Intel’s incoming CEO, Lip-Bu Tan, is considering major changes to the company’s chip manufacturing strategy and artificial intelligence (AI) roadmap.
According to sources cited by Reuters, Tan is looking to streamline operations and refocus Intel’s manufacturing business in an effort to revitalise the struggling technology giant.
Among the key priorities under consideration are restructuring Intel’s AI approach and reducing what he reportedly sees as an overly complex and slow-moving management structure. Sources indicate that Tan intends to overhaul the company’s foundry operations, which had historically been dedicated to producing chips solely for Intel but in recent years have opened up to external clients such as NVIDIA.
The strategy, while still in development, could include cost-cutting measures and efforts to improve Intel’s chip manufacturing processes. Investors responded positively to the news, with Intel’s shares rising by more than 8% during mid-day trading on Nasdaq.
In the short term, Tan is expected to focus on improving Intel Foundry’s performance by attracting new customers, including major players such as Microsoft and Amazon. There are also plans to reinvigorate Intel’s AI chip development, looking beyond AI servers into broader applications such as software, robotics, and foundational AI models.
Following his appointment, Tan reportedly addressed Intel employees in a town hall meeting, acknowledging that “tough decisions” would need to be made.
His early approach appears to be an evolution of his predecessor Pat Gelsinger’s turnaround plan, which aimed to position Intel as a competitive contract chip manufacturer alongside Taiwan Semiconductor Manufacturing Co. (TSMC). Gelsinger had invested heavily in new fabrication plants in the US and Europe, but Intel was forced to scale back its ambitions as demand for its core products waned.
Industry analysts believe Intel’s contract manufacturing efforts will only succeed if Tan secures at least two major customers willing to produce chips at high volume. Reuters reported that both NVIDIA and Broadcom have already initiated early test runs with Intel, while AMD is also evaluating its processes.
A key focus for Tan will be improving the manufacturing yield to make Intel a more attractive foundry option.
Looking ahead, Intel aims to transition to an annual AI chip release cycle, mirroring NVIDIA’s approach. However, sources indicate that it could take until at least 2027 for Intel to develop a compelling new AI chip architecture capable of competing in the rapidly evolving market.
