A persistent shortage of memory chips, driven by high demand from AI data centers, is reshaping the GPU market.
Nvidia and AMD are adjusting their product strategies, with reports indicating production shifts toward lower-VRAM models, potential discontinuation of some higher-VRAM cards, and price increases across multiple brands. Both companies have acknowledged strong demand and constrained memory supply.
Nvidia Product Strategy Adjustments
Production Focus Shift
According to industry reports, Nvidia is adjusting its RTX 50 series production strategy for 2026. The company is reportedly reducing shipments of higher-VRAM models, including the RTX 5070 Ti and RTX 5060 Ti 16 GB, and shifting focus toward the RTX 5060 Ti 8 GB and RTX 5060 8 GB models. Reports indicate these 8 GB cards are positioned as the primary sales drivers for the RTX 50 series, particularly in the Chinese market.
The RTX 5060 is expected to be the most widely available GPU in the RTX 50 lineup, with the RTX 5060 Ti 8 GB following closely. This strategy is reportedly linked to the 8 GB models requiring fewer GDDR7 memory modules (four modules compared to eight for 16 GB models), making them more cost-effective to produce under current market conditions.
Discontinuation Reports
Hardware Unboxed reported that Nvidia has effectively discontinued the RTX 5070 Ti and RTX 5060 Ti 16 GB due to memory supply shortages. According to the report, ASUS informed Hardware Unboxed that the RTX 5070 Ti is experiencing a supply shortage and has been placed into "end of life" status, with no further production planned. Australian retailers reportedly stated the RTX 5070 Ti is unavailable from partners and distributors, a situation expected to continue through at least Q1. ASUS also reportedly indicated it no longer plans to produce the RTX 5060 Ti 16 GB.
Nvidia's Official Statement
"Demand for GeForce RTX GPUs is strong, and memory supply is constrained. We continue to ship all GeForce SKUs and are working closely with our suppliers to maximize memory availability."
Nvidia has provided the above statement to multiple outlets. The company did not confirm or deny the discontinuation of the RTX 5070 Ti and RTX 5060 Ti 16 GB.
Older GPU Production
Reports indicate Nvidia is exploring options to increase production of older GPUs, such as the RTX 3060. A leak on the Board Channels forum suggests Nvidia may re-release the RTX 3060 with 12 GB of VRAM, with an estimated launch timeframe of June or July 2026. The RTX 3060, originally released over five years ago, uses a GA106 Ampere GPU with GDDR6 VRAM and a 12 GB memory buffer. The re-release faces questions about pricing, as the RTX 3060 would need to be priced substantially below $250 to compete with current-gen cards, but material costs remain high due to the ongoing memory shortage.
Nvidia VRAM Variant Rumors
RTX 5060 and 5060 Ti
According to reports from TechPowerUp and leakers, Nvidia may be planning to launch new versions of its RTX 5060 and RTX 5060 Ti cards featuring 9 GB of VRAM, an increase from the typical 8 GB configuration. This change would reportedly use new 3 GB GDDR7 memory chips. A separate report suggests the memory bus width for these cards may be adjusted from 128-bit to 96-bit, which could result in reduced overall memory bandwidth compared to the original 8 GB versions. Both sources cited claim the new cards could become available by the end of May or early June 2026, with a potential debut at the Computex trade show.
RTX 5050
A leak indicates the Nvidia RTX 5050 desktop GPU may receive a VRAM upgrade to 9 GB using 3 GB GDDR7 modules on a 96-bit bus. Hardware leaker Zed__Wang (MEGAsizeGPU) reported on April 17, 2026, that the rumored RTX 5050 9 GB has been delayed, and its launch is now uncertain. According to the leaker, Nvidia has not officially canceled the card but appears hesitant about its release.
RTX 5070 Mobile
Nvidia has announced a new 12 GB version of the mobile GeForce RTX 5070, in addition to the existing 8 GB version. Both configurations share the same core specifications: a GB206 chip based on the Blackwell architecture with 4,608 CUDA cores, 36 ray tracing cores, 144 Tensor cores, and a 128-bit memory interface. The 12 GB variant uses 3 GB GDDR7 memory chips instead of 2 GB chips. The company stated that the expansion addresses strong demand for GeForce RTX laptop GPUs and constrained memory supply. Laptops with the 12 GB RTX 5070 are expected to ship starting in June 2025.
GPU Pricing Trends
Current Market Pricing
Market observations show significant price increases for existing Nvidia RTX 50 series models. The RTX 5090, RTX 5080, and RTX 5070 Ti models are selling above their Manufacturer's Suggested Retail Price (MSRP) at retailers like Newegg and Amazon, with some high-end cards reaching up to $5,000-$8,000. While RTX 5070 and some RTX 5060 Ti variants were previously available below MSRP, they have also experienced price hikes.
Board Partner Price Increases
Asus and Gigabyte are reportedly planning to increase prices for RX 9000 and RTX 50-series graphics cards by up to 15% by the end of January, according to NotebookCheck citing a report from Taiwan. The extent of the price increase is expected to depend on VRAM capacity, with models with 16 GB or more potentially seeing a full 15% increase, while those with 8 GB could experience a 10% increase. MSI has already implemented price increases on its latest-generation GPU models.
AMD Price Adjustments
AMD is reportedly planning a second price increase for its Radeon GPUs in Q1 2026, following an initial price adjustment in January 2026. In January 2026, most AMD brand manufacturers raised their cost prices for the majority of Radeon GPU models by approximately 5-10%. Further price increases are anticipated by some manufacturers starting in February or March 2026.
ASUS has increased prices for its Radeon RX 9070 XT OC Edition and TUF GAMING models in the United States. The ASUS Prime Radeon RX 9070 XT OC Edition saw its price rise from $799.99 to $939.99 (a 17.5% increase), and the ASUS TUF GAMING Radeon RX 9070 XT experienced a price increase from $849.99 to $989.99 (a 16.5% increase). ASUS attributed these changes to AMD's GPU price increases amidst the memory crisis.
Zotac Korea's Supply Warning
Zotac Korea has informed customers that both memory and GPU availability are expected to decline. The company stated that memory supply will not be sufficient, GPU shipments will be reduced, and some graphics card models may become unavailable for extended periods. Zotac also confirmed a reduction of reward points for new GPU purchases to 0%, stating this measure aims to mitigate price increases driven by current market conditions.
Industry Context
The demand for RAM and other computer components from data center infrastructure companies, driven by the AI boom, has led memory manufacturers to shift production towards high bandwidth memory. This shift has resulted in increased prices for consumer RAM kits, GPUs, and SSDs. Micron Technology announced in December that it would wind down its consumer-facing Crucial brand to focus on supplying components to the AI industry.
ASUS is the first of Nvidia's add-in board (AIB) partners to comment on the memory shortage. A recent rumor suggested that Nvidia may require its board partners to source memory independently, rather than providing both the GPU die and memory as historically done.