NVIDIA has issued a price increase notice to downstream graphics card manufacturers for its GPU bundles — the combined packages of a graphics processor and its video memory (VRAM) that board partners turn into finished cards — citing rising memory prices as the main driver (United Daily News).

The notice, reported on 29 July 2026, follows a pattern of escalating costs on NVIDIA's professional and data-centre products. In mid-June, NVIDIA raised the official price of its RTX PRO 6000 Blackwell GPU — a high-end card built for professional workloads like 3D rendering and AI development — to US$13,200, an increase of roughly 55% (United Daily News).

Memory is the stated culprit. The chips that sit alongside a GPU's processing die, known as VRAM, have become more expensive as demand from AI data centres strains global supply. NVIDIA passes those costs to the manufacturers who build consumer graphics cards under brands like ASUS, MSI and Gigabyte, who in turn decide how much of the increase reaches retail shelves.

The price pressure on professional hardware stands in contrast to what is happening on the rental side of NVIDIA's business. Rental prices for the company's B200 GPU — a data-centre chip used for AI training and inference — dropped more than 30% from their peaks, and prediction markets have bet against a return to May 2026 highs (United Daily News).

For gamers, the picture is complicated. NVIDIA planned to reduce gaming GPU supply in 2026, marking the first time in 30 years the company would not launch a new gaming GPU in a calendar year (United Daily News). Fewer chips reaching the gaming channel, combined with rising memory costs passed through to board partners, puts upward pressure on the prices consumers pay for the next graphics card they buy.

NVIDIA has continued showcasing new partner products through 2026 — GeForce RTX cards and laptops at CES in January and again at Computex in June — but no official consumer GPU pricing announcement has appeared on the company's own domains this year. The GeForce showcases listed partner products at various price points without setting recommended retail prices of its own (NVIDIA GeForce).

Beyond graphics cards, NVIDIA has stayed busy on other fronts. The company held its GTC 2026 conference in March with live updates on its official blog (NVIDIA Blog). In June, NVIDIA and Microsoft announced a collaboration to "reinvent Windows PCs for the age of personal AI," integrating NVIDIA's processors into the next generation of Windows hardware (NVIDIA Investor Relations). NVIDIA also reported its fiscal 2026 fourth-quarter results in February (NVIDIA Investor Relations).

What ties these threads together is supply. Memory prices are climbing. Gaming GPU supply is tightening. Data-centre rental rates are easing. For anyone shopping for a graphics card — or watching the companies that make them — the question is how much of NVIDIA's rising input costs land at the checkout.