBM4 would drive an “AI memory supercycle,” and Samsung used Nvidia GTC in March to showcase HBM4E and its Nvidia partnership for next-generation data centers. (news.skhynix.com) (news.samsung.com) That helps explain why consumer graphics cards are being discussed in terms of supply cuts and delays. VideoCardz, citing board-channel reports and earlier reporting from The Information, said Nvidia had reduced planned RTX 50-series supply in 2026 amid memory shortages and a need to reserve components for its artificial intelligence chip business. (videocardz.com 1) (videocardz.com 2) GeForce has been central to Nvidia’s identity since the GeForce 256 launched in August 1999, and the company marked that history with a “25 Years of GeForce” campaign. A missed annual cadence would break with a long consumer rhythm even if Nvidia keeps selling existing RTX 50 products. (nvidia.com 1) (nvidia.com 2) Nvidia has not publicly announced a cancellation of a 2026 GeForce generation. What it has made public is a company increasingly organized around artificial intelligence systems, their suppliers and the capital needed to keep that buildout moving. (investor.nvidia.com) (finance.yahoo.com)
