How can Intel be defined other than as an eternally complicated case? Your humble writer has already dwelled on the chip designer and manufacturer's current strategy, drawing up a non-exhaustive list of the many risks facing the group. Unhappy with the current outlook, Lip-Bu Tan, newly appointed head of Intel, is reportedly considering a complete overhaul of the strategy, even if it means abandoning the 18A etching process Pat Gelsinger's flagship technology, to refocus on 14A and attempt to overtake TSMC.

Back to square one?

Sources close to the matter have told Reuters that Lip-Bu Tan is seriously considering a repositioning. Nothing has been decided yet, but the possibility of freezing additional orders for the 18A process is being considered in order to refocus efforts on 14A. Two problems immediately arise: the net loss associated with the unprofitable development of 18A (estimated at between several hundred million and one billion dollars), and the persistent feeling of déjà vu.

This approach is strangely reminiscent of the one advocated by Pat Gelsinger: don't bet on 3 nm in order to better focus on 18A. At the time, Intel had outsourced the production of 3 nm chips to TSMC. In the scenario envisaged today, Intel would honor its current commitments on 18A, including Panther Lake chips and orders already signed, but would no longer open this process to new customers.

Abandoning 18A would be a major decision for Intel, which would then be betting most of its future on the success of 14A. While the strategy is similar to that of its predecessor in substance, it differs in execution. Lip-Bu Tan has overhauled the internal organization, streamlined certain procedures, recruited new engineers, and replaced part of the management team.

A range already buried?

Intel declined to comment on the Reuters report, dismissing the revelations as mere rumors. In a statement, the group said: "Lip-Bu and the management team are committed to strengthening our roadmap, building trust with our customers, and improving our financial position for the future. We have identified key priorities and will take the necessary steps to turn the business around."

Nothing is set in stone yet, but feedback from customers such as Microsoft and Amazon, who are already 18A customers, could prove decisive in future decisions.

According to the same Reuters source, the rapid development of 14A is primarily aimed at offering a more competitive technology than TSMC in order to win back major customers such as Apple and Nvidia. This is one of Intel Foundry's biggest challenges: its first and only external customer... is Intel itself.