Intel Joins Forces with NVIDIA to Boost AI Chip Production Amid Soaring Demand

August 3, 2024
Intel Joins Forces with NVIDIA to Boost AI Chip Production Amid Soaring Demand
  • Intel Foundry Services (IFS) is set to begin supplying NVIDIA as early as next month, establishing itself as a crucial partner in addressing NVIDIA's packaging needs.

  • This partnership comes at a time when NVIDIA is struggling to meet the soaring demand for AI chips, with existing suppliers like TSMC unable to fully satisfy the requirements.

  • The increasing demand for NVIDIA's AI products has prompted the company to seek new collaborations to ensure optimal supply and enhance ordering experiences for its clients.

  • As part of this collaboration, Intel will produce 5,000 H100 wafers per month to help meet the demand for NVIDIA's AI GPUs.

  • This output significantly surpasses the capabilities of competitors like TSMC, positioning Intel as a formidable player in the market.

  • NVIDIA is particularly interested in Intel's Foveros 3D stacking technology, which competes with TSMC's CoWoS-S packaging process, and will be crucial in fulfilling its requirements.

  • By adding Intel Foundry as a supplier for its CoWoS packaging technology, NVIDIA aims to expand its supply chain and mitigate risks associated with reliance on a single source.

  • This partnership represents a significant opportunity for Intel to enhance its financial standing amid ongoing challenges in the semiconductor market.

  • Intel's CEO, Pat Gelsinger, has emphasized that IFS aims to evolve into an 'AI factory,' focusing on advanced packaging technologies tailored for the AI market.

  • Under its IDM 2.0 strategy, Intel is heavily investing to position IFS as a global leader in semiconductor manufacturing, which has attracted NVIDIA's attention.

  • The collaboration may also encompass NVIDIA's Hopper generation AI products, including the highly sought-after H100 GPUs, thereby significantly enhancing NVIDIA's production capacity.

  • NVIDIA has faced supply bottlenecks due to its dependence on TSMC for high-demand products like the H100 and A100, making this partnership with Intel even more critical.

Summary based on 2 sources


Get a daily email with more AI stories

More Stories