VFF - The signal in the noise
News

TSMC Eyes Advanced Memory Chips to Deepen AI Supply Chain Control

Read original
Share
TSMC Eyes Advanced Memory Chips to Deepen AI Supply Chain Control

TSMC is planning to manufacture advanced 3D DRAM memory chips at Nvidia's request, marking a potential expansion into memory production for the chipmaker. The move responds to severe memory chip shortages driven by AI hardware demand, which has pushed prices higher and prompted major chip designers to seek additional supply sources. Apple has already raised device prices citing rising memory costs. This would consolidate more of the AI hardware supply chain under TSMC's control.

  • TSMC is drawing up plans to manufacture 3D DRAM chips, a type of advanced memory, at Nvidia's request
  • Memory chip shortages driven by AI hardware boom have sent prices soaring
  • Chip designers including Nvidia and Apple are seeking additional memory supply sources
  • Apple raised device prices earlier this year citing rising memory costs

Memory chip shortages have become a critical bottleneck in AI infrastructure buildout. If TSMC enters advanced memory manufacturing, it would consolidate even more of the AI supply chain under a single company, raising questions about concentration risk and supply chain resilience. The move reflects how AI demand is reshaping semiconductor manufacturing priorities.

For customers like Nvidia and Apple, TSMC's potential entry into memory manufacturing could ease supply constraints and moderate price pressures. For competitors in memory manufacturing, it represents a significant threat to market share. For TSMC, it offers a path to capture more AI-driven spending and deepen customer relationships.

  • TSMC would expand its role in the AI supply chain beyond logic chips into memory, increasing its leverage with customers
  • Memory chip pricing pressure could ease if TSMC successfully brings 3D DRAM production online
  • Existing memory chip manufacturers face competitive pressure from TSMC's potential entry into their market

Monitor whether TSMC formally announces 3D DRAM production plans and any timeline for manufacturing to begin. Track memory chip pricing trends and whether shortages persist or ease. Watch for responses from established memory manufacturers and whether other chip designers besides Nvidia request similar capacity from TSMC.

OneUpAI
OneUp Your Business. Get More Done. OneUp Your Business. Get More Done. OneUp Your Business. Get More Done.
Learn More
Share

Subscribe to the newsletter

The latest stories and analysis, delivered to your inbox.

Free. No spam. Unsubscribe any time.

Related stories

NVIDIA Open Source Tools Drive AI-Native 6G Research
TrendingNews

NVIDIA Open Source Tools Drive AI-Native 6G Research

NVIDIA's open source wireless software tools, including Aerial and Sionna, have reached 1 million downloads and adoption across 8,000 research institutions as developers use them to build 6G networks. The tools enable researchers to test AI-driven approaches to wireless communications, including semantic communications for robotics, neural receivers for signal processing, and power management. Three concrete research projects demonstrate practical applications: King's College London reduced robot task completion time by up to 52.6% using semantic communications, SoftBank achieved 60% throughput gains with neural receivers, and ETH Zurich reduced block error rates by one-third through localized receiver tuning.

by Soma Velayutham· NVIDIA Blog (AI)
NVIDIA Vera CPU Targets AI Agent Bottleneck with Speed and Efficiency
TrendingModel Release

NVIDIA Vera CPU Targets AI Agent Bottleneck with Speed and Efficiency

NVIDIA has released Vera, a CPU designed to handle AI agent workloads that require both fast execution and high capacity. The chip delivers faster agent performance than the fastest x86 CPUs while also supporting 1.6x more agents per watt than comparable processors. Early adopters including Perplexity, DeepInfra, and Modal report performance gains ranging from 1.5x to 2.2x on production agent tasks.

by Hannah Coutand· NVIDIA Blog (AI)
NVIDIA Tackles AI Memory Crunch With Full-Stack Design

NVIDIA Tackles AI Memory Crunch With Full-Stack Design

NVIDIA is addressing AI's growing memory constraints through full-stack codesign, optimizing GPUs, CPUs, networking, storage and software together rather than simply adding more high-bandwidth memory. The approach targets inefficiencies across the memory hierarchy, from model weights to KV cache to retrieval-augmented generation data. Results show agentic throughput per dollar increased up to 30x from Blackwell Ultra to Vera Rubin despite only 3% HBM capacity gains per GPU.

by Amr Elmeleegy· NVIDIA Blog (AI)
Bluetooth gets 3.75x speed boost with new HDT standard

Bluetooth gets 3.75x speed boost with new HDT standard

The Bluetooth Special Interest Group has formally adopted Bluetooth High Data Throughput (HDT) as an optional feature in Bluetooth Core 7.0, tripling the maximum bit rate from 2 Mbps to 7.5 Mbps. The upgrade will enable higher resolution audio streaming while also improving battery life and reliability. First devices supporting HDT are expected in 2027, with wireless audio as the primary initial application.

by Andrew Liszewski· The Verge AI