
Intel is considering reducing its emphasis on its main chipmaking process in an effort to better compete with rival TSMC and attract major clients, such as Apple and NVIDIA, according to sources. While it has been developing its 18A chipmaking process for several years, some external customers have reportedly grown cautious about the chips it produces.
Why Intel’s CEO is concerned about the 18A chipmaking process
CEO Lip-Bu Tan, who took over from longtime boss Pat Gelsinger in March, has begun expressing concerns about 18A, anonymous sources told Reuters. They added that Intel may focus more on 14A, an alternate fabrication process that some engineers believe offers advantage over TSMC’s offerings.
However, the sources stated that scaling back plans for future sales of 18A chips, as well as its variant 18A-P, would result in Intel incurring a huge financial loss, which industry analysts told Reuters could amount to billions of dollars.
The 18A process involves a new method of delivering energy to chips and a new type of transistor, but it is also comparable to TSMC’s N3 process, which has been producing a high volume of chips since late 2022. Another issue is that 18A has faced rollout delays, potentially related to low yield rates, while TSMC’s N2 process remains on track for production.
What happens to 18A if the proposals are approved
If Tan’s proposals are approved, Intel would stop marketing 18A to new clients and recommend 14A as an alternative, while allocating more foundry staff and design partners to its advancement and support.
Sources told Reuters that even redirecting efforts to 14A at this stage wouldn’t guarantee the process will be production-ready in time to secure contracts from the tech giants, so Intel may ultimately decide against the switch. Tan could discuss the matter at a board meeting this month, but the decision is so significant that it may not be made until autumn.
Regardless of the meeting’s outcome, Intel plans to proceed with 18A chip production plans that are already underway. These include chips designed in-house specifically for the 18A process, as well as a small volume that Intel has promised for Amazon and Microsoft in the near future.
Intel declined to comment on what the sources said when contacted by Reuters. “Lip-Bu and the executive team are committed to strengthening our roadmap, building trust with our customers, and improving our financial position for the future,” a spokesperson told the news outlet. “We have identified clear areas of focus and will take actions needed to turn the business around.”
Tan aims to restore Intel’s reputation
Once a dominant force in the chip industry, Intel has struggled amid the AI boom. Unlike rivals who specialise in either chip design or manufacturing, Intel continues to operate in both sectors, a strategy that has left its fabrication efforts lagging behind TSMC.
In 2024, Intel’s stock declined by 60%, and the company fell from first to second place on Gartner’s list of top global semiconductor vendors by revenue growth. It was also its first unprofitable year since 1986, posting a net loss of $18.8 billion.
Gelsinger, who had led Intel for 30 years, departed after his turnaround plan — focused on funnelling money into new fabs and investing heavily in 18A — failed to deliver significant market share gains or profitability. He was replaced by Tan, who has a background in both chip design and manufacturing, and was expected to revamp Intel’s production strategy.
“Together, we will work hard to restore Intel’s position as a world-class products company, establish ourselves as a world-class foundry and delight our customers like never before,” Tan said in a letter to Intel employees upon his appointment.
In March, he paused on a Gelsinger-era plan to build two advanced semiconductor plants in Ohio citing limited immediate need. The following month, Tan announced a sweeping transformation plan that included hiring new engineering talent, flattening team hierarchies, increasing in-office workdays, and eliminating unnecessary internal processes such as superfluous reporting and meetings. He also revealed his intention to spin off assets that aren’t part of Intel’s core mission, and expand its core business using AI and Software 2.0, using AI to write code.
Under Tan’s leadership, the company has reportedly agreed to let TSMC take over some of its fabs, following months of rumours that it would be split between the Taiwanese chipmaker and Broadcom, or that NVIDIA and AMD would also be involved in a joint acquisition. If it comes to fruition, TSMC will hold a 20% stake in the company and provide value through sharing its chipmaking practices and training Intel staff. TSMC has denied any involvement in a joint venture.
Read more about Intel’s new project with SoftBank and SAI.TECH to develop a new kind of memory chip.