Nvidia Rubin GPU Faces Possible Delay As HBM4 Supply Falls Short, While Google TPU Demand Surges In Race For TSMC Capacity: Report

Nvidia Corp.'s (NASDAQ:NVDA) next-generation Rubin AI GPU platform could reportedly face production delays due to supply constraints in next-generation memory.

HBM4 Supply Constraints Could Slow Nvidia's Rubin Rollout

Nvidia's Blackwell successor, the Rubin GPU platform, is seeing downward revisions to wafer starts due to next-generation HBM4 memory supply coming in below expectations, Taiwan's Commercial Times reported, citing supply chain sources.

Suppliers are reportedly redesigning certain base-die components used in the memory stacks, a technical adjustment that could delay shipments by roughly one quarter.

As a result, Nvidia is said to be scaling back initial Rubin wafer production while increasing output of its current Blackwell GPUs instead of releasing manufacturing capacity.

Nvidia did not immediately respond to Benzinga's request for comments.

Cloud Giants Accelerate Custom AI Chip Development

At the same time, cloud service providers are intensifying efforts to develop their own AI chips to reduce reliance on Nvidia.

Alphabet Inc.'s (NASDAQ:GOOG) (NASDAQ:GOOGL) Google's TPU processors are seeing rapidly rising demand, driven by their lower total cost of ownership for large-scale AI workloads, the report said.

The company's next-generation TPU chips — developed with partners including MediaTek and Broadcom Inc. (NASDAQ:AVGO) — are expected to enter mass production soon.

Meanwhile, Meta Platforms Inc. (NASDAQ:META) plans to launch four new MTIA-series AI chips over the next two years as it expands its in-house AI infrastructure.

Industry insiders say Google's growing chip demand could make it one of Taiwan Semiconductor Manufacturing Co.'s (NYSE:TSM) largest customers in the coming years.

Google did not immediately respond to Benzinga's request for comments

TSMC Expands Advanced Manufacturing And Packaging

To support surging AI demand, TSMC is expanding its advanced manufacturing capacity, including multiple new 2-nanometer fabrication plants scheduled to begin mass production in the near future.

The company is also reportedly increasing advanced packaging capacity, including CoWoS technology used in AI accelerators, to help ease bottlenecks as competition for semiconductor production intensifies.

TSMC did not immediately respond to Benzinga's request for comments.

Price Action: Shares of Nvidia closed at $180.25, down 1.59% on Friday, with the stock edging slightly lower in after-hours trading to $180.20, down 0.029%, according to Benzinga Pro.

Benzinga Edge Stock Rankings indicate that NVDA is showing weakness across the short, medium and long-term timeframes, while its Growth score stands in the 97th percentile.

Disclaimer: This content was partially produced with the help of AI tools and was reviewed and published by Benzinga editors.

Photo courtesy: bluestork on Shutterstock.com