Intel and its manufacturing processes are in a state of flux. While we were told that 18A was the sacrosanct process that would put IFS back on track, it would appear that this is not quite the case. Without actually stopping it, we learn that the blueprints would stop marketing it. Why is this? Simply not enough customers and/or a desire to prioritize its use over internal production (CPU/GPU).
18A: not attractive enough, Intel opts for 14A!
Broadly speaking, 18A is the process Intel will be using for its Panther Lake processors and Xeon Clearwater Forest. The company’s figures for this process seem rather encouraging, with claims of a 25% improvement in performance and a 40% reduction in power consumption, but these figures don’t really seem to be convincing potential customers. A leaked report indicates that the company is planning to shelve the process… Or, to be more precise, to stop marketing it externally. This means that the few customers who use it will see their orders honored, while the rookies will also use it internally for their own CPUs. What’s more, paradoxically enough, variants will continue to be developed, with an 18A-P and 18A-PT scheduled for 2026 and 2028 respectively. By then, however, these processes could be cancelled if the company doesn’t see a market for them. This has already been done with the 20A. Instead, the rookies prefer to concentrate on a more promising node: the 14A. All this is in line with the new policy of the company’s current CEO Lip-Bu Tan. He seems determined to get the company back on track following last year’s colossal loss (2024). This involves a massive redundancy plan to clean up the company’s books. But the company also doesn’t seem to be bent on working on a process, costly to implement, if nobody wants it, or on outsourcing certain services.
