HBM shortage forces NVIDIA (NVDA) to prioritise AI accelerators, disrupting gaming GPU roadmap

HBM squeeze forces NVIDIA to prioritise AI accelerators

NVIDIA is reallocating scarce high-bandwidth memory (HBM) to its high-margin AI accelerators, prompting material changes to its gaming GPU roadmap, industry sources say. Reporting by The Information and comments from company spokespeople indicate the firm may postpone or skip a consumer-grade RTX 60 series this year as limited HBM supplies are channelled to Rubin-based AI chips already in mass production. NVIDIA CEO Jensen Huang is saying Rubin AI devices are on track to ship in the second half of the year, and the company is reportedly cutting production of GeForce RTX 50 gaming chips as memory constraints persist.

The memory shortfall is tightening launch timelines and lifting retail prices for current gaming GPUs, as manufacturers balance demand across gaming, data centre and mobile customers. A NVIDIA spokesperson is saying demand for GeForce RTX GPUs remains strong but memory supply is constrained, and that GeForce products are in stock and shipping. Analysts and supply-chain participants warn that expanding HBM capacity requires major wafer starts and packaging changes, meaning shortages could linger into 2025 without rapid industry investment in production scale-up.

NVIDIA’s prioritisation reflects a broader industry shift where HBM is increasingly critical to datacentre and generative AI workloads that deliver higher revenue per chip than consumer products. By directing scarce HBM to AI accelerators, NVIDIA is reinforcing its focus on the most profitable segments of its business while forcing partners and OEMs to adjust gaming and mobile product roadmaps. The reallocation also underscores how memory supply dynamics are becoming a strategic variable in chipmakers’ product planning and manufacturing decisions.

Memory strain ripples through device makers

The HBM crunch is producing knock‑on effects across the electronics supply chain. Qualcomm and Arm warn that HBM shortages could crimp smartphone production, Apple signals it will prioritise higher‑margin iPhone models, and Nintendo flags margin pressure from soaring memory costs. Industry notes and vendor warnings suggest shortages may distort pricing and product introductions across mobile, gaming and consumer electronics unless packaging and wafer‑start capacity expand quickly.

Cloud and AI vendors scale compute demand

At the same time, big tech is dramatically expanding AI compute commitments that increase demand for advanced memory. Alphabet is reporting robust Cloud growth and says enterprise AI licences are rising, while signalling very large capital spending to expand AI computing capacity. That surge in data‑centre investment is intensifying competition for HBM and other advanced components, linking memory supply directly to the pace of AI deployment across the industry.