There was the steady drip drip of stories about how far ahead of schedule TSMC’s 2nm process was proceeding and now another pall of gloom must have hit Santa Clara with TSMC’s announcement that it’s ahead of schedule with its 1.4nm process.
This year was the year Intel said it would take process technology leadership away from TSMC but now that now looks like a demented pipe-dream and Intel’s new aspiration is to try and get foundry customers for its 1.4nm process.
With TSMC saying it is going to start building a 1.4nm fab next month, a good many seeds of doubt will have been spread about Intel’s 1.4nm aspirations as well.
If Intel can’t show, at the very least, comparable process technology to TSMC, in the same time-frame, with a commercial yield, a similar price and a high volume capability it won’t have much of a chance of getting a significant customer for 1.4nm.
Meanwhile, another expected announcement out of Taipei which will add to Intel’s gloom is that TSMC will be putting its advanced process wafer price up 5-10% next year.
With Intel depending on TSMC to make its most advanced products, and with Intel’s gross margin already down from the heady days of 60%+ to 27.5% in Q2, a 10% increase in its wafer cost is going to hit Intel where it hurts – in the return on revenue it needs to make to fund its process development effort.
With Intel budgeting only $18 billion in capex this year compared to TSMC’s $38-42 billion, the contest for 1.4nm foundry customers already looks uneven and the Chips Act $5.7 billion won’t go far enough towards closing the gap.
Somehow you get the impression that someone, somewhere is steadily tightening the screw on Intel.