TSMC and Intel are competitors in the semiconductor market.
TSMC has long dominated the world's advanced manufacturing processes, but Intel has launched a counterattack by relying on the 18A process. It not only introduced two new technologies, RibbonFET and PowerVia, but also offered a lower price than TSMC to grab customers. In addition, the advantages of domestic manufacturing in the United States have allowed the market to re-see Intel's competitiveness. However, 18A mass production is only the first step. Yield, customers and ecosystem are the keys to whether Intel can truly shake up TSMC.
TradingKey reported that 18A is seen as an important battle for Intel to re-establish its competitiveness in advanced processes. 18A uses RibbonFET full surround gate transistor and PowerVia back power supply technology to improve chip performance and energy efficiency through changes in transistor structure and power supply method.
Where does the competitiveness of Intel 18A come from? At present, 18A has entered the mass production stage, and the initial products include Intel’s own Panther Lake processor. Market news points out that the current yield rate of 18A is about 55% to 60%. Although it is still far from mature large-scale mass production levels, it is enough to support initial shipments and customer verification.
For Intel, the fact that 18A can officially enter mass production is an important progress in itself, because it means that the company has regained a certain degree of say in advanced process technology.
In addition to new process technology, the main selling points of 18A include price and the advantages of domestic manufacturing in the United States. The U.S. government has actively promoted the reshoring of semiconductor manufacturing in recent years, creating more opportunities for Intel.
However, the start of mass production of 18A does not mean that Intel has succeeded. The real key lies in whether it can obtain enough external customers. Panther Lake can help verify the yield and performance of 18A, but the foundry still needs to improve the ecosystem of PDK, IP and design tools. There have been reports in the market that Broadcom has reservations about some processes and tools after testing.
In addition, the current 18A yield rate is about 55% to 60%, and output efficiency still needs to be further improved to reduce costs and ensure stable delivery. In terms of finance, Intel Foundry is still in the red. If there are insufficient external orders and low production capacity utilization, huge depreciation and R&D expenditures will continue to drag down profits. Therefore, improving yield rate, expanding external customers and improving foundry gross profit are the keys to whether 18A can lead Intel to turn around.
The report pointed out that from a technical perspective, Intel 18A has allowed Intel to regain the qualification to participate in advanced process competition, "but it is still very difficult to challenge TSMC." TSMC’s advantages are not just the process itself, but also include years of accumulated high-yield mass production experience, a huge customer base, and advanced packaging ecosystems such as CoWoS and 3DFabric.
For AI chip customers, the ability to obtain advanced packaging, HBM integration and stable production capacity is as important as process efficiency. In addition, customers such as Apple, Huida, AMD, Qualcomm and Broadcom have established long-term cooperation with TSMC. If chip designers want to change foundries, they must re-design and verify, so they will not transfer orders on a large scale just because the price is lower.
Therefore, 18A is more likely to become an important second source of supply besides TSMC in the short term, rather than directly replacing TSMC. In comparison, Intel 18A may put more direct pressure on Samsung first.
