NVIDIA is reportedly considering a significant shift in its supply chain, potentially selling graphics processing units (GPUs) without pre-attached video memory (VRAM) to its manufacturing partners, a move that could heighten costs and market instability for graphics cards worldwide.
This potential strategy aims to circumvent a severe global shortage of GDDR memory, a critical component for modern graphics cards.
Industry reports indicate that the cost of these memory modules has more than doubled in recent times.
By separating the GPU from the VRAM, NVIDIA could avoid making the final price of its graphics cards prohibitive or unviable for certain versions, especially given the escalating memory costs.
The rumored plan would involve NVIDIA supplying only the core graphics chip, leaving partners to procure and assemble the memory independently.
This approach offers NVIDIA greater flexibility in its contracts, allowing it to maintain centralized control over the GPU while partners negotiate memory acquisition based on market conditions, volume, or demand.
If implemented, the strategy could lead to several significant changes in the graphics card market.
Consumers might see higher or more unstable prices due to increased logistical complexities involved in separate component procurement.
The number of available graphics card variants could decrease, as smaller manufacturers might struggle to secure memory in sufficient quantities or at competitive prices.
This could also create a disparity between large and small manufacturing partners, with larger firms leveraging their purchasing power to maintain a stable supply, while smaller ones face scarcity or inflated prices.
Furthermore, graphics cards assembled with VRAM from different suppliers could exhibit minor variations in performance or stability, leading to less standardization than currently seen.
As of now, NVIDIA has not officially confirmed these rumors.
The speculations originated from a leak attributed to the account “Golden Pig Upgrade Pack,” which has a track record of accurate predictions.
The memory crisis extends beyond consumer GPUs, impacting demand for components in artificial intelligence, data centers, and other markets, further complicating the global supply chain.