Intel’s process tech catch-up looks unlikely

Intel’s hopes of overtaking TSMC in process technology this year or next year have taken a knock after a TSMC statement about its A16 process from CEO C.C.Wei that: “Volume production is scheduled for second half 2026.”

Meanwhile, rumours are resurrected about a takeover.

Intel hopes to start to ramp its 18A process at its Oregon development fab in H2 2025 which would suggest it will not be in volume production until 2026, while TSMC says it will be in high volume production on its equivalent N2 process in H2 2025.


Intel’s next process after 18A, 14A, is being developed using High-NA EUV at Oregon. There have been no recent projections about when, or if, 14A will be production-ready.


TSMC’s A16 is equivalent to Intel’s 14A and set to deliver an improvement over N2 of 8-10% in speedIntel’s process tech catch-up looks unlikely and 7-10% density. When A16 is in high volume production next year,  Intel could be a generation behind.

With hindsight, Intel’s problems started with its 2021 stated intention to catch up in process tech in four years. As Malcolm Penn, CEO of Future Horizons, said at last week’s IFS2025: “Once you fall behind it’s awfully difficult to catch up. The rule of thumb is it takes a decade.”

Last week, SemiAccurate reported it had seen an email saying that a potential buyer for the whole operation with enough money to pay for it was interested in acquiring the company,

Speculation about  possible buyers includes: Apple, Nvidia, Softbank.  Broadcom Apollo and Brookfield

There are indications that investors are getting fed up with  supporting Intel’s attempt to catch up in process technology which has seen the shares drop from 60 to 20 in the last three years

David Manners

David Manners

David Manners has more than forty-years experience writing about the electronics industry, its major trends and leading players. As well as writing business, components and research news, he is the author of the site's most popular blog, Mannerisms. This features series of posts such as Fables, Markets, Shenanigans, and Memory Lanes, across a wide range of topics.

Comments

No comments

  1. I don’t really see TSMC A16 as equivalent to Intel 14A. Ignoring the silly naming games our industry is famous for, Intel’s 18A already has backside-powering whereas TSMC A16 will be their first with it. So as with most other industry innovations (immersion, Hi-K, FinFet, etc) it is Intel going first and the others following. I suspect TSMC will hit some of the same problems Intel has hit with it and have to introduce a half-step fixing them ASAP.

    But agree Intel then make a total mess-up of getting a process into mass production, as we can see with some product cancellations and delays on shipping 18A devices in volume.

Leave a Reply

Your email address will not be published. Required fields are marked *

*