There is a positive sign with Intel's next-generation 14A process. The company's management recently stated that the defect density of this process is decreasing faster than their internal targets, indicating that their wafer manufacturing capabilities are improving. However, the foundry business has not yet managed to overcome significant losses, and there have been no decisive results in expanding external customer base.
14A Defect Density Improvement
Intel's Chief Financial Officer David Zinsner stated at the Deutsche Bank Technology Conference that the defect density of 14A performs better than the company's internal target curve. This is a rare sign of significant improvement for Intel since the 22-nanometer manufacturing process.
A decrease in defect density usually means that there is a greater chance of increasing the number of usable chips per wafer. However, this does not equate to the final yield meeting the standards. Chip area, design complexity, and defect distribution all affect whether a marketable product can ultimately be produced.
Mass production is still scheduled to begin after 2027.
According to Intel's plan, 14A will enter risk production in the second half of 2027, with the goal of achieving high-volume manufacturing in 2028. This process will combine second-generation RibbonFET transistors, PowerDirect backside power supply, as well as High - NA EUV lithography technology.
- Compared to 18A, performance is improved by 15% to 20% under the same power consumption.
- Under the same performance, power consumption is reduced by 25% to 35%.
- Transistor density increased by approximately 30%
However, these are still company goals and not independent test results for chips that have already been mass-produced.
External customers remain key.
From a financial data perspective, Intel's foundry business reported revenue of $5.765 billion in the second quarter, which is higher than the $4.417 billion in the same period last year; however, the department still incurred an operating loss of $2.089 billion for that quarter, with the loss amount accounting for more than 30% of its business revenue.
What is more noteworthy is that only $293 million of this revenue comes from external customers, while approximately $5.5 billion is related to internal business settlements within the company, mainly associated with Intel's own products. This means that whether the foundry business can truly regain its competitiveness depends on whether external customers are willing to entrust Intel with more chip designs for production.
Zinsner indicates that the discussions among potential 14A customers have progressed from viewing technical data to inquiring about how much production capacity Intel can provide. However, as of now, Intel has not announced any large-scale external production contracts related to 14A.
Intel's recent performance improvement has also been driven by the rebound in demand for AI chips. The company's revenue in the second quarter increased by 25% year-on-year to $16.1 billion. For this company, whether 14A can become a turning point for its foundry business depends not only on whether defects continue to decrease, but also on whether yield rates, external orders, and narrowing losses can progress in tandem.
