AMD invests $10 billion+ in Taiwan to accelerate next-gen AI infrastructure

AMD invests more than $10 billion across Taiwan’s semiconductor and AI ecosystem as it looks to expand advanced packaging capacity and push its

AMD invests news image
Lisa Su, chair and chief executive officer of Advanced Micro Devices Inc. (AMD), during the 2026 CES event in Las Vegas, Nevada, US, on Monday, Jan. 5, 2026. At CES, the biggest names in tech will make the case for artificial intelligence for a target audience of investors, corporate clients and, perhaps just as importantly, ordinary shoppers who have yet to be fully sold on the idea of AI-infused gadgets. Photographer: Bridget Bennett/Bloomberg via Getty Images

AMD invests more than $10 billion across Taiwan’s semiconductor and AI ecosystem as it looks to expand advanced packaging capacity and push its next-generation AI infrastructure plans forward. The move ties together CPU development, rack-scale systems and manufacturing partnerships at a time when the company is trying to build out the hardware needed for larger AI deployments.

What AMD is funding

The new investment package is aimed at strengthening strategic partnerships and scaling advanced packaging facilities. AMD says the work will support its wider AI infrastructure plans, with a particular emphasis on improving the technologies that connect and package chips for more efficient systems.

One major part of the effort focuses on next-generation interconnect technology for Venice CPUs. AMD is working with companies including ASE and Silicon Precision Industries on wafer-based 2.5D bridge technology, which is intended to improve power efficiency and help AI systems operate within practical power and cooling limits.

Why the packaging work matters

Advanced packaging may not grab as much attention as a new GPU or CPU launch, but it plays a key role in how modern AI hardware is built. Better interconnect and packaging technology can improve how chips communicate, while also making it easier to design systems that are more efficient at scale.

  • Supports development of next-gen interconnect technology
  • Helps improve power efficiency in AI systems
  • Targets more practical power and cooling requirements
  • Backs AMD’s broader AI infrastructure rollout

Helios and Venice move closer

AMD says the investment is also meant to support the rollout of its Helios rack-scale platform in the second half of this year. That platform will include Instinct MI450X GPUs, alongside the company’s wider AI infrastructure push.

At the same time, AMD and TSMC have reached another production milestone, with Venice Epyc CPUs now ramping on TSMC’s 2nm process in Taiwan. AMD also says it plans to ramp production at TSMC’s Arizona fab in the future.

The bigger picture

For AMD, this is a clear signal that the company wants to stay active across both chip design and the manufacturing ecosystem behind it. The scale of the investment underlines how much AI hardware depends on packaging, interconnects and foundry capacity, not just on the chips themselves.

Source

Source: PC Gamer Hardware