SK hynix has major HBM order to Broadcom: which is making AI chips for Google, Meta, ByteDance

SK hynix rumored to deliver huge HBM memory order to Broadcom in 2025: which is building new AI chips for Google, Meta, and ByteDance.

SK hynix has major HBM order to Broadcom: which is making AI chips for Google, Meta, ByteDance
Comment IconFacebook IconX IconReddit Icon
Gaming Editor
Published
2 minutes read time

As an Amazon Associate, we earn from qualifying purchases. TweakTown may also earn commissions from other affiliate partners at no extra cost to you.

TL;DR: ByteDance plans to develop two custom AI chips using TSMC's 5nm process by 2026. OpenAI is collaborating with Broadcom and TSMC to create its first AI chip, while Broadcom is introducing advanced 3.5D XDSiP technology with multiple compute tiles and HBM sites. Broadcom also unveiled a large AI chip with a new XPU design.

SK hynix is delivering a huge HBM memory order to Broadcom in 2025, which it will use on its new AI chips that it is making for cloud service providers (CSPs) including Google, Meta, and TikTok parent company ByteDance.

SK hynix has major HBM order to Broadcom: which is making AI chips for Google, Meta, ByteDance 61
2

SK hynix already has NVIDIA as a major client of its industry-leading HBM memory chips, but sources of TheElec report that the South Korean memory giant will deliver new HBM to Broadcom in 2H 2025, with the US chipmaker to use the AI memory chips in new AI chips for Google, Meta, and ByteDance.

The deal will see SK hynix push its HBM production capacity to 160,000 - 170,000 wafers in 2025, up from the previous target of 140,000 - 150,000. SK hynix is providing most of the HBM memory chips it makes to NVIDIA for use in H200, GB200, B200, GB200, and other AI chips and AI servers, but now Broadcom wants some of that delicious HBM memory pie for its CSP clients.

TheElec says that SK hynix might delay the installation of its new 1c DRAM equipment, in order to address the sudden surge in demand from Broadcom. Major cloud service providers (CSPs) have been pushing their own AI infrastructures, which is increasing the demand for AI-specific ASICs and HBM. Google is building HBM3E into its Trillium TPU, and Amazon AWS is using HBM in its in-house Trainium chipset for AI learning purposes.